AGCO Corporation (NYSE: AGCO) — A European Franchise Carrying an American Loss, Priced as Neither

Report date: 2026-08-22. Filings and transcripts accessed 2026-08-22; market and factor data as of the 2026-08-21 close ($106.81). Latest reported quarter: Q2 2026 (10-Q filed 2026-07-30). Every figure below is sourced to a public filing, release, transcript or dataset, cited inline and listed in Appendix B.

⚡ Claude’s Take

This section is the author’s own subjective opinion, offered as general information and not investment advice. It is the only part of this article that takes a position; the analysis in sections 1–15 below is deliberately position-free and sets no price target.

Verdict: BUY the cycle, not the bounce — accumulate on weakness below ~$100, size up below ~$90; at $106.81, three sessions after an 8.5% rip, I would be adding slowly rather than reaching. Directional valuation zone: on mid-cycle earnings power of ~$11.00–13.00 per share (management’s own “industry at ~85% of mid-cycle” anchor, run through a 10.0–11.0% mid-cycle margin) at a 11–13x mid-cycle multiple — a deliberate discount to the 13.6–15.3x Deere commands on its own mid-cycle framework — fair value lands around ~$125–155. Today’s $106.81 is ~6.0x mid-cycle EBITDA and ~8.4x mid-cycle EPS. The labeled scenario grid spans ~$80 bear / ~$127 base / ~$165 bull, so the price sits meaningfully below the base case. That is the bull argument, and it is a real one.

The framing is deep-cyclical value with a quality core, not a compounder — and the single most important thing to know is that the reported earnings are a mirage in both directions. FY2025’s headline GAAP EPS of $9.75 is more than half non-operating: a $251.9M gain on selling AGCO’s TAFE stake plus a $179.8M federal tax benefit from a legal-entity reorganization (the 10-K’s words, not mine). The vendor screens have not caught this — AZI reports trailing EPS of $10.44 and a P/E in the 16th percentile of AGCO’s own history, i.e. “screamingly cheap.” Rebuilt from the filings, trailing EPS is $7.21 and the true P/E is ~14.8x; the honest own-history read is the P/S percentile at 68.7, which says the stock is richer than two-thirds of its own past on sales. Anyone buying this on the screen P/E is buying a number that does not exist. Meanwhile the operating business underneath is genuinely split: Europe/Middle East earned $1,006.8M on $6,736.7M of sales — a 14.9% margin and 169% of consolidated operating income — while North America lost $112.0M. Fendt is a real franchise; the American business is a subscale challenger against a ~50–60%-share incumbent, and it is currently burning money.

So why constructive? Because the cyclical evidence is turning while the price still embeds the trough. Deere called fiscal 2026 the ag-cycle bottom on 2026-08-20 and pointed to improving 2027 orders; dealer inventories are falling in all three major regions (Europe and Latin America at ~3.5 months, North America to just below 7 against a 6-month target); FY2026 guidance embeds no further IEEPA tariff refunds despite $22M already recovered and claims outstanding; the buyback took the share count from 72.6M to 70.0M in six months; and — the datum I weight most — director Bob De Lange has bought stock on the open market five times since August 2024, including 1,000 shares at $100.71 on 2026-08-14, eight days ago and within two sessions of the 52-week low. In a corpus of 294 Form 4s containing only six open-market purchases, that is the closest thing to a conviction signal this sector produces.

Conviction: medium. The honest bear risks are specific, not vague: PTx North America’s goodwill headroom falls to ~3% on a one-point increase in the discount rate (the company’s own disclosed sensitivity — a live impairment trigger at trough demand); H1-2026 operating cash flow was −$245.0M against +$153.5M a year earlier, so FY2025’s $740M of free cash flow was destock-flattery, not run-rate; and management cut guidance on 2026-07-30. Flips me more bullish: North America returning to operating profit — the single cleanest proof that the loss is cyclical rather than structural. Flips me bearish: a PTx goodwill impairment, or a second consecutive guidance cut that pushes the recovery past 2027. The market bought the headline and is now selling the reality; the business is better than the tape and worse than the screen.

📈 Stock Price Action — Five-Year Event Map

Prices are AZI dividend-adjusted closes (pulled 2026-08-22). Price moves are FACT; attributed causes are INTERPRETATION, cross-referenced to the 8-K corpus, earnings releases and the news feed. No price target, no recommendation.

AGCO has round-tripped a full agricultural cycle and then some. From ~$86 in early 2021 the stock ran to a five-year closing high of $139.71 on 2026-02-13 — the day the FY2025 10-K was filed — before falling 23.5% to $106.81 (2026-08-21). Along the way it bottomed at $74.58 on 2025-04-08 in the tariff crash. The 52-week range is $98.45–$139.71, and the 52-week low was set four sessions ago, on 2026-08-18. Calendar-year total returns tell the cycle plainly: 2021 +16.8%, 2022 +23.7%, 2023 −6.9%, 2024 −20.8%, 2025 +15.5%, 2026 YTD +1.7% — a stock that has gone nowhere for five years while the underlying business rose 30% and then fell 30%.

Cycle narrative.

1. The boom (2021–early 2023). Record grain prices lifted US net farm income to historic highs and equipment demand with it. AGCO’s revenue compounded from $11.1B to $14.4B and GAAP EPS reached $15.63 in FY2023, with ROIC of 22.9% — genuinely franchise-level returns at the peak (FACT — EDGAR XBRL; ROIC.ai).
2. Ukraine (March 2022). AGCO is the most European of the three majors, with two-thirds of sales in EME; the invasion hit it harder on the day than its US-centric peers, and the exposure has been a standing discount factor since.
3. The down-cycle (2023–2024). Crop prices retreated, the replacement trigger switched off, and a glut of used high-horsepower equipment blocked new sales. FY2024 revenue fell 19.1% and the company posted a GAAP loss of $(424.8)M on a $354.1M goodwill impairment and $157.8M of restructuring (FACT — EDGAR XBRL).
4. The tariff crash (April 2025). The five-year low of $74.58 was set on 2025-04-08 in a market-wide event, then recovered 13.3% on the April-9 pause. Pure macro beta, not company news.
5. The false dawn (H2 2025 – February 2026). The stock rallied ~35% into a five-year high of $139.71 on 2026-02-13. The proximate catalysts were real — the TAFE dispute settled with mutual releases and AGCO’s stake sold for $260M; a $1.0B buyback authorized in July 2025 with a $250M ASR in November — but the peak coincided precisely with the FY2025 10-K reporting $9.75 of headline EPS. More than half of that figure was the $251.9M TAFE gain and the $179.8M legal-entity tax benefit (FACT — FY2025 10-K). INTERPRETATION: the market re-rated on an earnings number that the operating business did not produce.
6. The fade and the cut (February–July 2026). As 2026 guidance framed a second trough year, the stock gave back 17%, then fell a further 7.8% on 2026-07-30 when management cut the adjusted-EPS outlook to $5.50–5.75, citing “weaker-than-expected industry conditions, currency fluctuations and a more cautious outlook” (FACT — Q2-2026 earnings release).
7. The Deere bounce (2026-08-20/21). Four sessions ago the stock set a 52-week low of $98.45; two sessions later Deere reported a Q3 beat, raised its outlook and called fiscal 2026 the ag-cycle bottom, and the entire ag complex jumped. INTERPRETATION — and this matters: Deere’s beat was driven by Construction & Forestry, a segment AGCO does not own, with Production Agriculture still weak. The read-across to a pure-play ag manufacturer is therefore weaker than the +5.4% move implies. That CAT fell 3.3% over the same window confirms the move was ag-sentiment-specific rather than machinery-wide.

📊 Price & Factor Positioning — Quantitative Overlay

Sources: AZI price CSV and the FactorsToday factor model, both pulled 2026-08-22 (loadings dated 2026-08-21); AZI valuation_index dated 2026-08-21. Third-party computations, not primary data. Loadings, returns and drawdowns are FACT; regime and continuation reads are INTERPRETATION, labeled. An overlay subordinate to the thesis — no price targets, no chart-pattern reading.

Trend and position (FACT)

• Close 2026-08-21: $106.81 — above the 21-day EMA ($104.66) but below the 50-day ($109.10) and 200-day ($112.03) EMAs. The three-session bounce has recovered the shortest average only; the stock remains in a downtrend on the intermediate and long averages.
• −23.5% below the five-year closing high ($139.71, 2026-02-13). The 52-week low of $98.45 was set on 2026-08-18 — four sessions ago. 52-week range $98.45–$139.71.
• Relative strength: rs_6m −22.1, rs_12m −4.4, rs_ytd +3.2, rs_peak −23.6. Beta 0.98; alpha −0.221 — negative.
• Dividend yield 1.10% ($1.16 annualized, $0.29/quarter).

Risk-adjusted track record (FactorsToday leaderboard, 2026-08-22; every horizon annualized)

This is the profile of a beaten-down cyclical, not a compounder and not a momentum vehicle. Three-, five- and one-year returns are all negative; Sharpe is negative at every horizon out to five years. Only the ten-year and lifetime windows are positive, and the lifetime maximum drawdown of −78.7% is the honest statement of what an agricultural down-cycle can do to this equity. INTERPRETATION: an investor here is underwriting cycle mean-reversion, and the historical record says the ride is violent.

Factor identity (FactorsToday; read within one model — betas are not comparable across nested models)

Base model (style tilts net of Market and macro; R² = 0.322): Market +0.83, DividendYield +0.70, SmallSize +0.30, Value +0.21, Quality +0.04, Momentum −0.04.

All-Factors model (thematic attribution; R² = 0.465): Market +0.88, Sector: Industrials +0.67, DividendYield +0.59, SmallSize +0.35, Sector: Materials +0.29, Value +0.21, Sector: Financials −0.28, Industry: Aerospace & Defense −0.31.

• What AGCO is in factor space (FACT): a market-sensitive industrial cyclical with a dividend-yield and small-size tilt and a mild value lean. Momentum is effectively zero (−0.04) and Quality is effectively zero (+0.04). The Materials loading (+0.29) is the closest thing the model has to an agricultural-commodity proxy — there is no dedicated ag factor.
• INTERPRETATION: the zero Momentum loading is informative in both directions. AGCO is not a crowded momentum trade that can unwind (the KGS failure mode), and it is not yet classified as a deep-value name despite a 23.5% drawdown — the Value beta of +0.21 is mild. The market has not made up its mind what this is, which is consistent with a stock that has gone nowhere for five years.
• Idiosyncrasy (FACT): All-Factors R² of 0.465 leaves ~53% of variance stock-specific; specific volatility 27.2% annualized. Segment execution — specifically whether North America stops losing money — is a company-level story the factor model cannot price.

Factor-similar peers (comp-set cross-check)

Cosine similarity: CNH 0.964 (top match), TEX 0.955, ST 0.937, KMT 0.936, TKR 0.936, MSM 0.933, HLIO 0.927, DE 0.916, VEGI 0.915 (the global agribusiness ETF), RRX 0.914, PCAR 0.913, TTC 0.894. This factor-validates the fundamental comp set: CNH is the closest statistical twin and Deere the relevant quality benchmark, exactly as the fundamental analysis assumes. It is also a two-way confirmation — prior published analysis of CNH found AGCO at 0.95 similarity from that side.

Current factor regime (factor-returns, pulled 2026-08-22; |z| ≥ 2 = extreme)

• The regime is currently working in AGCO’s favour, which is unusual for this name. Value 21d +2.3% (z +0.75), 63d +5.3% (z +1.03); SmallSize 21d +2.7% (z +0.69), 63d +8.1% (z +1.24); DividendYield 63d +5.7% (z +1.00). Those are AGCO’s three positive style loadings, and all three are being paid.
• The offsetting headwind is Sector: Industrials at 21d −2.9% (z −0.98), against AGCO’s +0.67 loading. Momentum is falling (21d −3.1%, z −1.22) but AGCO’s near-zero loading makes that close to irrelevant here — a genuine difference from a high-momentum name.
• INTERPRETATION (regime-caveated): for the first time in this cycle the style backdrop suits the stock. That is a tailwind to a re-rating, not a cause of one — the fundamental trigger remains farm income, and it has not turned.

Own-history valuation percentiles (AZI valuation_index, dated 2026-08-21) — read with care

Sanity check passes on price: AZI latest price $106.81 ties exactly to the CSV close. The percentiles do not pass: composite 39.5th, P/E 10.23x @ 15.8th, P/B 1.80x @ 33.9th, P/S 0.76x @ 68.7th (n = 3, ~20-year history).

The P/E leg is unusable and the divergence is the tell. AZI’s trailing EPS input of $10.4365 cannot be reconciled to the filings: rebuilt from reported quarterly diluted EPS (Q3-2025 $4.09 + Q4-2025 ~$1.28 + Q1-2026 $0.76 + Q2-2026 $1.08) trailing EPS is $7.21, making the true trailing P/E ~14.8x, not 10.23x. And even $7.21 is inflated, because Q3-2025 still carries part of the FY2025 legal-entity tax benefit. A P/E in the 16th percentile of own history is therefore an artifact twice over. Read P/S at the 68.7th percentile as the honest own-history signal — its denominator does not collapse or inflate — with P/B at the 33.9th as the middle view. The three metrics disagree by more than 50 percentile points, which is itself diagnostic: when P/E and P/S point in opposite directions this violently, the earnings denominator is the thing that is broken.

Synthesis (INTERPRETATION, for Claude’s Take and Variant Perception)

AGCO is a five-year-flat industrial cyclical sitting four sessions off a 52-week low, with negative alpha, negative Sharpe out to five years, no momentum loading and only a mild value tilt — a stock the market has neither crowded into nor capitulated on. The tape has two things going for it that it did not have six months ago: the style regime (value, small size, dividend yield) is paying, and the sector sentiment trigger fired on 2026-08-20 when Deere called the ag-cycle bottom. It has one thing going against it: the bounce came on a peer’s beat driven by a construction segment AGCO does not own, so the +5.4% move rests on borrowed evidence. This is neither a falling knife (the 200-day is 4.9% away, not 40%) nor a trend to ride (the 50- and 200-day both sit above the price) — it is a cyclical at the point of maximum narrative ambiguity, which is generally where the returns are made and lost.

1. Executive Summary

AGCO is the world’s third-largest agricultural-equipment manufacturer and the only pure-play among the majors: Deere and CNH both carry construction franchises, AGCO does not. It sells tractors (66% of FY2025 sales), combines (2%), hay/forage, planters, implements and precision-agriculture technology through independent dealers in roughly 140 countries under four principal brands — Fendt (the European premium franchise), Massey Ferguson, Valtra, and the PTx precision-ag stack (Precision Planting plus the 85%-owned PTx Trimble joint venture). Farm equipment is financed through AGCO Finance, a joint venture with Rabobank accounted for under the equity method — an important structural difference from CNH, whose consolidated captive book makes its headline leverage and enterprise value analytically useless. AGCO’s balance sheet requires no such surgery.

The company is two businesses wearing one ticker, and the segment disclosure is stark. In FY2025, Europe/Middle East generated $6,736.7M of sales and $1,006.8M of operating income — a 14.9% margin and 169% of consolidated GAAP operating income of $595.7M. North America generated $1,665.5M of sales and an operating loss of $(112.0)M, a $196.0M swing from the prior year’s $84.0M profit. South America contributed $51.4M on $1,115.6M (4.6%). Fendt in Europe is a genuine premium franchise with demonstrated pricing power; the North American business is a subscale challenger against an incumbent holding roughly 50–60% of large-tractor and combine share, and at trough volumes it does not cover its fixed cost. Any thesis on AGCO is a judgment about whether that American loss is cyclical or structural.

The industry is in a deep, well-documented trough that has not yet turned. AGCO’s revenue fell from $14,412.4M in FY2023 to $11,661.9M in FY2024 (−19.1%) and $10,082.0M in FY2025 (−13.5%) — 30.0% below peak. Management states industry demand is running at ~85% of mid-cycle. The demand trigger is net farm income, and USDA’s 2026 forecast of $153.4B is down 2.6% in real terms, with direct government payments up 45.2%, farm debt up 5.2% and working capital down 9.2% — support, not health. Both Deere and CNH frame 2027, not 2026, as the earliest recovery. On 2026-08-20 Deere called fiscal 2026 the ag-cycle bottom and pointed to improving 2027 orders, lifting the whole complex; the read-across is real but qualified, since Deere’s beat came from Construction & Forestry, which AGCO does not own.

The reported earnings must be normalized before any valuation work is credible. FY2025 GAAP diluted EPS of $9.75 contains a $251.9M pre-tax gain on the sale of AGCO’s stake in Tractors and Farm Equipment Limited (TAFE) and an income tax benefit of $(77.4)M that, per the 10-K, “includes a net federal tax benefit of $179.8 million related to a legal entity reorganization” plus roughly $82.0M of further reorganization-related benefits. Together those approximate $4.90–5.00 per share — more than half the headline. The quarterly pattern is unmistakable: Q2-2025 reported EPS of $4.22 against operating income of only $164.0M. Vendor screens have not adjusted: AZI reports trailing EPS of $10.4365 and a P/E in the 15.8th percentile of AGCO’s own ~20-year history. Rebuilt from the filings, trailing EPS is $7.21 and the true trailing P/E is ~14.8x. The honest own-history read is the P/S percentile at 68.7 — richer than two-thirds of its own past.

Financial quality is mixed and the cash flow is the weak point. Gross margin has held remarkably well through a 30% revenue collapse (26.2% FY2023 → 25.5% FY2025), which is evidence of genuine pricing discipline. Peak-cycle returns were franchise-grade — ROIC of 22.9% and ROE of 27.4% in FY2023. But FY2025 free cash flow of $740.2M (operating cash flow $988.1M less capex $247.9M) was destock-flattered: H1-2026 operating cash flow was −$245.0M against +$153.5M in H1-2025, a $398.5M adverse swing, and management’s own FY2026 conversion target of 75–100% of adjusted net income implies roughly $300–400M for the year — under half of FY2025. Engineering expense is simultaneously rising into the trough, from $487.7M in FY2025 to a guided ~5% of sales, because precision agriculture is the strategic bet and management is refusing to cut it.

Capital allocation is better than the sector norm on returns of capital and unresolved on the big acquisition. A $1.0B buyback was authorized in July 2025 and a $250M accelerated repurchase executed in November; shares outstanding fell from 72,629,310 to 70,002,207 between 2025-12-31 and 2026-06-30 (−3.6% in six months). The decade-long TAFE dispute was settled with mutual releases and the stake monetized for $260M. Against that, FY2024 carried a $354.1M goodwill impairment and $157.8M of restructuring, and the PTx Trimble bet now carries a disclosed and uncomfortable sensitivity: a one-point increase in the discount rate would cut the PTx North America reporting unit’s goodwill headroom to approximately 3%. The insider record is the sector’s best: of 294 Form 4s containing only six open-market purchases, five belong to director Bob De Lange — including 1,000 shares at $100.71 on 2026-08-14, within two sessions of the 52-week low.

Valuation. At $106.81 the enterprise value is ~$9.54B (70.0M shares, net debt $2,066.1M) — ~0.94x FY2026E sales, ~8.7x FY2026E EBITDA and 18.6–19.4x the cut FY2026 adjusted-EPS guidance of $5.50–5.75. That is expensive on trough earnings and cheap on normalized ones: mid-cycle EPS of ~$11–13 (derived from management’s own 85%-of-mid-cycle anchor) puts the stock at ~8.4x mid-cycle, against the 13.6–15.3x Deere commands on its own mid-cycle framework. The scenario grid in section 10 spans ~$80 bear / ~$127 base / ~$165 bull. The variant perception, in one sentence: the market re-rated this stock in February on an EPS number that was half one-time items, has spent six months un-rating it, and in the process has arrived at a price that under-rates the mid-cycle earnings power of a business whose European core earns 14.9% margins.

The analysis below takes no position and sets no price target; the single labeled exception is the opinion block above.

2. Business Overview

What AGCO does

AGCO Corporation designs, manufactures and distributes agricultural machinery and precision-agriculture technology, selling through independent dealers and distributors in approximately 140 countries (FACT — FY2025 10-K). It is the third-largest player in the global ag-equipment oligopoly behind Deere and CNH, and the only pure-play: Deere carries Construction & Forestry, CNH carries a construction segment currently under strategic review, AGCO carries neither. That purity is the single most important structural fact about the company — it means no cushion when agriculture turns down, and maximum operating leverage when it turns up. The August 2026 episode illustrated it precisely: Deere’s Q3 beat was driven by construction demand tied to data centers and energy infrastructure, an earnings stream AGCO simply does not have.

Revenue by product (FY2025, percentage of net sales — FACT, 10-K):

Two things are visible in that table. Tractors have risen from 61% to 66% of the mix while combines have halved from 4% to 2% — combines are the most discretionary, highest-ticket item in the catalogue and are the first thing farmers stop buying. And the residual category shrank, which is where precision agriculture sits; the mix has concentrated toward the core tractor franchise as the cycle deepened.

The brands, and which one matters

• Fendt — the premium European tractor brand, and the profit engine of the entire company. Fendt commands genuine price premiums in Germany, France and the Nordics and is being rolled out into North and South America as a “full-line” premium offering.
• Massey Ferguson — the global volume brand, strongest in emerging markets, Brazil and India, and the brand at the centre of the TAFE dispute.
• Valtra — Nordic and Latin American strength, a specialist utility/forestry position.
• PTx — the precision-agriculture stack, combining Precision Planting (retrofit planter technology) with PTx Trimble, the 85%-owned joint venture formed in April 2024 from Trimble’s agriculture assets. PTx is explicitly retrofit-first and mixed-fleet: the 10-K describes solutions that “enable farmers who use almost any brand” to operate, with “retrofit, factory-fit and Original Equipment Manufacturer (‘OEM’) solutions that work across mixed fleets.” Precision-ag revenue ran approximately $860 million in 2025, guided flat to modestly up for 2026 (FACT — Q2-2026 earnings call).

How it makes money, and the shape of the revenue

AGCO sells equipment to independent dealers, who sell to farmers. The 10-Q is explicit about the mechanics that matter: “To the extent practicable, we attempt to sell products to our dealers and distributors on a level basis throughout the year,” and there is a timing gap “between our sale of the equipment to the dealer and the dealer’s sale to a retail customer.” Reported revenue is therefore wholesale shipments, not retail demand — and in a down-cycle the two diverge violently, because the manufacturer must under-ship retail to drain dealer inventory. That is precisely what AGCO is doing now, and it is why reported revenue has fallen faster than the underlying retail market. It also means reported revenue will over-shoot on the way up, once destocking ends (INTERPRETATION).

Revenue quality sits in three tiers:

1. Whole goods (the ~68% in tractors and combines) — cyclical, big-ticket, financed, and deferrable. Farmers can and do run equipment an extra season.
2. Replacement parts and service — the annuity layer. AGCO does not disclose parts revenue separately, which is a genuine disclosure gap relative to Deere, but the installed base across 140 countries generates a materially more stable revenue stream that partially offsets whole-goods cyclicality (INTERPRETATION — the mechanism is industry-standard; the magnitude is not disclosed).
3. Precision agriculture (~$860M) — the strategic growth layer, sold both factory-fit and, unusually, retrofit into competitors’ fleets. The Q2-2026 call noted that “the retrofit channel has continued to stay relatively strong compared to the overall OEM industry” — a meaningful observation, because retrofit demand is partially counter-cyclical: a farmer who will not buy a new tractor may still buy technology to make the old one more productive.

Geography — and why the segment table is the whole story

AGCO reports four geographic segments. FY2025 (FACT — 10-K MD&A):

The arithmetic deserves emphasis: EME alone earns $1,006.8M — 169% of the consolidated $595.7M — and the gap is closed by North America’s loss, a small South American contribution, Asia/Pacific/Africa, and corporate and unallocated costs (amortization, restructuring, stock compensation and central overhead; the APA figure is derived by subtraction and the residual is not separately disclosed in the extracted MD&A — ASSUMPTION flagged). Over two-thirds of revenue and effectively all of the operating profit comes from Europe. AGCO is, financially, a European tractor company with a North American challenger operation attached and a Brazilian business that swings with soft-commodity economics.

That concentration cuts both ways and both must be stated. On the positive side, EME’s 14.9% margin is Deere-grade profitability and it rose in FY2025 — income up $45.5M on roughly flat sales, “as a result of positive net pricing, partially offset by lower sales and production volumes and higher warranty costs” (FACT — 10-K). A franchise that holds and improves price through a downturn is demonstrating something real. On the negative side, it is a single-region dependency: European farm policy, European farm income and the euro sit directly on AGCO’s earnings, and the March-2022 invasion shock (−9.3% in a single session) is a reminder that the exposure is geopolitical as well as economic.

North America deserves its own sentence. A $196.0M year-over-year swing into a $(112.0)M loss on $1,665.5M of sales, attributed in the 10-K simply to “lower sales and production volumes,” is what negative operating leverage looks like in a business that must keep dealer, engineering and distribution infrastructure in place through the trough. North American sales fell 27.5% in FY2025 alone.

Customers, seasonality and financing

End customers are farmers, and their capital spending is driven almost entirely by net farm income and the tax treatment of equipment purchases. There is no meaningful customer concentration at the dealer level. Sales are seasonal, weighted to planting and harvest cycles by hemisphere — which is one reason AGCO’s Brazilian business partially offsets Northern-Hemisphere seasonality.

Financing runs through AGCO Finance, a joint venture with Coöperatieve Rabobank U.A., accounted for under the equity method and contributing $39.6M of equity income in FY2025 ($46.4M FY2024, $68.2M FY2023 — FACT, EDGAR XBRL). This is a genuine structural advantage for analytical purposes: AGCO’s reported net debt of ~$2.07B is industrial net debt, requiring none of the disaggregation that makes CNH’s consolidated $23.4B figure analytically useless. It is also a modest strategic disadvantage — AGCO does not capture the finance spread Deere’s captive earns, and it has less control over retail financing terms when it wants to stimulate demand. The equity-income line falling 42% from $68.2M to $39.6M across the down-cycle is itself a clean read on farmer credit conditions.

Verdict

AGCO is a structurally coherent but geographically lopsided pure-play on global agriculture. The business model — wholesale equipment through independent dealers, supported by a parts annuity and an increasingly important retrofit-first technology stack — is the industry standard, and AGCO executes it well enough to hold a 14.9% operating margin in its core European market through a 30% revenue collapse. What it is not is diversified: there is no construction segment to cushion the trough, no consolidated finance arm to smooth earnings, and two-thirds of the revenue with effectively all of the profit sits in one region. The reported-revenue line measures shipments rather than retail demand, so both the reported decline and the eventual recovery overstate the underlying swing. Investors are buying concentrated, unhedged, high-operating-leverage exposure to the world agricultural cycle, run by a management team that has kept its European franchise’s pricing intact while its American operation moved into loss. That is a legitimate thing to own at the right price; it is not a defensive holding and should never be mistaken for one.

3. Industry Dynamics — Global Agricultural Equipment

Structural framing draws on the primary filings plus prior published analysis of Deere (2026-08-14) and CNH (2026-07-17), both of which analysed the same industry from the #1 and #2 positions. No suitable third-party industry primer was available for this sector, so the structure below is built from filings and peer analysis.

3.1 Structure: a profitable oligopoly with a very unequal profit pool

Global agricultural equipment is a genuine oligopoly. Deere is the clear #1, holding roughly 50–60% of North American large tractors and combines; CNH (Case IH / New Holland) is #2; AGCO (Fendt / Massey Ferguson / Valtra) is #3; Kubota dominates small-ag and compact at low cost; CLAAS is a strong European private competitor, particularly in harvesting; Mahindra and regional players fill out emerging markets. Barriers to entry are high and well understood: dealer density, brand, emissions-compliance R&D scale, precision-technology integration, and sheer capital intensity.

The critical structural feature is that the profit pool is skewed dramatically toward the #1. Deere’s peak-cycle equipment operating margins run in the high teens to low twenties; AGCO’s consolidated FY2023 peak was 11.9%; CNH’s FY2025 trough industrial EBIT margin was 4.3% against Deere’s ~12.6% in the same period. Scale in this industry is not merely helpful — it compounds through dealer economics (a denser dealer network sells more per outlet, supports better parts availability, and attracts the better dealers), through R&D amortization (Deere spends roughly $2.3B a year against CNH’s ~$1.0B and AGCO’s ~$0.5B), and through the precision-ag data flywheel, where installed base begets data begets product quality.

The one place the hierarchy is genuinely contestable is Europe, where Deere is not dominant, CLAAS is strong, and Fendt is a premium brand in its own right. AGCO’s 14.9% EME operating margin is the direct financial evidence: in its home market, AGCO earns margins comparable to the global leader’s. That is not true anywhere else it operates.

3.2 Market size, and where we are in the cycle

The demand driver is net farm income, which determines farmer capital spending with a lag. The current position is unambiguous and documented from multiple directions:

• USDA forecasts 2026 US net farm income of $153.4B, down 2.6% in real terms, with crop-profit margins the poorest since the 2016–2020 stretch. Corn has been trading around $4.20–4.50 — well below the levels that triggered the 2021–23 boom (FACT — USDA data as compiled in prior published analysis of Deere and CNH).
• The composition of that income is deteriorating even as the headline holds: direct government payments up 45.2%, farm debt up 5.2%, and working capital down 9.2%. INTERPRETATION: farm income is being propped by transfer payments while farmer balance sheets weaken — which supports current consumption but not a capital-spending recovery.
• Industry unit declines have been severe. Peer disclosure documents North American declines on the order of 27–39% in four-wheel-drive tractors, 20%+ in high-horsepower tractors and 5–10% in combines across recent periods.
• A glut of used high-horsepower equipment blocks new sales: the used market must clear before new orders resume, and used values have to stabilise before trade-in economics work.
• Management’s own anchor: industry demand is running at “around 85% of mid-cycle demand” (FACT — AGCO Q2-2026 earnings call). This is the single most useful normalization datum in the report and section 10 uses it directly.

Both Deere and CNH have framed 2027, not 2026, as the earliest recovery. On 2026-08-20 Deere went further and called fiscal 2026 the ag-cycle bottom, pointing to improving 2027 order trends — the proximate cause of AGCO’s 8.5% three-session bounce. The qualification, again, is that Deere’s actual earnings beat came from Construction & Forestry, not Production Agriculture.

3.3 Competitive intensity and pricing

Competitive intensity is moderate and, importantly, rational. The striking evidence is that AGCO’s gross margin barely moved through a 30% revenue collapse: 26.2% (FY2023) → 24.9% (FY2024) → 25.5% (FY2025) (FACT — ROIC.ai, reconciled to filings). A 130-basis-point round trip on a 30% volume decline is remarkable, and it tells us something specific: the majors have not resorted to destructive price competition to defend volume. AGCO’s EME segment income actually rose $45.5M in FY2025 on “positive net pricing.”

What the industry does instead of discounting is cut production — under-shipping retail demand to drain dealer inventories. AGCO is explicit about it: production hours are guided “slightly lower than 2025 levels as we continue to align our output with retail demand.” That is the disciplined response, and it protects price at the cost of absorption (which is precisely what drove North America to a loss).

Where competition is intensifying is precision agriculture, and this is the strategic battleground. Third-party estimates put Deere’s precision-ag share at ~15–18%, AGCO’s at ~7–9%, and CNH’s lower and unquantified. Deere’s strategy is closed and integrated (its own machines, its own tech); AGCO’s is deliberately the opposite — retrofit-first, mixed-fleet, “works across almost any brand.” This is a genuinely differentiated go-to-market and the one place AGCO is arguably ahead of CNH, which bought its position (Raven, $2.1B in 2021) and subsequently took a $172M IPR&D impairment on it.

3.4 Regulation and structural factors

• Emissions compliance (EU Stage V, US Tier 4 Final and successors) is a scale barrier: engine and aftertreatment R&D is a fixed cost spread over units, and it structurally disadvantages the smallest players.
• Tariffs are now a first-order earnings variable. AGCO expects gross tariff-related costs of approximately $115 million in 2026, reduced to a net $95 million by $22M of recovered IEEPA refunds — and that net figure is an increase of $50 million versus 2025 (FACT — Q2-2026 call). The legal position is fluid: the US Supreme Court ruled in February 2026 that IEEPA does not authorize the tariffs; the Court of International Trade ordered refunds on 2026-03-04 and expanded the scope on 2026-03-27; the refund-filing system opened 2026-04-20; and the administration has both appealed to the Federal Circuit and imposed replacement tariffs under alternative statutes (FACT — Q2-2026 10-Q). See section 8 and section 10 for why this creates unrecognized optionality.
• Right-to-repair is a live regulatory theme across the industry, addressed at length in the prior Deere work. It bears less directly on AGCO, whose retrofit-first, open-architecture technology strategy is commercially aligned with the direction of regulation rather than against it (INTERPRETATION) — a small but real strategic advantage over Deere’s closed system.
• European agricultural policy (CAP payments) underwrites the farm incomes that drive two-thirds of AGCO’s revenue. It is stable but politically contestable, and a materially less generous CAP settlement would hit AGCO harder than either peer.

3.5 Capital-cycle read (Marathon)

Marathon’s supply-side lens asks where capital is going and what returns it is chasing. The reading here is favourable, and it is the strongest structural argument for owning any of these names:

Supply is being withdrawn, not added. All three majors are cutting production, and AGCO has taken $157.8M (FY2024) and $46.2M (FY2025) of restructuring charges plus a $354.1M goodwill impairment — capacity and capitalised value coming out of the system. Nobody is building tractor plants. There is no capital-market lubrication: no ag-equipment IPO wave, no private-equity entrant, no capacity flooding in. Industry M&A has been consolidative (AGCO/Trimble ag, CNH/Raven) rather than capacity-additive.

Demand is depressed but not structurally impaired. The world still needs food; the installed fleet still ages; the replacement cycle is deferred, not cancelled. What is missing is the trigger — farm income — and the USDA data says it has not turned.

Where Marathon counsels caution: the aging-fleet argument has been made by bulls for two years and the trigger still has not flipped, and government transfer payments propping farm income are not the same as a genuine demand recovery. INTERPRETATION: this is a classic Marathon setup — supply discipline is real and the returns on incremental capital in the industry are being protected — but the timing is unresolved, and the honest position is that the capital cycle says the eventual recovery will be profitable without saying when it starts.

3.6 Verdict: a structurally good industry, entered at the wrong point of the cycle by anyone who bought early

Agricultural equipment passes the structural tests convincingly. It is a consolidated oligopoly with high and durable entry barriers (dealer density, brand, emissions R&D scale, precision integration, capital intensity), share that moves only slowly and mostly through M&A, and — the strongest evidence — demonstrated pricing discipline through a 30% volume collapse, with AGCO’s gross margin moving only 130 basis points and its European segment margin actually improving. That is not a commodity industry; commodity industries surrender price when volume goes.

The honest offsets are three. First, the profit pool is grossly unequal — Deere takes a disproportionate share, and the #3 position earns structurally lower through-cycle returns, which is why AGCO’s FY2023 peak ROIC of 22.9% sits below what Deere earns and why its trough is a loss-making North American segment rather than a merely thinner margin. Second, the demand driver is exogenous and currently negative: net farm income is forecast down in real terms, supported increasingly by government payments while farm debt rises and working capital falls; no amount of manufacturer execution changes that. Third, the cycle’s timing is genuinely unknown — 2027 is the consensus earliest recovery, that consensus has already slipped once, and a second slip would extend a trough that has already run two years.

Net: a structurally attractive industry in the late stage of a painful down-cycle, where supply discipline has protected pricing and the returns will be good when demand returns — but where the return date is the variable that matters and nobody, including management, can currently name it.

4. Competitive Position

The moat question in Greenwald’s taxonomy

Run the three-advantage test honestly, and run it by region, because AGCO does not have one competitive position — it has two very different ones.

Supply/cost advantages — largely absent. AGCO buys engines, transmissions and components from the same supplier base as its competitors, and in fact purchases fully manufactured tractors from third parties — the 10-K names TAFE (until termination), SDF S.p.A., Carraro S.p.A. and Iseki & Company. A manufacturer that outsources whole units is not claiming a manufacturing cost advantage. AGCO’s R&D of ~$488M (FY2025 engineering expense) is roughly 21% of Deere’s ~$2.3B — a structural scale disadvantage in the most R&D-intensive part of the product (emissions, autonomy, precision). There is no cost moat here; if anything there is a cost disadvantage versus the leader.

Demand advantages (customer captivity) — present, real, and strongly regional. This is where the franchise lives, and the evidence is financial rather than asserted:

• Fendt in Europe commands price. EME operating margin of 14.9% in FY2025, with segment income rising $45.5M on “positive net pricing” during the worst ag downturn in a decade, is the single most persuasive moat datum in this report. Customers paid more for Fendt in a year when they were buying less of everything. Brands without pricing power do not do that.
• Dealer relationships and switching costs are genuine but bounded. A farmer’s choice of brand is substantially a choice of dealer — parts availability, service response during harvest, and the trade-in relationship. Downtime during a planting or harvest window is catastrophically expensive, so service density outranks list price. This creates real captivity, measured in equipment-replacement cycles (5–10 years) rather than decades.
• The parts annuity reinforces it. An installed base across 140 countries generates recurring parts and service demand that follows the machine, not the next purchase decision.

Economies of scale plus captivity — present in Europe, absent in North America. This is the crux. Greenwald’s strongest advantage type is local scale combined with captivity, and AGCO has exactly that in Europe: dense dealer coverage, a premium brand, and the leading position in several national markets. In North America it has the mirror image — subscale distribution against an incumbent holding 50–60% share, which is precisely why the segment lost $112.0M on $1,665.5M of sales in FY2025 while EME earned 14.9%. The same company, the same products, the same management: the difference is local scale. That is Greenwald’s thesis demonstrated inside a single income statement.

Network effects — one genuine candidate, and it is not yet earning. Precision agriculture is the only place a real data network effect can exist in this industry: more machines generate more agronomic data, which improves the models, which improves the product. Deere has it and AGCO does not, at least not yet — third-party estimates put Deere’s precision-ag share at ~15–18% against AGCO’s ~7–9%. Do not underwrite an AGCO network effect on current evidence. What AGCO has instead is a deliberately different architecture, discussed below.

The PTx bet: retrofit-first as a flanking strategy

AGCO’s precision strategy is explicitly mixed-fleet and retrofit-first — the 10-K describes technology that “enable[s] farmers who use almost any brand to increase profitability,” delivered through “retrofit, factory-fit and OEM solutions that work across mixed fleets.” Precision-ag revenue is roughly $860M (2025), guided flat to modestly up for 2026.

The strategic logic is sound and worth stating properly. Deere’s closed system monetizes only Deere’s installed base. AGCO’s open system addresses the entire global fleet, including Deere’s — a far larger addressable market, and one that is partially counter-cyclical, because a farmer deferring a $600,000 tractor may still spend $40,000 making the existing one more productive. Management confirmed the mechanism on the Q2-2026 call: “the retrofit channel has continued to stay relatively strong compared to the overall OEM industry.” It is also, incidentally, aligned with the direction of right-to-repair regulation, where Deere’s closed architecture is exposed.

The honest counter is threefold. First, retrofit is a harder commercial motion: no captive dealer channel pushing it, longer sales cycles, and a customer who must be convinced rather than defaulted into the purchase. Second, $860M against a ~$10.1B revenue base is 8.5% — strategically central, financially not yet decisive. Third, and most concretely, the market’s carrying value of this bet is under visible strain: the FY2025 10-K’s critical audit matter discloses that a one-point increase in the discount rate would reduce the PTx North America reporting unit’s goodwill headroom to approximately 3%. A strategy this important should not be one rate-assumption away from an impairment two years after the deal closed.

Market-share stability and ROIC tests

Share stability — passes, at the industry level. Leadership positions in ag equipment have been stable for decades; share moves through acquisition (AGCO/Trimble ag 2024, CNH/Raven 2021) rather than competitive displacement. AGCO has held the #3 global position throughout, and management claims share gains in the current downturn — “gaining market share in key regions… including high-horsepower offerings in North America” (Q2-2026 call), with the caveat that the same call conceded losing “a little share in France” while gaining in Germany. INTERPRETATION: modest share gains are credible and consistent with the price discipline evidence, but they are not transformational and they are self-reported.

ROIC test — passes at the peak, fails through the cycle. The record (ROIC.ai, third-party, reconciled directionally to filings):

*FY2025 ROE is flattered by the one-time TAFE gain and tax benefit; on normalized earnings it is roughly 11–12% (INTERPRETATION). Note also that ROIC.ai’s operating-margin field is an adjusted figure — it reports +7.95% for FY2024 against an actual GAAP operating loss — so the margin column above is computed from EDGAR XBRL, not the vendor.

A peak ROIC of 22.9% genuinely clears Greenwald’s 15–25% franchise bar. The problem is that it clears it only at the peak. Averaged across FY2020–FY2025 the returns are mid-single-digit-to-low-teens, and two of six years produced returns below any reasonable cost of capital. That is the signature of a good business in a cyclical industry, not a franchise — the distinction matters, because a franchise earns excess returns through the cycle and AGCO does not.

Head-to-head

The last row is the thesis in one line. The market prices AGCO essentially identically to CNH (0.76x vs 0.73x sales) and at one-sixth of Deere’s multiple. Yet on the evidence above AGCO is not CNH: it has a genuine premium franchise earning 14.9% margins in its largest region, roughly double CNH’s precision-ag share, a differentiated open-architecture technology strategy, a peak ROIC of 22.9%, and a clean balance sheet requiring no captive-finance disaggregation. CNH’s own trough industrial margin was 4.3% against AGCO’s 5.9% consolidated — and AGCO’s includes a loss-making North America.

Verdict

A real but narrow and regionally-confined advantage — Fendt-plus-dealer-density in Europe — inside a good industry, attached to a subscale North American operation that is currently destroying value. The moat is identifiable, financially visible and falsifiable: if it were absent, EME could not have raised segment income on positive net pricing during the worst downturn in a decade, and consolidated gross margin could not have held within 130 basis points through a 30% revenue collapse. Those are the two facts that keep this out of the “no moat” bucket.

Three qualifications keep it out of the “durable franchise” bucket. (1) The advantage does not travel. The same company earns 14.9% in Europe and −6.7% in North America; local scale, not corporate capability, is doing the work. (2) Returns clear the franchise bar only at the peak. ROIC of 22.9% in FY2023 against negative returns in FY2024 averages to something ordinary, and an investor is compensated for cyclicality, not for a moat. (3) The one genuine network-effect opportunity belongs to a competitor. Deere has roughly twice AGCO’s precision-ag share and five times its R&D budget; AGCO’s open-architecture counter-strategy is intelligent and differentiated but remains 8.5% of revenue with a goodwill cushion the company itself discloses as thin.

Stated plainly: AGCO is a better business than CNH and a materially worse one than Deere, and it is priced as though the first half of that sentence were untrue.

5. Growth History and Forward Opportunities

The historical record: one boom, one bust, no secular growth

(FACT — EDGAR XBRL companyfacts and the Q2-2026 earnings release.)

The honest reading of that table is that AGCO has not grown. FY2026E revenue of ~$10.15B is 12% above FY2019’s $9.04B — a compound growth rate of about 1.7% a year across seven years — and that overstates organic performance, because the period includes the PTx Trimble acquisition and excludes the divested Grain & Protein business. This is a cyclical business oscillating around a slowly-rising base, not a growth company, and it should never be valued as one.

What has improved is the structural quality of the revenue. FY2019’s $9.0B carried a 3.9% operating margin; FY2026E’s ~$10.15B carries a guided 7.5% adjusted margin at 85% of mid-cycle demand. The portfolio surgery is the reason: the majority of the low-margin Grain & Protein business was divested in 2024, precision agriculture was added at higher margin, and the mix shifted toward Fendt. Comparing trough to trough rather than peak to peak is the fair test, and on that test the business is meaningfully better than it was.

Organic versus acquired

• FY2021–FY2023 growth was overwhelmingly organic and price-led — the ag boom lifted volumes and AGCO took price. Operating income nearly tripled from $599.7M to $1,700.4M on a 58% revenue increase, an incremental margin above 30%.
• FY2024 was the portfolio pivot: AGCO “fundamentally shifted our portfolio through the PTx Trimble joint venture and the divestiture of the majority of our Grain & Protein business” (10-K). PTx Trimble added precision-ag revenue at 85% ownership; G&P removed low-margin revenue. The net revenue effect was negative; the net margin effect was positive.
• FY2025–FY2026 is pure cycle, with no material M&A and revenue determined by farm economics and destocking.

Forward opportunity 1: the cyclical recovery (the largest and least controllable)

Management states industry demand is at ~85% of mid-cycle. Simple arithmetic: if AGCO’s FY2026E revenue of ~$10.15B corresponds to 85% of mid-cycle, mid-cycle revenue is roughly $11.9B — about $1.75B of latent revenue that returns with the cycle rather than with any management action. At incremental margins of 25% (conservative against the 30%+ achieved on the way up), that is roughly $440M of incremental operating income, taking mid-cycle operating income to ~$1.2B and mid-cycle margin to ~10%. This is the dominant driver of any AGCO investment case, and section 10 builds the valuation on it.

The supporting evidence that the cycle is bottoming rather than deepening:

• Dealer inventories are falling in all three major regions: Europe to ~3.5 months (from just under 4 in Q1), Latin America to ~3.5 months (from 4), North America to just below 7 months against a 6-month target (FACT — Q2-2026 call). Europe and Latin America are approaching normal; North America is not, and remains the laggard.
• H1-2026 net sales were $4,952.6M, +5.7% versus H1-2025 — the first period of year-over-year growth since the peak (though Q2 alone was −1.0%, so the growth was front-loaded into Q1).
• Deere called fiscal 2026 the ag-cycle bottom on 2026-08-20 and cited improving 2027 order trends.

The counter-evidence, which is equally concrete: management cut FY2026 guidance on 2026-07-30 citing “weaker-than-expected industry conditions… and a more cautious outlook”; USDA forecasts 2026 real net farm income down 2.6%; and production hours are guided lower than 2025.

Forward opportunity 2: precision agriculture

Precision ag is ~$860M of revenue (2025), guided flat to modestly up in 2026, against Deere’s ~15–18% category share versus AGCO’s ~7–9%. The growth thesis rests on the retrofit-first, mixed-fleet architecture addressing the entire global installed base rather than only AGCO’s own machines — a structurally larger market than Deere’s closed system can serve, with partially counter-cyclical demand. Management continues to fund it aggressively: engineering expense rose to $487.7M in FY2025 and to $273.8M in H1-2026 (+17.1% year over year), guided to ~5% of sales — increasing R&D intensity at the bottom of a cycle while cutting production. New product evidence includes the Outrun autonomous retrofit platform, with tillage in beta and fertilization in alpha testing (FACT — 10-K).

INTERPRETATION: this is the right strategy and it is being funded with conviction, but it is not yet moving the consolidated needle. At 8.5% of revenue, flat year over year, with a goodwill cushion the company discloses as thin, precision ag is currently a cost line supporting an option. Treat it as optionality, not as an earnings driver, until the revenue inflects.

Forward opportunity 3: North American margin recovery

The clearest company-specific opportunity is simply North America ceasing to lose money. The segment swung $196.0M into a $(112.0)M loss in FY2025. Restoring even a mid-single-digit margin on $1,665.5M of sales would add roughly $200M of operating income — equivalent to about 20% of consolidated FY2025 operating profit — without requiring any growth in Europe. The Q2-2026 call reported “gaining share in key markets, including high-horsepower offerings in North America,” and a $22M IEEPA tariff refund landed entirely in the North American segment. This is the metric to watch, and section 14 makes it the primary bull falsification test.

Forward opportunity 4: the buyback

With shares down from 72,629,310 to 70,002,207 in six months and $1.0B authorized, share-count reduction is a genuine per-share growth lever at a depressed price. At the current pace the count is shrinking roughly 5–7% annually. INTERPRETATION: buying back stock at ~8x mid-cycle earnings is accretive if the mid-cycle estimate is right — and value-destructive if the cycle does not turn and management is buying a permanently impaired earnings stream. It is a leveraged bet on the same call the equity holder is making.

Verdict

Low-quality growth history, genuine mid-cycle recovery potential, and one strategically important but financially immaterial growth engine. AGCO has compounded revenue at under 2% a year since FY2019 and is not a growth company by any definition; anyone underwriting secular growth here is mistaken. What the investment case actually rests on is cyclical normalization — roughly $1.75B of latent revenue and ~$440M of latent operating income implied by management’s own 85%-of-mid-cycle statement — supplemented by two company-specific levers: North America returning to profitability (worth ~$200M, entirely within management’s control given the share-gain evidence) and continued share repurchase at a depressed price.

Precision agriculture is the strategically correct long-term bet, differentiated from Deere by architecture rather than scale, and management deserves credit for increasing engineering spend through the trough rather than harvesting it. But at 8.5% of revenue and flat year over year, it earns the label option, not driver. The quality of the growth on offer is therefore entirely a function of whether the cycle turns: if it does, the incremental margins are excellent and the operating leverage is real; if it does not, there is no secular engine underneath to carry the company. That is an honest, unhedged cyclical proposition, and it should be underwritten as one.

6. Financial Quality

All figures reconcile to SEC filings (EDGAR XBRL and the FY2025 10-K / Q2-2026 10-Q). ROIC.ai is used as a cross-check only, and its divergences are documented below. Accessed 2026-08-22.

6.1 The headline earnings are not real — establish this before anything else

This is the most important analytical point in the report, and every valuation conclusion depends on it.

FY2025 reported GAAP diluted EPS was $9.75. More than half of it was non-operating. Two items:

1. A $251.9 million pre-tax gain on the sale of AGCO’s ownership interest in Tractors and Farm Equipment Limited (TAFE). Under a Buyback Agreement, TAFE repurchased AGCO’s shareholding for $260 million, completing 2025-09-30, with after-tax proceeds of approximately $230 million. The gain was booked in “Other expense (income), net” (FACT — FY2025 10-K, Note 18). The near-equivalence of the $260M price and the $251.9M gain tells you the stake was carried at close to zero — this was a legacy holding, not an operating asset.
2. An income tax benefit of $(77.4) million for the full year, versus a $98.4M expense in FY2024 and $230.4M in FY2023. The 10-K is explicit: the provision “includes a net federal tax benefit of $179.8 million related to a legal entity reorganization, which excludes approximately $82.0 million of additional benefits related to a change in the Company’s permanent reinvestment assertion and state tax benefits associated with the reorganization” (FACT).

Together these approximate $4.90–5.00 per share of the $9.75 reported — the TAFE gain contributing roughly $2.50 after tax and the disclosed federal tax benefit roughly $2.41 on ~74.5M average diluted shares.

The quarterly pattern proves it beyond argument. Compare operating income to net income:

In Q2 and Q3 of 2025, net income was roughly double operating income — before interest expense. That is arithmetically impossible from operations. In Q1 and Q2 of 2026 the relationship is normal. The 2025 “earnings” were a tax reorganization and an asset sale.

Consequence — the vendor screens are wrong. AZI’s valuation_index reports trailing EPS of $10.4365, a P/E of 10.23x, and a P/E percentile of 15.8 — i.e. cheaper than 84% of AGCO’s own ~20-year history. Rebuilt from reported quarterly diluted EPS (Q3-2025 $4.09 + Q4-2025 ~$1.28 + Q1-2026 $0.76 + Q2-2026 $1.08), trailing EPS is $7.21 and the true trailing P/E at $106.81 is ~14.8x. Cross-checked on net income: $305.7M + ~$95.5M + $55.0M + $77.2M = $533.4M. And even $7.21 remains inflated, because Q3-2025 still carries part of the reorganization benefit.

The three own-history percentiles disagree by more than fifty points — P/E 15.8, P/B 33.9, P/S 68.7 — and that disagreement is itself the diagnosis. When the sales-based and earnings-based measures point in opposite directions this violently, the earnings denominator is broken. Read P/S at the 68.7th percentile as the honest own-history signal. On sales, AGCO is richer than roughly two-thirds of its own past — the precise opposite of what the P/E screen advertises.

6.2 Revenue, margins and the pricing evidence

(Gross margin FY2021–FY2025 per ROIC.ai; all other figures EDGAR XBRL or the Q2-2026 10-Q. H1 operating income derived from the 10-Q income statement.)

The pricing evidence is the strongest quality signal in the financials. Gross margin went 26.2% → 24.9% → 25.5% across a revenue decline from $14,412.4M to $10,082.0M — a 30% volume collapse absorbed with 130 basis points of gross-margin movement, and margin actually recovered in the second down year. Businesses without pricing power do not do that; they discount to hold volume and the gross line collapses. AGCO instead cut production. EME segment income rose $45.5M in FY2025 on “positive net pricing.”

Below gross margin, the operating leverage is brutal in the other direction. GAAP operating margin fell from 11.8% to −1.0% to 5.9%. The gap between a resilient gross margin and a collapsing operating margin is fixed cost: engineering ($487.7M), SG&A, and factory absorption. AGCO is deliberately not cutting the first of these — engineering rose 17.1% year over year in H1-2026 — which suppresses current earnings in favour of the precision-ag option.

Q2-2026 detail (FACT — 10-Q and earnings release): net sales $2,609.7M, −1.0% year over year and −3.7% excluding a 2.7% favourable currency translation. Gross profit $645.9M (24.75%) versus $658.6M (25.0%). Reported operating margin 5.4%, down 80bp; adjusted operating margin 6.6%, down 170bp. Reported EPS $1.08; adjusted EPS $1.43.

6.3 The tariff refund — a peer-wide flattery that must be normalized out

Machinery earnings across the sector are currently flattered by non-recurring IEEPA tariff refunds running through adjusted results, and AGCO is no exception. The specifics (FACT — Q2-2026 call and 10-Q):

• Gross tariff-related costs of approximately $115 million expected in 2026.
• $22 million of IEEPA tariff refunds recognized in Q2-2026, reducing the net tariff impact to $95 million for the year — itself an increase of $50 million versus 2025.
• The $22M landed entirely in the North America segment: “North America operating results were generally in line with the prior [year] including a benefit of approximately $22 million from certain IEEPA tariff refunds.”

Normalize it. On Q2 sales of $2,609.7M, $22M is 84 basis points of margin. The reported adjusted operating margin of 6.6% is therefore approximately 5.8% ex-refund — meaning the year-over-year decline was closer to 250bp than the reported 170bp. Per share, $22M pre-tax on 71.1M diluted shares is roughly $0.31 before tax and ~$0.23 after — about 16% of the quarter’s $1.43 adjusted EPS.

Two mitigating facts deserve equal weight. First, this is smaller and better-flagged than at peers: Deere’s comparable quarter included a $272M tariff refund. Second — and this is genuinely favourable — AGCO’s FY2026 guidance explicitly excludes any further recovery: the net $95M figure “does not assume any potential benefits related to future IEEPA refunds,” while additional claims remain filed and pending CBP review. The guidance is clean of optimism on this line, which creates unrecognized upside (see section 10).

6.4 Cash flow — the weakest part of the story

FY2025’s $740.2M of free cash flow — a 9.9% yield on today’s market capitalisation — is not a run-rate, and treating it as one is the single easiest mistake to make on this name. It was produced by destocking: as revenue fell 13.5%, receivables and inventory released cash. The reversal is already visible. H1-2026 operating cash flow was $(245.0)M against $+153.5M in H1-2025 — a $398.5 million adverse swing in a period when sales actually rose 5.7%. Working capital that released cash on the way down consumes it on the way back up. This is the identical pattern flagged on CNH, where FY2025 industrial free cash flow was destock-flattered and reversed to consumption as production normalized.

Management’s own FY2026 target is free-cash-flow conversion of 75–100% of adjusted net income. On guided adjusted EPS of $5.50–5.75 and ~70.5M shares, adjusted net income is roughly $390–405M, implying FY2026 free cash flow of roughly $300–400M — under half of FY2025’s $740.2M. Capital expenditure guidance was reduced to $300–325M, itself a demand signal. The honest free-cash-flow yield is therefore ~4.0–5.3% on the market capitalisation, not 9.9%.

6.5 Balance sheet

At 2026-06-30 (FACT — Q2-2026 10-Q debt note and balance sheet):

Net debt of ~$2.07B against FY2026E adjusted EBITDA of roughly $1.09B is about 1.9x — comfortable, and importantly this is industrial net debt requiring none of the captive-finance disaggregation that makes CNH’s balance sheet unusable. Liquidity is ample: $960.0M available under the $1.25B revolver plus a €200M uncommitted facility. The maturity ladder is manageable, though the $400M 2027 notes and the 2027 revolver expiry cluster.

Two observations. The 0.800% notes due 2028 ($683.9M) are legacy cheap euro debt that will refinance at a materially higher coupon — a modest, dateable earnings headwind. And goodwill of $1,883.2M represents 46% of stockholders’ equity, which is where section 9’s impairment discussion begins.

6.6 Returns on capital

Peak-cycle returns were genuinely franchise-grade — ROIC 22.9% and ROE 27.4% in FY2023 — but they do not persist. FY2024 produced a −10.1% ROE; FY2025’s reported 18.1% ROE is flattered by the TAFE gain and tax benefit and is closer to 11–12% normalized (INTERPRETATION). Averaged across FY2020–FY2025 the returns are ordinary. This is a good business in a cyclical industry, not a compounder, and the return profile should be underwritten on a full-cycle average rather than on either the peak or the trough.

A methodological warning that matters: ROIC.ai’s oper_margin field is an adjusted figure, not GAAP. It reports FY2024 operating margin of +7.95% against an actual GAAP operating loss of $(122.1)M, and FY2025 of 6.93% against GAAP 5.91%. Any analysis built on the vendor’s margin series will systematically overstate profitability at exactly the points in the cycle where accuracy matters most. All margin figures in this article are computed from EDGAR XBRL.

Verdict: do economics improve with scale?

Yes at the gross-margin line, emphatically no at the operating line, and the cash flow is currently the weakest link.

The evidence for quality is real and should not be dismissed: a 30% revenue collapse absorbed with only 130 basis points of gross-margin erosion, a European segment that raised its operating income on positive net pricing during the worst of it, a peak ROIC of 22.9%, a clean and comfortably-levered balance sheet at 1.9x with no captive-finance obfuscation, and a management team funding engineering upward through the trough rather than harvesting it.

The evidence against is equally concrete. GAAP operating margin swung from 11.8% to −1.0% to 5.9%, which is the mathematical statement that this company’s fixed-cost base is large relative to its gross profit. Free cash flow of $740.2M in FY2025 was destock-flattery and has already reversed to a $(245.0)M H1-2026 outflow, with the company’s own conversion target implying roughly half the FY2025 figure for the full year. And the reported earnings themselves have been unreliable — FY2025’s headline EPS was over half non-operating, a distortion the vendor screens have propagated rather than caught.

The single sentence: AGCO’s unit economics hold up under stress far better than its consolidated earnings or cash flow suggest, but the operating leverage runs violently in both directions and nothing in the FY2025 reported numbers should be extrapolated without first stripping out $430M-plus of one-time items.

7. Capital Allocation

AGCO’s capital-allocation record over the past three years is unusually eventful for a company of its size: a portfolio restructured through one large acquisition and one large divestiture, a decade-old shareholder dispute settled and monetized, a $354.1M goodwill impairment taken, a $1.0B buyback authorized and started, and a dividend held flat. The verdict is mixed but tilts positive — with one large unresolved item.

7.1 The 2024 portfolio pivot

The FY2025 10-K states it directly: “In 2024, we fundamentally shifted our portfolio through the PTx Trimble joint venture and the divestiture of the majority of our Grain & Protein (‘G&P’) business.”

PTx Trimble (April 2024) — the big bet, and the open question. AGCO acquired 85% of a joint venture formed from Trimble’s agriculture assets and technologies, combining them with Precision Planting under the PTx brand. Strategically this is defensible and arguably clever: it bought AGCO a mixed-fleet, retrofit-first precision-ag platform addressing the entire global installed base rather than only AGCO machines — a structurally different and larger market than Deere’s closed system serves, and a faster route to relevance than out-spending a competitor with 5x the R&D budget.

Financially the verdict is not yet in, and the early evidence is not encouraging. Precision-ag revenue is roughly $860M (2025), guided flat to modestly up in 2026 — no visible inflection two years after closing. Goodwill arising from the JV was assigned to four new reporting units across the geographic segments, and the FY2025 10-K’s critical audit matter discloses that for the PTx North America reporting unit, a one-point increase in the discount rate “would have decreased the headroom to approximately 3%” (a one-point decrease would raise it to ~32%). INTERPRETATION: an acquisition whose carrying value sits that close to impairment on a single assumption, two years in and with revenue flat, has not yet earned its cost of capital. It has not failed either — precision-ag adoption is a decade-long story and management is funding it through the trough. But this is the item on which the capital-allocation verdict will ultimately turn, and today it is unproven. It also sits alongside the $354.1M goodwill impairment taken in FY2024 (with $157.8M of restructuring), which demonstrates that this management team’s acquisition history already contains at least one write-down.

Grain & Protein divestiture (2024) — a good decision. Selling the majority of a lower-margin, non-core business simplified the portfolio and improved mix. FY2019’s operating margin of 3.9% versus FY2026E’s guided 7.5% adjusted margin at 85% of mid-cycle demand is the cumulative evidence that the portfolio work raised the structural margin, though PTx and pricing also contribute.

7.2 The TAFE resolution — genuinely good work

The dispute with Tractors and Farm Equipment Limited, AGCO’s long-time Indian partner, Massey Ferguson licensee and a substantial AGCO shareholder, was a multi-year overhang involving arbitration, Indian litigation, brand-ownership questions and governance conflict. AGCO terminated all commercial relationships with TAFE on 2024-04-26. The resolution, executed 2025-06-30, comprised an Arbitrations Settlement Agreement and an India Litigation Settlement Agreement under which “the parties agreed to mutually release any and all claims against one another,” an intellectual-property agreement covering the Massey Ferguson brand in India and certain other countries, and a Buyback Agreement under which TAFE repurchased AGCO’s shareholding in TAFE for $260 million — completed 2025-09-30, generating after-tax proceeds of ~$230M and a book gain of $251.9M (FACT — FY2025 10-K, Note 18).

INTERPRETATION: this is clean capital allocation. A non-core, non-controlled legacy stake carried at close to zero was converted into $230M of after-tax cash, litigation risk was extinguished with mutual releases, brand rights were clarified, and a hostile ~10%+ shareholder was removed from the register. The only caveat is a presentational one: the $251.9M gain flowed through GAAP earnings and, as section 6 establishes, has been widely mistaken for operating profit.

7.3 Returns of capital

Buybacks — accelerating, and at the right price. The board authorized a $1.0 billion repurchase program on 2025-07-09 with no expiration. Execution history: $53.0M ASR (November 2023), $22.0M ASR (November 2024), and a $250.0M ASR in November 2025 with initial delivery of 1,997,204 shares. The effect on the share count is material and recent: 72,629,310 shares outstanding at 2025-12-31 fell to 70,002,207 at 2026-06-30 — a 3.6% reduction in six months, implying roughly $260M of H1-2026 repurchase at an average around $100.

INTERPRETATION: buying back ~5–7% of the company annually at ~8x mid-cycle earnings is accretive if the mid-cycle estimate is right, and it is the single highest-return use of capital available to management at this price — better than acquisitions at this point in the cycle. The honest caveat is that it is a leveraged bet on the same cyclical call the equity holder is making: if the cycle does not turn, management will have spent ~$500M buying a permanently impaired earnings stream. Note also the timing pattern — the ASRs cluster in November, and the $250M November-2025 repurchase was executed while the stock was climbing toward its February-2026 five-year high of $139.71, not at the August-2026 lows.

Dividend — flat and small. $0.29 per quarter, $1.16 annualized, a 1.09% yield at $106.81, with a payout ratio around 12% of FY2025 reported earnings. The dividend has not been cut through the down-cycle, which matters, but it is not a meaningful part of the return proposition. Capital is deliberately going to buybacks and engineering instead — the right priority order at this valuation.

7.4 Incentives and alignment — the best-designed plan in recent comparable analysis

The 2026 proxy sets out a compensation structure that is materially better than the machinery-sector norm (FACT — DEF 14A filed 2026-03-12):

• Annual Incentive Plan (cash): 40% adjusted operating margin, 40% Return on Net Assets (RONA), 10% employee engagement, 10% customer satisfaction. Both financial metrics are “adjusted on a sliding scale based upon changes in industry conditions to address agricultural equipment industry cyclicality.”
• Long-Term Incentive Plan: ~60% performance shares plus ~40% RSUs. Performance shares are 50% three-year revenue growth and 50% three-year RONA, subject to a relative TSR modifier of ±20%.
• Double-trigger change-of-control vesting, a NYSE-compliant clawback policy, and stock ownership requirements.

INTERPRETATION: RONA appears in both the annual and the long-term plan, and the TSR modifier is relative, not absolute. That combination is genuinely good design — it pays for return on capital rather than for size, and the relative TSR gate means a management team cannot be rewarded simply for a sector-wide re-rating. It compares favourably with plans built on absolute EBITDA and low absolute-TSR hurdles, which reward empire-building and rising tides.

The one legitimate criticism is the sliding scale. Normalizing targets for industry conditions is intellectually defensible in a cyclical industry — executives should not be punished for weather and grain prices — but it also means the plan can pay out respectably in years when absolute results are poor, and it introduces committee discretion into what is otherwise a formulaic scorecard. A second, smaller point: 20% of the annual bonus rests on “employee engagement” and “customer satisfaction,” which are soft and unverifiable from outside.

7.5 SEC filings sweep — insider transactions and corpus integrity

Corpus. The trailing 60-month EDGAR corpus (since 2021-08-22) was mirrored in full: 413 documents, with the script reporting “Integrity check OK: 413 files on disk for 413 mirrored manifest rows.” Composition: 5 10-Ks, 15 10-Qs, 63 8-Ks, 5 DEF 14As, 294 Form 4s.

Full Form 4 code histogram (all 294 filings): F 223 (tax withholding on vesting), A 215 (grants), S 27, M 13, P 6. Only six open-market purchases exist in the entire five-year corpus.

The buy side is the finding. Five of the six belong to one person — director Bob De Lange — and they are persistent and recent:

De Lange has bought roughly $852,000 of stock across two years, in five separate transactions, at prices from $87.91 to $100.71 — including eight days before this report and within two sessions of the 52-week low of $98.45.

The sell side is dominated by a non-signal. Recorded code-S selling since 2024-01-01 totals ~$111.5M, but 97.7% of it ($108.9M) is TAFE — the former shareholder exiting under the settlement described above. Genuine officer sales are immaterial: SVP/GM Massey Ferguson Felli ~$1.65M, SVP Engineering Bennett ~$572K, SVP CHRO Harris ~$362K. There is no pattern of executive distribution.

INTERPRETATION: this is the best insider picture in recent comparable analysis. On Timken the diagnostic was that not one insider bought at the low; here a sitting director has bought five times across two years including at the current lows, while officer selling is trivial and the one large block is a resolved-dispute exit rather than a conviction signal. It is one director, the sums are small against his likely net worth, and directors are not always right — but in a corpus containing only six purchases in five years, repeated buying by the same person at successively higher and then lower prices is about as clear a revealed preference as this data set produces. Net insider signal: mildly positive.

Verdict

Capital allocation is above average for the sector, with one large unresolved bet. The credit side is substantial: the TAFE dispute settled and the stake converted to $230M of after-tax cash; a low-margin non-core business divested; a $1.0B buyback authorized and executed at a 3.6%-in-six-months pace while the stock trades near multi-year lows; a dividend maintained through the trough; engineering spending increased rather than harvested; a balance sheet held at a comfortable ~1.9x net debt to EBITDA with no captive-finance obfuscation; and an incentive plan built on return on net assets in both the annual and long-term programs with a relative — not absolute — TSR modifier, which is materially better than the peer norm.

The debit side is concentrated in one place. PTx Trimble is the defining capital-allocation decision of this management team and it is not yet working. Two years after closing, precision-ag revenue is flat at ~$860M, and the company’s own disclosure puts the PTx North America reporting unit’s goodwill headroom at ~3% on a one-point discount-rate increase. That sits next to a $354.1M goodwill impairment already taken in FY2024. A management team that allocates capital well in every other respect has placed its largest strategic bet on an asset whose carrying value is one assumption away from a write-down.

The honest summary: management has done the small and medium things notably well and the large thing is still in doubt. [OPEN QUESTION: whether PTx generates a return above its cost of capital — the first real evidence will be precision-ag revenue breaking out of its ~$860M plateau, or an impairment confirming it will not.]

8. Changes and Headwinds — Last Two Years

Window: mid-2024 → August 2026. FACT items trace to SEC filings, earnings releases or transcripts; INTERPRETATION labels judgment. Accessed 2026-08-22.

Timeline of material changes

The changes that matter

1. North America went from profit to loss (weakens, and it is the central fact). A $196.0M swing into a $(112.0)M operating loss on sales down 27.5% is the clearest evidence of AGCO’s structural vulnerability: a subscale position against a ~50–60%-share incumbent does not cover its fixed cost at trough volumes. Management’s response has been production alignment and cost discipline rather than price cutting, which protects the franchise but deepens the absorption hit. INTERPRETATION: this is the swing factor in both directions — the single largest source of latent earnings if it is cyclical, and the single largest structural concern if it is not. section 14 makes it the primary bull test.

2. The February 2026 re-rate rested on a mirage (weakens, and explains the tape). The stock reached a five-year high of $139.71 on 2026-02-13 — the day the FY2025 10-K was filed reporting $9.75 of GAAP EPS. As section 6 establishes, more than half of that was the $251.9M TAFE gain and the $179.8M legal-entity tax benefit. The subsequent 23.5% decline is, in substantial part, the market discovering what it had actually bought. INTERPRETATION: this is why the price action and the fundamentals appear to disagree — the fundamentals never supported the February price, and the current price is being set against a guidance number ($5.50–5.75) that is the first clean earnings figure the company has published in two years.

3. The guidance cut (weakens). On 2026-07-30 management lowered FY2026 adjusted EPS to $5.50–5.75, citing “weaker-than-expected industry conditions, currency fluctuations and a more cautious outlook for the balance of the year,” with farmers facing “elevated operating costs, uneven crop economics and broader macroeconomic uncertainty, resulting in delayed equipment investments and limited visibility into demand recovery.” Production hours are now guided below 2025. INTERPRETATION: the recovery has slipped, and this is the strongest single piece of evidence for the bear case. A second cut would push recovery past 2027 and materially damage the thesis.

4. Tariffs became a first-order earnings variable, with a legal option attached (mixed). Net tariff cost of $95M in 2026 is up $50M year over year — a real and rising headwind against a ~$761M adjusted operating income base. But the legal position inverted in AGCO’s favour in February 2026 when the Supreme Court ruled IEEPA does not authorize the tariffs, and $22M has already been recovered. Critically, guidance “does not assume any potential benefits related to future IEEPA refunds” while further claims remain pending. INTERPRETATION: an unusual situation in which a headwind carries embedded, unrecognized upside. The administration has appealed to the Federal Circuit and imposed replacement tariffs under other statutes, so this is not free money — but the guidance contains no optimism on the line, which is the right way for a company to handle it.

5. The TAFE resolution (strengthens). A decade-old dispute involving arbitration, litigation, brand rights and a hostile ~10%+ shareholder was settled with mutual releases, the stake converted to ~$230M after tax, and the shareholder removed from the register through ~$108.9M of open-market sales. This removed both a legal overhang and a supply of stock. INTERPRETATION: unambiguously good, and its completion means the Form 4 selling data from 2024–2026 should not be read as insider distribution.

6. Deere’s August print (mixed, and widely misread). Deere beat, raised, and called fiscal 2026 the ag-cycle bottom with improving 2027 orders — a genuine sentiment catalyst that moved AGCO +5.4% and CNH +13.7% over six sessions. But Deere’s beat came from Construction & Forestry, driven by data-center and energy-infrastructure demand; Production Agriculture remained weak. That CAT fell 3.3% over the same window confirms the move was ag-sentiment rather than machinery-wide. INTERPRETATION: the directional read-across (the cycle is bottoming) is legitimate and comes from the best-informed operator in the industry. The earnings read-across is not — AGCO has no construction segment to deliver the same beat.

What did not change

No CEO or CFO change: Eric Hansotia remains Chairman, President and Chief Executive Officer throughout the window. No auditor change, no restatement, no covenant issue, no dividend cut. The four-segment geographic reporting structure is unchanged. The Rabobank finance joint venture continues on an equity-method basis. AGCO continues to purchase fully manufactured tractors from third parties (SDF, Carraro, Iseki following the TAFE termination), so the outsourced-manufacturing dependency is unchanged.

Verdict

The trailing two years net-weaken the near-term earnings picture and net-strengthen the structural position — a genuine divergence that explains why the stock and the story feel misaligned. Structurally, AGCO exits this window with a cleaner portfolio (G&P out, PTx in), a settled and monetized TAFE relationship, a $1.0B buyback shrinking the count 3.6% in six months, an improved margin structure, and a tariff-refund claim that guidance ignores. Operationally, it exits with a loss-making North American segment, a cut guidance number, a rising net tariff burden, negative half-year operating cash flow, and a precision-ag bet that has not inflected while its goodwill cushion has thinned to ~3% on a one-point rate move.

The one change that dominates all others is the discovery — by the market, over six months — that FY2025’s headline earnings were half non-operating. That single fact reconciles the 23.5% drawdown from February with a business whose European franchise raised prices through the worst of the cycle. The thesis is now correctly framed as two questions rather than one: does the cycle turn, and does North America come back with it.

9. Risk Analysis

Likelihood = probability of materializing within the investment horizon (~2–3 years); Impact = severity on earnings, cash flow or the thesis if it does. FACT-based evidence; INTERPRETATION where judgment is applied. Accessed 2026-08-22.

Risk matrix

The top risks in prose

1. The cycle fails to turn in 2027 (Medium / High). Everything in the bull case is a claim about mid-cycle normalization: management’s own “~85% of mid-cycle” anchor implies roughly $1.75B of latent revenue and ~$440M of latent operating income. That is worth nothing if the trigger never fires. The trigger is net farm income, and the evidence says it has not: USDA forecasts 2026 US net farm income at $153.4B, down 2.6% in real terms, with crop margins the poorest since 2016–2020 and corn around $4.20–4.50. Worse, the composition is deteriorating — direct government payments up 45.2% while farm debt rises 5.2% and working capital falls 9.2%. INTERPRETATION: transfer payments sustain farm consumption; they do not fund $600,000 tractor replacements. The 2027-recovery consensus has already slipped once (AGCO cut guidance on 2026-07-30 and Deere/CNH have moved their language before), and a second slip would extend a trough that has already run two years and would invalidate the normalization arithmetic on which section 10’s base case rests.

2. PTx North America goodwill impairment (Medium / Medium-High). This is the most specific, most dateable risk in the report, and the company disclosed it itself. The FY2025 10-K’s critical audit matter states that for the PTx North America reporting unit, an increase in the discount rate of 1.0 percent “would have decreased the headroom to approximately 3%” (a 1.0-point decrease would raise it to ~32%). Base-case headroom is therefore modest — somewhere in the mid-teens — and asymmetrically exposed. Total goodwill is $1,883.2M, or 46% of stockholders’ equity. Set that against the operating facts: precision-ag revenue is flat at ~$860M two years after the JV closed, North America is the loss-making segment, and management already took a $354.1M goodwill impairment in FY2024. INTERPRETATION: rates need not rise for this to fire — a downward revision to the forecast revenue growth embedded in the model does the same work, and a third consecutive year of flat precision-ag revenue would force exactly that revision. A write-down would be non-cash and largely anticipated by anyone reading the audit matter, but it would damage the capital-allocation narrative and remove ~46%-of-equity’s worth of cushion from the book value.

3. North America’s loss proves structural rather than cyclical (Medium / High). The bull case implicitly assumes the $(112.0)M segment loss is a volume phenomenon that reverses. The bear reading is that a subscale distributor competing against a ~50–60%-share incumbent with 5x the R&D budget cannot earn an acceptable return in North America at any point in the cycle, and that FY2024’s $84.0M profit was the peak-cycle exception rather than the norm. Supporting the bear: NA dealer inventory remains at just below 7 months against a 6-month target — the worst of the three regions, where Europe and Latin America are already near normal at ~3.5 months. Supporting the bull: management reports gaining share “including high-horsepower offerings in North America,” and the region absorbed the full $22M tariff refund. INTERPRETATION: unresolved, and it is worth roughly $200M of operating income — about 20% of consolidated FY2025 operating profit. This is why section 14 makes NA’s return to profitability the primary bull falsification test.

4. Single-region dependency on Europe (Low-Medium / High). EME contributes 66.8% of sales and $1,006.8M of segment income against consolidated operating income of $595.7M. There is no diversification here in any meaningful sense: if Fendt’s European pricing power weakens, or CAP support is reduced, or European farm income falls the way North America’s has, there is nothing else in the portfolio to absorb it. The likelihood is low — Fendt raised price through the worst of the global downturn, which is strong evidence of durability — but the impact would be existential to the thesis rather than merely damaging. The 2022 invasion shock is a reminder that this concentration carries geopolitical as well as economic beta.

5. Working capital and free cash flow (High, already present / Medium). This risk is not prospective; it is happening. H1-2026 operating cash flow was $(245.0)M against +$153.5M in H1-2025 — a $398.5M adverse swing — in a half-year when sales rose 5.7%. FY2025’s $740.2M of free cash flow was destocking, not earning power, and management’s own conversion guidance of 75–100% of adjusted net income implies roughly $300–400M for FY2026. INTERPRETATION: the risk is not solvency — net debt of ~$2.07B at ~1.9x EBITDA with $960M of undrawn revolver is comfortable — but that a buyer attracted by a headline ~10% free-cash-flow yield is buying a ~4–5% one. It also constrains the buyback: management cannot repurchase 5–7% of the company annually out of $300–400M of free cash flow indefinitely without adding leverage.

Verdict

The risk profile is cyclical and concentrated rather than existential. Solvency is not in question, the dividend is covered many times over, liquidity is ample, and no covenant or refinancing wall is visible before 2027. What can go wrong is more specific: the cycle refuses to turn on the 2027 schedule everyone now assumes; the loss-making North American segment turns out to be a structural rather than a volume problem; and the PTx goodwill — the company’s own disclosure puts its North American headroom at ~3% on a one-point rate move — gets written down, confirming that the defining acquisition of this management team did not earn its cost of capital.

Notably, the risks and the opportunities are the same three items viewed from opposite sides. That symmetry is what makes AGCO a genuine cyclical judgment rather than a quality-versus-price question: the investor is not being asked whether the business is good, but whether three specific, observable and short-dated things resolve favourably. The matrix argues for underwriting this on mid-cycle economics with an explicit haircut for the possibility that North America never fully recovers — which is precisely how section 10’s base case is constructed.

10. Valuation Discussion — Embedded Expectations

Embedded-expectations framing — no price target, no BUY/SELL. Third-party data reconciled to filings; the filing wins on disagreement. Market data as of the 2026-08-21 close.

10.1 Enterprise value

FACT: AGCO closed at $106.81 on 2026-08-21. Shares outstanding 70,002,207 at 2026-06-30 (Q2-2026 10-Q balance sheet). Gross debt $2,639.5M and cash $573.4M at the same date, giving net debt of $2,066.1M.

Noncontrolling interests (Trimble’s 15% of PTx Trimble) ran $7.5M in FY2025 — immaterial to EV. Critically, AGCO Finance is a Rabobank joint venture carried under the equity method, so this net debt figure is genuine industrial leverage requiring none of the captive-finance disaggregation that renders CNH’s consolidated $23.4B analytically useless. The EV above is directly comparable across the peer set on a like-for-like basis; CNH’s is not without adjustment.

10.2 Multiples on guided (trough) earnings

FY2026 guidance (FACT — Q2-2026 earnings release, revised downward 2026-07-30): net sales $10.1–10.2B, adjusted operating margin ~7.5%, adjusted EPS $5.50–5.75, capex $300–325M, FCF conversion 75–100% of adjusted net income.

INTERPRETATION: on guided earnings, AGCO is not cheap. Nineteen times trough EPS and 8.7x trough EBITDA is a full price for a cyclical, and a 4–5% free-cash-flow yield is unremarkable. Anyone buying on the vendor’s 10.2x P/E and 15.8th-percentile screen is buying a number that does not exist. The entire investment case must therefore rest on normalization — which is legitimate for a cyclical at 85% of mid-cycle demand, but it must be argued explicitly rather than smuggled in through a broken screen.

10.3 Own-history percentiles — and why only one of the three is usable

AZI valuation_index, dated 2026-08-21 (~20-year history, n = 3 components). The price input ties exactly to the CSV close, so the data is current; the earnings input does not tie to the filings.

The 53-percentile-point spread between P/E and P/S is itself the diagnosis. When earnings-based and sales-based own-history measures disagree this violently, the earnings denominator is broken — and section 6 shows exactly how: FY2025’s headline EPS was over half TAFE gain and legal-entity tax benefit. Read P/S at the 68.7th percentile as the honest signal: on sales, AGCO is richer than roughly two-thirds of its own twenty-year history.

That is a genuine caution, and it deserves a genuine counter-argument. AGCO’s historical P/S range was set by a structurally lower-margin business: FY2019 operating margin was 3.9%; the guided FY2026 adjusted margin is 7.5% at 85% of mid-cycle demand, after divesting Grain & Protein and adding precision agriculture. A higher sales multiple is arguably warranted by a better margin structure. But that argument requires the improved margin to be real and durable — which is the same bet as the rest of the thesis. The percentile does not settle it; it simply removes “obviously cheap on its own history” from the bull case.

10.4 The normalization — what mid-cycle actually looks like

This is the load-bearing calculation. Management’s own statement is the anchor: “industry demand remains at a low level, operating at around 85% of mid-cycle demand” (FACT — Q2-2026 call).

At $106.81 that is ~8.4x mid-cycle EPS (range 8.2–9.7x) and ~6.0x mid-cycle EBITDA (mid-cycle EBITDA ~$1,530M against EV of $9,544M).

Two sanity checks. First, FY2023 — the actual cycle peak, at revenue of $14,412.4M — produced GAAP EPS of $15.63 and ROIC of 22.9%; a mid-cycle figure of ~$12.25 on revenue of ~$11.9B sits sensibly below it. Second, the implied mid-cycle operating margin of ~10.1% is below the FY2022 achieved margin of 10.0%–FY2023’s 11.8%, so the estimate does not require heroic margin assumptions — it requires volume, and modest incremental margins on that volume.

The cross-read that frames this: prior published analysis of Deere establishes that Deere trades at 13.6–15.3x its own management-stated mid-cycle EPS framework of $40–45. AGCO at ~8.4x mid-cycle is therefore at roughly a 40–45% discount to Deere on comparable normalized earnings. Some discount is unambiguously deserved — Deere is the #1 with a structurally higher through-cycle margin, 5x the R&D, twice the precision-ag share, a wholly-owned captive finance arm and a construction segment for diversification. Whether the deserved discount is 40–45% or something closer to 25–30% is the entire debate.

10.5 Scenario grid

All assumptions explicit; per-share outputs are scenario translations, not price targets.

INTERPRETATION: the price sits ~16% below the base case and ~33% above the bear, which is a reasonable but not extraordinary skew — roughly 1.3:1 in favour of the upside on these assumptions. Note what the bear case actually is: it is not a depression, it is “the cycle takes longer and North America never earns a normal margin,” and it still produces ~$80. The bull requires no heroics either — an 11.5% mid-cycle margin is below the FY2023 achieved 11.8%.

The load-bearing assumptions are transparent. For the base case: that mid-cycle revenue really is ~$11.9B (management’s own 85% figure) and that incremental margins of 25% hold. For the bear: that North America’s $(112.0)M loss is structural, which costs roughly 200bp of consolidated mid-cycle margin. For the bull: that North America returns to a mid-single-digit margin and precision agriculture finally inflects above its ~$860M plateau.

10.6 Two things the market is not paying for

The IEEPA tariff refunds. FY2026 guidance embeds the $22M already recognized but explicitly “does not assume any potential benefits related to future IEEPA refunds”, while additional claims sit filed and pending CBP review. Gross 2026 tariff cost is ~$115M; if a material further share were recovered, it would flow straight to earnings against a ~$761M adjusted operating income base. This is genuinely unpriced optionality, appropriately excluded by management, and legally uncertain (the administration has appealed to the Federal Circuit and imposed replacement tariffs under other statutes). It should be treated as an option, not as a forecast — but it is an option with no premium in the price.

The buyback’s per-share compounding. The share count fell from 72,629,310 to 70,002,207 in six months (−3.6%). If management sustains even half that pace against a $1.0B authorization, the mid-cycle EPS denominator shrinks meaningfully by the time the cycle turns — which is when it matters most. At ~8.4x mid-cycle earnings, repurchase is the highest-return capital use available, and the arithmetic compounds in the shareholder’s favour precisely because the price is depressed.

10.7 Verdict — what the market is underwriting

Correctly: that the agricultural cycle has not turned and that USDA’s farm-income data offers no evidence it is about to; that AGCO’s FY2026 guidance was cut and could be cut again; that the North American segment is losing money and that a subscale challenger against a 50–60%-share incumbent deserves a discount; that PTx has not yet earned its cost of capital and its goodwill cushion is thin; and that FY2025’s free cash flow was destocking rather than earning power. A 19x multiple on trough EPS and a 4–5% free-cash-flow yield are not the marks of an unloved stock.

Incorrectly, or at least incompletely: the market appears to be extrapolating the trough rather than discounting the cycle. At ~8.4x mid-cycle EPS and ~6.0x mid-cycle EBITDA, the price embeds little recovery in a business whose own management puts industry demand at 85% of mid-cycle, whose European segment raised operating income on positive net pricing during the worst of the downturn, whose gross margin moved only 130 basis points through a 30% revenue collapse, and which is repurchasing 5–7% of its shares annually. The 40–45% discount to Deere’s normalized multiple prices AGCO as though it were CNH — 0.76x sales against CNH’s 0.73x — when AGCO has a genuine 14.9%-margin European franchise, roughly double CNH’s precision-ag share, a peak ROIC of 22.9% and a balance sheet that needs no disaggregation.

None of that says the price must rise. It says the market is paying for a trough that management, the peer group and the dealer-inventory data all suggest is closer to its end than its beginning — while charging nothing for the tariff-refund option and giving no credit to a shrinking share count. The single sentence: expensive on what AGCO earns today, meaningfully cheap on what it earns at 100% of mid-cycle, and the gap between those two statements is the investment.

11. Variant Perception

Consensus, bull, bear, and the assumptions that matter. No new facts are introduced; every figure traces to a preceding section. No BUY/SELL, no price target; per-share figures are section 10’s labeled scenario outputs, not targets.

11.1 Consensus belief — what the market is underwriting

FACT: at $106.81 AGCO carries an enterprise value of ~$9,544M — 0.94x FY2026E sales, ~8.7x FY2026E EBITDA and 18.6–19.4x cut FY2026 adjusted-EPS guidance of $5.50–5.75. The stock is 23.5% below its five-year high of $139.71 (2026-02-13), set its 52-week low of $98.45 four sessions ago on 2026-08-18, and has produced negative returns over one, three and five years with negative Sharpe ratios across all three. It trades at 0.76x sales against CNH’s 0.73x and roughly one-sixth of Deere’s multiple.

INTERPRETATION: the consensus being priced is straightforward and defensible: this is a structurally disadvantaged #3 in a cyclical industry whose down-cycle has not ended, whose guidance was just cut, whose North American business is losing money, and whose big acquisition has not worked. The market is extrapolating the trough. It is not pricing a franchise, it is not pricing the European segment’s 14.9% margin as durable, and it is explicitly not pricing any tariff-refund recovery.

Importantly, consensus here is not crowded in either direction. The factor model shows Momentum at −0.04 and Quality at +0.04 — both effectively zero — with only a mild Value tilt of +0.21 despite a 23.5% drawdown. This is a stock nobody is excited about and nobody has capitulated on: no momentum crowd to unwind, but no deep-value crowd underwriting it either.

11.2 The strongest bull case — steelmanned

• The normalization arithmetic is management’s own, not the analyst’s. Industry demand is running at ~85% of mid-cycle. That implies ~$11.9B of mid-cycle revenue against ~$10.15B guided, ~$1.75B of latent revenue, and — at a conservative 25% incremental margin versus the 30%+ achieved on the way up — roughly $440M of latent operating income. Mid-cycle EPS of ~$12.25 puts the stock at 8.4x, against the 13.6–15.3x Deere commands on its own mid-cycle framework.
• The pricing evidence says this is not a commodity business. Gross margin moved from 26.2% to 24.9% to 25.5% across a 30% revenue collapse — 130 basis points. The European segment’s operating income rose $45.5M in FY2025 “as a result of positive net pricing.” Companies without pricing power discount to hold volume; AGCO cut production instead and held margin. That is a franchise behaving like one.
• The cyclical turn has begun to show in the data, not just the narrative. Dealer inventories fell in all three major regions (Europe and Latin America to ~3.5 months, North America to just below 7 against a 6-month target); H1-2026 sales rose 5.7% year over year; and on 2026-08-20 Deere — the best-informed operator in the industry — called fiscal 2026 the ag-cycle bottom and cited improving 2027 orders.
• Two levers are entirely within management’s control. North America returning to even a mid-single-digit margin adds ~$200M of operating income — about 20% of consolidated FY2025 operating profit — with management already reporting share gains in high-horsepower there. And the share count fell 3.6% in six months against a $1.0B authorization, compounding the per-share effect of any recovery.
• Two things are free. FY2026 guidance excludes any further IEEPA tariff recovery while claims sit pending, against gross 2026 tariff costs of ~$115M. And the incentive plan pays on RONA in both the annual and long-term programs with a relative TSR modifier — management is paid to earn returns on capital, not to grow.
• The insider signal is the sector’s best. In 294 Form 4s containing only six open-market purchases, director Bob De Lange accounts for five — ~$852K across two years, including 1,000 shares at $100.71 on 2026-08-14, within two sessions of the 52-week low.

11.3 The strongest bear case — steelmanned

• The demand trigger has not fired, and its composition is deteriorating. USDA forecasts 2026 US net farm income of $153.4B, down 2.6% in real terms, with crop margins the poorest since 2016–2020 and corn at ~$4.20–4.50. The support is increasingly artificial: direct government payments up 45.2%, farm debt up 5.2%, working capital down 9.2%. Transfer payments sustain farm consumption; they do not fund $600,000 tractor replacements.
• Management cut guidance five weeks ago, citing “weaker-than-expected industry conditions… elevated operating costs, uneven crop economics… delayed equipment investments and limited visibility into demand recovery,” and guided production hours below 2025. The 2027-recovery consensus has already slipped once.
• The reported earnings have been unreliable, and the market was fooled by them. FY2025’s headline EPS of $9.75 was over half the $251.9M TAFE gain and the $179.8M legal-entity tax benefit. The five-year high of $139.71 was set on the very day the 10-K disclosing that number was filed. Vendor screens still publish trailing EPS of $10.4365 against an actual $7.21. A stock whose valuation case requires this much correction before it can even be assessed carries a real analytical-risk premium.
• On the one own-history measure that works, the stock is not cheap. P/S at 0.76x sits at the 68.7th percentile of a ~20-year history — richer than two-thirds of its own past. And on guided earnings it trades at 18.6–19.4x, with a free-cash-flow yield of 4.0–5.3%, not the 9.9% the destock-flattered FY2025 figure advertises.
• Cash flow has already turned negative. H1-2026 operating cash flow was $(245.0)M against +$153.5M a year earlier — a $398.5M adverse swing in a half-year when sales rose.
• The defining acquisition is one assumption from a write-down. The 10-K’s own critical audit matter puts PTx North America’s goodwill headroom at ~3% on a one-point discount-rate increase, with total goodwill at $1,883.2M — 46% of stockholders’ equity — precision-ag revenue flat at ~$860M two years post-close, and a $354.1M impairment already taken in FY2024.
• The structural position is genuinely inferior. ~$0.5B of R&D against Deere’s ~$2.3B; ~7–9% precision-ag share against ~15–18%; no construction segment; no captive finance spread; two-thirds of profit in one region; and a North American business that loses money against a 50–60%-share incumbent.

11.4 The assumptions that matter most, and what falsifies each side

1. The ag cycle turns on the 2027 schedule. Falsifies the bull: a second FY2026 or an FY2027 guidance cut; USDA’s next farm-income forecast down again in real terms; dealer inventories stopping their decline. Falsifies the bear: Q3/Q4-2026 order books and 2027 pre-season commitments firming across AGCO, Deere and CNH.
2. North America’s loss is cyclical, not structural. Falsifies the bull: the segment still loss-making at materially higher volumes, or NA dealer inventory stuck above the 6-month target through 2027. Falsifies the bear: NA returning to positive operating income — the single cleanest test in the report.
3. Fendt’s European pricing power is durable. Falsifies the bull: EME segment margin compressing below ~12% while volumes are flat or rising, i.e. price rather than volume doing the damage. Falsifies the bear: EME holding ~15% through the trough and expanding on the recovery — currently the observed state.
4. PTx earns its cost of capital. Falsifies the bull: a goodwill impairment at PTx North America, or a third consecutive year of precision-ag revenue flat at ~$860M. Falsifies the bear: precision-ag revenue breaking above the plateau with disclosed margin accretion.
5. The normalized-EPS estimate is right. Falsifies the bull: mid-cycle margins proving structurally capped near 8% because North America is permanently subscale — which produces ~$8.00 of EPS and the ~$80 bear output. Falsifies the bear: incremental margins on the recovery matching the 30%+ delivered in FY2021–23 rather than the 25% assumed.

11.5 What the tape is pricing — the factor-positioning read (input, not a price call)

FACT: AGCO is, in factor space, a market-sensitive industrial cyclical with dividend-yield and small-size tilts and a mild value lean — Market +0.88, Sector: Industrials +0.67, DividendYield +0.59, SmallSize +0.35, Value +0.21, with Momentum at −0.04 and Quality at +0.04, both effectively zero. Specific volatility is 27.2% and ~53% of variance is idiosyncratic. Returns are negative over one year (−3.5%), three years (−2.5%/yr) and five years (−2.4%/yr), with negative Sharpe at each; only the ten-year (+10.9%/yr) and lifetime (+9.1%/yr) windows are positive, and lifetime maximum drawdown is −78.7%. Alpha is −0.221. The factor-similar peer set is CNH 0.964, TEX 0.955, DE 0.916, VEGI 0.915 — statistically validating the fundamental comp set from the other direction.

INTERPRETATION (regime-caveated): three things follow. First, there is no crowded trade to unwind here — a zero momentum loading after a 23.5% drawdown means the marginal seller has largely already sold, which is a materially different risk profile from a high-momentum name where the downside is a positioning cascade. Second, the style regime has turned in AGCO’s favour for the first time this cycle: Value (63d +5.3%, z +1.03), SmallSize (63d +8.1%, z +1.24) and DividendYield (63d +5.7%, z +1.00) are all being paid, and those are precisely AGCO’s three positive loadings; the offsetting headwind is Sector: Industrials at 21d −2.9% (z −0.98). Third, and most important for interpreting the last week, the +8.5% bounce off the 52-week low rests on borrowed evidence — Deere’s beat came from Construction & Forestry, a segment AGCO does not own, and CAT fell 3.3% over the same window. Where consensus may be offsides in positioning terms is in treating a stock with negative five-year returns and zero momentum loading as a value trap, when the factor regime and the insider record both point the other way.

11.6 Where consensus is right and where it is wrong

Right: that the cycle has not turned and USDA’s data offers no evidence it is about to; that guidance was cut and could be cut again; that a subscale, loss-making North American segment against a 50–60%-share incumbent deserves a discount to Deere; that PTx has not earned its cost of capital and its goodwill cushion is disclosed as thin; that FY2025’s free cash flow was destocking; and that a business with ~$0.5B of R&D cannot out-innovate one spending ~$2.3B.

Wrong, or at least incomplete:

• On what AGCO is. The market prices AGCO at 0.76x sales against CNH’s 0.73x — essentially identically. But AGCO has a 14.9%-margin European franchise that raised prices through the worst downturn in a decade, roughly double CNH’s precision-ag share, a differentiated open-architecture technology strategy against CNH’s bought-and-impaired one, a peak ROIC of 22.9% against CNH’s materially lower figures, and a clean balance sheet requiring none of the captive-finance disaggregation CNH’s demands. Pricing these two businesses the same is the clearest mispricing in the report.
• On trough versus cycle. At ~8.4x mid-cycle EPS and ~6.0x mid-cycle EBITDA the price extrapolates the trough rather than discounting the cycle — in a business whose own management puts industry demand at 85% of mid-cycle and whose dealer inventories are falling in every region.
• On what is being given away. The tariff-refund option carries no premium in the price and is explicitly excluded from guidance; a 3.6%-per-half-year share-count reduction gets no credit; and an incentive plan paying on return on net assets with a relative TSR modifier is materially better than the sector norm and goes unremarked.

The net variant perception: the market has correctly identified that AGCO’s reported earnings were a mirage and has spent six months un-pricing them — but in doing so it has arrived at a valuation that treats a genuine European franchise as though it were CNH, and treats a documented 85%-of-mid-cycle trough as though it were the run-rate. The falsification schedule is short-dated: North American segment profitability, EME margin durability, the next USDA farm-income forecast, and any PTx impairment will settle most of this within four quarters.

12. Fact vs. Interpretation Table

FACT = stated in a filing, release, transcript or primary dataset. INTERPRETATION = our reading of facts. ASSUMPTION = an input we chose, which a reader may replace. OPEN QUESTION = not answerable on available evidence.

13. Open Questions

Questions this analysis could not settle from public evidence, each with what would answer it and roughly when.

1. Is North America’s operating loss cyclical or structural? This is the most valuable unanswered question in the report and it is worth roughly $200M of operating income — about 20% of consolidated FY2025 operating profit. FY2024 produced +$84.0M and FY2025 produced $(112.0)M on sales down 27.5%, which is consistent with pure negative operating leverage; but it is equally consistent with a subscale distributor that only earns in boom conditions against a 50–60%-share incumbent. AGCO does not disclose North American segment margins at comparable volumes across enough cycles to separate the two. Answered by: the segment’s result as NA volumes recover — specifically whether it reaches positive operating income before consolidated revenue regains ~$11B. Expected: FY2027.

2. What is the base-case goodwill headroom at PTx North America? The 10-K discloses the sensitivity (~3% at +1.0pt discount rate, ~32% at −1.0pt) but not the base-case cushion itself. Linear interpolation suggests mid-teens, but goodwill sensitivities are not linear and the disclosure is deliberately partial. The distinction matters: mid-teens headroom with flat revenue is a warning; low-single-digit headroom is a near-certainty of impairment. Answered by: the FY2026 10-K’s critical audit matter, or the impairment itself. Expected: February 2027.

3. What did PTx Trimble actually cost, and what is it earning? The FY2025 10-K describes the joint venture and assigns its goodwill across four reporting units but does not isolate the purchase consideration, the standalone revenue, or the segment profitability of the precision-ag business in the extracted disclosure. Precision-ag revenue of ~$860M is a management figure given on a call, not an audited segment. Without a disclosed margin it is impossible to compute a return on the acquisition — the single most important capital-allocation judgment on this management team. Answered by: AGCO would have to begin reporting precision agriculture as a segment, which it has shown no intention of doing. This may remain permanently unanswerable from outside.

4. What proportion of revenue is replacement parts and service? AGCO does not disclose it. This is a genuine disclosure gap relative to Deere, and it matters because the parts annuity is the principal stabiliser of a cyclical revenue base — it determines how much of the trough revenue is genuinely defensive. Every estimate of AGCO’s “true” cyclicality is weaker for its absence. Answered by: not answerable from public filings.

5. How much of the remaining IEEPA tariff exposure will actually be refunded, and when? AGCO has recovered $22M and filed further claims, but states the process “remains subject to CBP review,” that the administration has appealed to the Federal Circuit “contesting both the scope of the refund obligation and the reliquidation of finally liquidated entries,” and that “it remains uncertain when, or to what extent, such refunds will ultimately be collected.” Against gross 2026 tariff costs of ~$115M, the unrecovered balance is material. Answered by: Federal Circuit ruling and CBP processing. Expected: uncertain — plausibly 2027.

6. Why is the sliding scale in the incentive plan calibrated as it is? The AIP and LTI both adjust adjusted-operating-margin and RONA targets “on a sliding scale based upon changes in industry conditions.” The concept is defensible for a cyclical, but the proxy does not disclose the scale itself, so it is impossible to judge whether FY2025’s payouts were earned against a demanding normalized target or a generous one. Answered by: the proxy would need to publish the adjustment mechanism; it does not.

7. Is Fendt’s 14.9% EME margin sustainable at higher volumes, or is it partly a mix artifact? In a downturn, AGCO shipped proportionally more high-margin Fendt and fewer low-margin volume units — so some of EME’s margin resilience may be mix rather than pricing. Management attributes it to “positive net pricing,” but the 10-K does not decompose price versus mix versus volume. If it is mix, the margin will fall as volume returns, which would materially damage the normalization arithmetic in section 10. Answered by: EME margin behaviour through the first two quarters of genuine volume recovery.

8. What is AGCO’s actual global and regional market share? Share figures used throughout (Deere ~50–60% of NA large tractors and combines, AGCO ~7–9% precision-ag) are third-party estimates compiled in prior published analysis, not census data. Management asserts share gains “in key markets” while conceding losses in France. No independent measurement was located. Answered by: a purchased industry data set; not resolvable from public sources.

9. Will the buyback be sustained if free cash flow stays at ~$300–400M? Repurchasing 5–7% of the company annually is not fundable from $300–400M of free cash flow alongside a $1.16 dividend without adding leverage. Management has not addressed the trade-off publicly. Answered by: the H2-2026 and FY2027 repurchase pace against reported free cash flow.

14. What Must Be True

Each side reduced to conditions that can be checked, with an explicit falsification test and a date. This section exists so a future reader can score the memo rather than re-argue it.

For the bull case to work

B1. The agricultural cycle must turn, beginning in 2027. Every element of the base case depends on the ~$1.75B of latent revenue implied by management’s “~85% of mid-cycle” statement. Without the cycle, mid-cycle EPS of ~$12.25 is fiction and the stock is simply 19x trough earnings.

Falsification test: a second consecutive guidance reduction (FY2026 having already been cut on 2026-07-30), or USDA’s next US net farm income forecast down again in real terms from $153.4B, or dealer inventories ceasing to fall. Check at: Q3-2026 (November 2026), Q4-2026/FY2027 guidance (February 2027), and the USDA February forecast.

B2. North America must return to operating profit. The $(112.0)M segment loss must prove to be negative operating leverage rather than structural subscale. This is worth roughly $200M of operating income and about 200bp of consolidated mid-cycle margin — the difference between the base case (~$127) and the bear (~$80).

Falsification test: the segment still loss-making once North American revenue recovers toward ~$2.0B, or NA dealer inventory stuck above the 6-month target through 2027 while Europe and Latin America sit at ~3.5 months. Check at: quarterly segment disclosure through FY2027.

B3. Fendt’s European pricing power must be durable and not a downturn mix artifact. EME’s 14.9% margin and $1,006.8M of income carry the entire company. The bull case needs that margin to hold — ideally expand — as volume returns.

Falsification test: EME segment margin compressing below ~12% while EME volumes are flat or rising, which would show the margin was mix rather than price. Check at: the first two quarters of genuine European volume recovery. Currently tracking bull: EME income rose $45.5M in FY2025 on “positive net pricing.”

B4. Incremental margins on the recovery must be at least ~25%. The normalization assumes $440M of incremental operating income on $1.75B of incremental revenue. AGCO achieved above 30% on the way up in FY2021–23.

Falsification test: two quarters of recovering volume delivering incremental operating margins below ~20%, which would indicate the cost base has permanently inflated — engineering at ~5% of sales is the obvious candidate. Check at: FY2027 quarterly results.

For the bear case to work

R1. The recovery must slip past 2027. The bear does not require a depression — only delay. A third trough year compounds the damage because it would coincide with negative free cash flow, a stretched buyback and a goodwill test under pressure.

Falsification test: AGCO, Deere and CNH all confirming firming 2027 order books and pre-season commitments. Currently tracking bear, weakly: AGCO cut guidance five weeks ago; but Deere called fiscal 2026 the ag-cycle bottom on 2026-08-20 with improving 2027 orders, which cuts the other way.

R2. North America must prove structurally unprofitable. The bear’s ~$80 output requires mid-cycle margins capped near 8% because the second-largest region never earns.

Falsification test: North America posting positive operating income in any quarter of a recovering market. Check at: FY2027.

R3. PTx must fail to earn its cost of capital — and preferably be written down. The bear needs the defining acquisition to be confirmed as value-destructive.

Falsification test: precision-ag revenue breaking decisively above the ~$860M plateau with disclosed margin accretion. Currently tracking bear: revenue flat two years post-close, and the company’s own disclosure puts PTx North America headroom at ~3% on a +1.0pt discount-rate move.

R4. The market must continue to price AGCO as CNH’s equal. The bear implicitly requires the 0.76x-versus-0.73x sales parity to persist — i.e. that the European franchise, the precision-ag share advantage and the 22.9% peak ROIC are never re-recognised.

Falsification test: any multiple expansion toward Deere’s normalized 13.6–15.3x, or a strategic event (activist, break-up interest, or a bid for the European franchise) that forces a sum-of-the-parts revaluation.

Scoring summary

The honest read: the bull case is winning on pricing power and losing on the acquisition, while the two largest questions — the cycle’s timing and North America’s profitability — are genuinely unresolved and are the same question asked twice. Four of the eight conditions turn on North American segment profitability or the cycle date, which is why this article treats the quarterly segment disclosure through FY2027 as the highest-information data stream available. Notably, the bear case rests substantially on things that have already happened (the guidance cut, the flat precision-ag revenue, the thin goodwill headroom) while the bull case rests on things that must still happen — an asymmetry that argues for patience and for sizing into weakness rather than reaching on strength.

15. Source Appendix

See Appendix B — Source Appendix, which accompanies this article and lists every primary and secondary source relied on, with URLs and access dates.

Appendix A — Diligence Questionnaire

AGCO Corporation (NYSE: AGCO | CIK 0000880266) — Report date 2026-08-22

Every question is answered from the underlying analysis. Labels: [FACT] = sourced to a filing/transcript/named data source; [INTERPRETATION] = analyst judgment on facts; [ASSUMPTION] = estimate without primary-source confirmation. Sources accessed 2026-08-22; market data as of the 2026-08-21 close ($106.81). Full citations in Appendix B. No recommendation, no price target.

1. General

What thoughtful questions have other investors asked about this company? The recurring ones, and our answers: (i) Is AGCO structurally disadvantaged as the #3? Yes globally, but not in Europe, where Fendt earns a 14.9% segment margin [FACT]. (ii) Was PTx Trimble worth it? Unproven and currently under strain — precision-ag revenue flat at ~$860M and PTx North America goodwill headroom of ~3% on a +1.0pt discount-rate move [FACT]. (iii) Why is the P/E so low? Because it is computed on a distorted denominator; trailing EPS is $7.21, not the $10.4365 vendors publish [FACT]. (iv) Is the dividend safe? Yes — $1.16 annualized against ~$300–400M of guided free cash flow is roughly 4x covered [INTERPRETATION]. (v) Why does North America lose money? Subscale distribution against a ~50–60%-share incumbent, plus factory under-absorption from deliberate production cuts [INTERPRETATION].

2. Cyclicality & Earnings Nature

Are earnings at a cyclical high or low? A pronounced low. Revenue fell 30.0% from $14,412.4M (FY2023) to $10,082.0M (FY2025); GAAP operating margin fell from 11.8% to −1.0% to 5.9%; management states industry demand is at ~85% of mid-cycle [FACT].

Driven by the external environment or internal actions? Overwhelmingly external. The driver is net farm income — USDA forecasts 2026 US net farm income of $153.4B, down 2.6% in real terms, with crop margins the poorest since 2016–2020 [FACT]. Internal actions have been defensive and, on the evidence, well executed: production cut rather than price cut, gross margin defended within 130bp [INTERPRETATION].

How stable are revenues? Poor stability at the whole-goods level (~68% of sales in tractors and combines — big-ticket, financed, deferrable). Partially stabilised by a parts and service annuity whose size AGCO does not disclose — a genuine gap versus Deere [FACT/OPEN QUESTION]. Note that reported revenue is wholesale shipments, not retail demand, so it overshoots the underlying cycle in both directions [FACT — 10-Q].

Outlook for products/services? Tractors have risen to 66% of sales from 61%, combines fallen to 2% from 4% — the mix has concentrated on the core as the cycle deepened [FACT]. Precision agriculture (~$860M, guided flat to modestly up) is the strategic growth vector [FACT].

How big will this market be — growing, shrinking, domestic or international? International: EME 66.8% of sales, North America 16.5%, South America 11.1%, APA ~5.6% [FACT]. The market is mature and GDP-plus at best over a full cycle — AGCO’s revenue has compounded at under 2% a year since FY2019. The cyclical amplitude, not the secular trend, is what matters [INTERPRETATION].

3. Business Quality & Competitive Moat

Is the industry getting more or less competitive? Broadly stable and, importantly, rational — the strongest evidence being that gross margin moved only 130bp through a 30% volume collapse, i.e. nobody bought volume with price [FACT/INTERPRETATION]. Competition is intensifying in precision agriculture, where Deere leads on scale and AGCO counters on architecture.

How profitable is the business (ROIC, ROE)? Peak-cycle returns are franchise-grade and trough returns are not: ROIC 22.9% and ROE 27.4% at the FY2023 peak; ROE −10.1% in FY2024; FY2025’s reported 18.1% ROE is flattered by one-time items and is ~11–12% normalized [FACT/INTERPRETATION]. Through-cycle, the returns are ordinary — a good business in a cyclical industry, not a compounder.

How profitable is the industry — how many competitors, what barriers to entry? A profitable oligopoly — Deere, CNH, AGCO, Kubota, CLAAS — with the profit pool skewed heavily to Deere. Barriers: dealer density, brand, emissions-compliance R&D scale, precision integration, capital intensity [INTERPRETATION, built from filings and prior peer work].

Can the business be easily understood? Yes at the operating level. No at the reported-earnings level — FY2025’s $9.75 headline EPS was over half TAFE gain and legal-entity tax benefit, and vendor screens still propagate the error [FACT].

Can it be undermined by foreign low-cost labor? Only at the low end. Compact and utility tractors face low-cost competition (Kubota, Mahindra); the high-horsepower and precision segments where AGCO’s profit sits are protected by dealer networks, service requirements and emissions engineering [INTERPRETATION]. Note AGCO itself buys fully manufactured tractors from third parties (SDF, Carraro, Iseki) — it is a participant in outsourcing, not a victim of it [FACT].

Do brands matter? Decisively, and this is the single best moat evidence in the report: Fendt commands price, with EME segment income rising $45.5M in FY2025 on “positive net pricing” during the worst downturn in a decade [FACT].

What is the nature of competition? Dealer-led. A farmer’s brand choice is largely a dealer choice — parts availability and service response during a harvest window outrank list price, because downtime is catastrophically expensive [INTERPRETATION].

Customers’ switching costs? Real but bounded — measured in equipment replacement cycles (5–10 years) rather than decades. Reinforced by the parts annuity and, increasingly, by precision-ag data. Note AGCO’s own precision strategy is deliberately anti-lock-in: retrofit-first and mixed-fleet, designed to work on competitors’ machines [FACT].

4. Financial Condition & Balance Sheet

Assets not fully recognized on the balance sheet? The Fendt brand and the global dealer network are the principal ones — internally generated brand value is not capitalised, and Fendt’s demonstrated pricing power suggests it is worth considerably more than its carrying value [INTERPRETATION]. The AGCO Finance JV is carried at equity-method value rather than economic value.

Off-balance-sheet liabilities? No material ones identified. The Rabobank finance JV is a genuine joint venture, not a consolidation-avoidance structure, and AGCO does not carry the retail credit book [FACT].

How conservative is the accounting? Mixed. Positives: the goodwill sensitivity is disclosed candidly (the ~3% headroom figure is voluntary detail most issuers omit), and FY2024’s impairment was taken promptly. Negative: goodwill of $1,883.2M is 46% of stockholders’ equity [FACT], and FY2025’s reported earnings were dominated by non-operating items presented without prominent normalization [INTERPRETATION].

How CapEx-hungry is the business? Moderately. FY2025 capex $247.9M against depreciation $256.5M — roughly maintenance level; FY2026 guided down to $300–325M. The heavier commitment is engineering expense of $487.7M (4.8% of sales), guided to ~5% and rising 17.1% year over year in H1-2026 — R&D, not capex, is where the capital intensity sits [FACT].

5. Capital Allocation & Management

How much FCF does the business generate, how does management use it, what is the philosophy? FY2025 free cash flow was $740.2M — but destock-flattered: H1-2026 operating cash flow was $(245.0)M against +$153.5M a year earlier, and management guides FY2026 conversion at 75–100% of adjusted net income, implying ~$300–400M [FACT]. Uses, in priority order: engineering (~$500M), buybacks, dividend. That order is correct at this valuation [INTERPRETATION].

Significant acquisitions recently? PTx Trimble (85% JV, April 2024) — the defining decision, and unproven: revenue flat at ~$860M and goodwill headroom disclosed as thin. Also the 2024 divestiture of the majority of Grain & Protein, which improved mix [FACT].

Buying back shares? Yes, materially. $1.0B authorized 2025-07-09; a $250M ASR in November 2025; shares outstanding fell from 72,629,310 to 70,002,207 between 2025-12-31 and 2026-06-30, −3.6% in six months [FACT]. At ~8.4x mid-cycle earnings this is the highest-return use of capital available [INTERPRETATION].

Issuing large amounts of new shares to insiders? No. The Form 4 corpus shows 215 grant events across five years but the share count is falling, so grants are more than offset by repurchase [FACT].

Compensation policy of directors/management? Better than the sector norm. AIP: 40% adjusted operating margin, 40% RONA, 10% employee engagement, 10% customer satisfaction. LTI: ~60% performance shares (50% three-year revenue growth, 50% three-year RONA) with a relative TSR modifier of ±20%, plus ~40% RSUs; double-trigger vesting and a clawback policy [FACT]. Return on capital appears in both plans and the TSR gate is relative, not absolute — a materially better design than plans built on absolute EBITDA. The criticism: the targets are adjusted “on a sliding scale based upon changes in industry conditions,” and the scale itself is not disclosed [FACT/OPEN QUESTION].

Motivations of management? Aligned on returns rather than size, per the plan above. The insider record supports it: of 294 Form 4s containing only six open-market purchases, five are director Bob De Lange’s (~$852K), including 1,000 shares at $100.71 on 2026-08-14 — within two sessions of the 52-week low. Officer selling is immaterial; 97.7% of code-S volume is TAFE exiting under the settlement [FACT].

6. Valuation & Market Data

Is the stock an ADR, MLP, or K-1 issuer? No — AGCO Corporation is a Delaware corporation listed on the NYSE, issuing a standard Form 1099 [FACT].

Dividend policy? $0.29 quarterly, $1.16 annualized, a 1.09% yield, payout ~12% of FY2025 reported earnings. Maintained through the down-cycle, not raised [FACT].

How profitable is the business? See section 3 above. FY2026E adjusted operating margin ~7.5% at 85% of mid-cycle, against 11.8% GAAP at the FY2023 peak [FACT].

Is net income diverging from cash from operations? Yes, and materially — in both directions. FY2025: net income $726.5M against operating cash flow $988.1M, the gap being destocking. H1-2026: net income $132.2M against operating cash flow of $(245.0)M — a $377M divergence the other way as working capital rebuilt [FACT]. This is the single most important accrual-quality observation on the name: neither the FY2025 free-cash-flow yield of 9.9% nor the H1-2026 outflow is a run-rate [INTERPRETATION].

7. Risks & Downside

What factors would cause the stock to decline? In order: (1) a second guidance cut pushing recovery past 2027; (2) North America remaining loss-making into recovering volumes; (3) a PTx goodwill impairment; (4) EME margin compression revealing that 14.9% was mix rather than price; (5) further tariff escalation or an adverse Federal Circuit ruling on refunds [INTERPRETATION].

Risk of a catastrophic loss? Low. Net debt of ~$2,066.1M at ~1.9x FY2026E EBITDA, $960.0M of undrawn revolver, no maturity wall before 2027, a dividend covered ~4x, and a business generating positive operating income even in the worst year of a 30% revenue collapse [FACT/INTERPRETATION]. The realistic severe case is the section 10 bear scenario of ~$80 — roughly 25% below the current price — driven by a delayed cycle plus a structurally unprofitable North America, not by insolvency.

Chance of a total loss? Negligible within the horizon. This is an investment-grade-profile industrial with real assets, a 66.8%-of-sales European franchise earning 14.9% margins, and modest leverage. The lifetime maximum drawdown of −78.7% [FACT — FactorsToday] is a statement about cyclical volatility, not solvency.

8. Recent News & Events

Has the business environment changed recently? Yes, in three ways: the US Supreme Court ruled in February 2026 that IEEPA does not authorize the tariffs, creating a refund claim ($22M recovered, more filed and pending, none assumed in guidance); management cut FY2026 guidance on 2026-07-30; and on 2026-08-20 Deere called fiscal 2026 the ag-cycle bottom with improving 2027 orders, re-rating the complex — though Deere’s beat came from Construction & Forestry, which AGCO does not own [FACT/INTERPRETATION].

Significant acquisitions? None since PTx Trimble (April 2024). The notable transaction since is a disposal: AGCO’s stake in TAFE sold back for $260M, completing 2025-09-30 [FACT].

Change in accounting policies? None identified. No auditor change, no restatement, no covenant issue across the five-year corpus [FACT].

Recent changes — new markets, facilities, management? No CEO or CFO change: Eric Hansotia remains Chairman, President and CEO throughout the window [FACT]. The strategic changes are the 2024 portfolio pivot (PTx in, Grain & Protein out), the TAFE termination and settlement, and continued Fendt full-line expansion into the Americas [FACT].

Appendix A notes on evidence quality

• The largest third-party, non-company estimates repeated here are market shares — Deere ~50–60% of North American large tractors and combines, precision-ag share of ~15–18% (Deere) versus ~7–9% (AGCO). These were compiled during prior published analysis of Deere and CNH; no independent census exists and they should be treated as directional.
• Management statements treated as hypotheses (management commentary is a hypothesis, not evidence), not evidence: “industry demand… at around 85% of mid-cycle,” share gains “in key markets,” and the ~$860M precision-ag revenue figure (given on a call, not an audited segment).
• Vendor data was materially wrong on this name and was overridden. AZI’s trailing EPS of $10.4365 (P/E 10.23x, 15.8th percentile) does not reconcile to the filings; the rebuilt figure is $7.21. Separately, ROIC.ai’s oper_margin field is an adjusted figure — it reports +7.95% for FY2024 against a GAAP operating loss — so all margins herein are computed from EDGAR XBRL.
• All AGCO financial figures reconcile to SEC filings. Where ROIC.ai and the filings disagreed, the filing won and the discrepancy is documented.

Appendix B — Sources

AGCO Corporation (NYSE: AGCO | CIK 0000880266) — Report date 2026-08-22

Sources are deduplicated and organized by type. All sources accessed 2026-08-22; market and factor data as of the 2026-08-21 close.

⚠ Prominent caveat — vendor data was wrong on this name

Two third-party data errors were identified and overridden. Any reader reproducing this analysis from vendor feeds will reach different — and incorrect — conclusions.

1. AZI valuation_index trailing EPS of $10.4365 (implying P/E 10.23x at the 15.8th percentile of own history) does not reconcile to the filings. Rebuilt from reported quarterly diluted EPS, trailing EPS is $7.21 and the true P/E is ~14.8x. The P/E percentile is unusable; P/S at the 68.7th percentile is the honest own-history read.
2. ROIC.ai’s oper_margin field is an adjusted figure, not GAAP. It reports FY2024 operating margin of +7.95% against an actual GAAP operating loss of $(122.1)M, and FY2025 of 6.93% against GAAP 5.91%. All operating margins in this report are computed from EDGAR XBRL.

Additionally, market-share figures (Deere ~50–60% of NA large tractors/combines; precision-ag share ~15–18% Deere vs ~7–9% AGCO) are third-party estimates compiled in prior published analysis, not census data.

1. SEC filings — AGCO Corporation (CIK 0000880266) — primary

The trailing 60-month corpus (since 2021-08-22) was mirrored in full via the SEC EDGAR interface: 413 documents, integrity check reported “OK: 413 files on disk for 413 mirrored manifest rows.” Composition: 5 × 10-K, 15 × 10-Q, 63 × 8-K, 5 × DEF 14A, 294 × Form 4, plus 8-K/A, ARS, Form 3 and DEFA14A.

SEC EDGAR XBRL companyfacts API, CIK 0000880266, via the SEC EDGAR interface — the authoritative source for all multi-year GAAP series used in this report: NetIncomeLoss, OperatingIncomeLoss, RevenueFromContractWithCustomerExcludingAssessedTax, EarningsPerShareDiluted, IncomeTaxExpenseBenefit, GoodwillImpairmentLoss, RestructuringCharges, IncomeLossFromEquityMethodInvestments, CashAndCashEquivalentsAtCarryingValue, LongTermDebtNoncurrent, StockholdersEquity, Goodwill, PaymentsForRepurchaseOfCommonStock, CommonStockDividendsPerShareDeclared, ResearchAndDevelopmentExpense.

2. Earnings call transcripts

3. Peer filings and prior published analysis — used for cross-read

4. Industry, macro and regulatory

5. Quantitative data services — third-party, cross-check only

6. Date-integrity note

All financial, market and factor data is current as of 2026-08-22, with market prices as of the 2026-08-21 close and the latest reported quarter being Q2 2026 (10-Q filed 2026-07-30). No 10-K, 10-Q or 8-K has been filed since 2026-08-03. Older sources are used strictly for historical context: the FY2021–FY2024 10-Ks for the multi-year series, and the 2023/2011/2009 Drive primers only to establish that no relevant primer exists. Third-party market-share estimates carry no date stamp and are labeled as directional throughout.