StoneX Group Inc. (NASDAQ: SNEX) — From Forgotten FCM to 4x Book: A Volatility-Levered Compounder Priced for the Cycle to Hold

Independent fundamental research. General information only — not investment advice. This article contains no buy/sell recommendation and no price target outside the explicitly-labeled “Author’s Take” block below.

⚡ Author’s Take

The author’s own independent opinion and general information — not investment advice. The analysis below carries no position and no price target.

Verdict: HOLD / accumulate-on-weakness. Not-a-short. Medium conviction. A genuinely good, long-mis-priced compounder that the market has finally — and correctly — re-rated, then arguably pushed a step too far at a moment when earnings are running hot. Fair-value zone ≈ $95–$125 (≈ 3.0–3.6x book / ≈ 14–17x normalized EPS); the current $139 prices in continuation of a volatility-flattered earnings run with little margin of safety.

For its entire public life StoneX was the definition of an under-followed, “cheap, commoditized broker” — it traded 1.5–1.9x book and 6–11x earnings every year from 2019 through 2024 while quietly compounding book value per share from ~$6 to ~$24 (~25%/yr) and net income from $170M to $306M. That gap between quality and price was the opportunity, and it has now closed violently: the stock is up ~7x in five years and ~155% in the last ten months, re-rating to 4.2x book (richest ever, 99.98th percentile of its own history) and ~24x trailing earnings. The re-rate is justified in direction — this is a ~20%-ROE, ~20%-book-grower that just made itself the #1 non-bank US futures broker via the RJ O’Brien deal — but it is generous in degree. My hesitation is the earnings base: FY26 is running ~80% ahead of FY25 on a stack of RJ O’Brien accretion (durable) plus a record metals/OTC-derivatives volatility quarter on the US–Iran conflict (not durable). Management itself flagged “moderation” entering Q3, the highest-margin retail/FX segment is mean-reverting, and insiders — including the founder, the CEO, the CFO and the CAO — sold into the all-time high with zero open-market buys. Pay 4x book for a volatility-levered intermediary at a volatility peak and you are making two bets at once (earnings hold and multiple holds).

Framing: quality-compounder-finally-discovered, now priced like one — not a falling knife (price and earnings are both at highs), not a momentum-factor crowd-trade (it doesn’t even load on the Momentum factor), but a high-beta small-cap financial whose risk is that the re-rate and the earnings normalize together. Conviction: medium. Bullish flip: two-plus more quarters proving the FY26 run-rate is a durable base — RJO cross-sell quantified and delivering, ROE holding ~20%+ as market volatility normalizes. Bearish flip: two consecutive quarters of QoQ net-operating-revenue decline (proving Q2-FY26 was a spike) alongside multiple compression back toward 2.5–3x book. Tag: the mid-market plumber the market finally noticed — and then some.

📈 Stock Price Action — Five-Year Event Map

StoneX has round-tripped from obscurity to discovery: split-adjusted, the stock has run from ~$20 (mid-2021) → $139.01 (18-Jun-2026), an all-time-high close, ~7x in five years and +155% off the 52-week low of $54.63 (06-Aug-2025). It now trades far above its 200-day EMA (~$85.5), 52-week range $54.63–$139.01, ~0% off its peak. The move is both a near-quadrupling of book value/earnings and a multiple re-rate from ~1.7x to 4.2x book. (Price moves are Fact; attributed drivers are Interpretation.)

Cycle narrative. (1–2) The 2022 commodity shock was StoneX’s first taste of how violently its volatility-levered model earns in turbulent markets, but the market treated it as a one-off and kept the multiple at ~1.8x book. (3) FY24 was the under-the-radar grind — earnings and book compounded but the multiple did not move, the classic “value trap that wasn’t.” (4–5) The RJ O’Brien announcement in April 2025 was the catalyst that began to change the narrative from “commoditized broker” to “consolidating #1 FCM,” though the stock based through the summer ahead of the close. (6–7) The real re-rate is the last nine months: RJO accretion plus a genuinely extraordinary FY26-Q2 — record net income of $174.3M on a US–Iran-conflict volatility surge and a precious-metals quarter management called the best ever — convinced the market this was a structurally larger, higher-quality franchise, and the multiple expanded from ~2.8x to 4.2x book. The split announcement (26-May-2026) added retail momentum. The open question the rest of this memo addresses: how much of event #7’s earnings is a new floor versus a cyclical spike.

1. Executive Summary

StoneX Group Inc. (formerly INTL FCStone) is a $10.5B-market-cap global financial-services intermediary — a “network” connecting ~54,000 commercial, institutional, payments and retail clients across more than 180 countries to ~40 derivatives exchanges, most global securities markets, ~185 currencies and thousands of OTC products. It earns money the way plumbing does: transaction commissions and bid-ask spreads on enormous volumes, physical-commodity margins (chiefly precious metals), cross-border payment spreads, and interest income on client float. FY25 operating revenue (the meaningful figure, net of $128.3B of physical-commodity cost pass-through) was $4,126.9M, up from ~$1.3B five years ago; net income was $305.9M, and the company earns a through-cycle ROE of ~20–26%.

The investment debate is not about business quality — it is about price and the durability of the current earnings level. Three facts frame it:

1. A real compounder, long mis-priced. Book value per share compounded from ~$6.36 (FY19) to ~$24 (FY25) and ~$33 today (~25%/yr) despite ~16% share dilution; net income nearly doubled in five years; ROE held ~20%+ across a ~10-deal acquisition run without mean-reverting. For all of that, the market paid 1.5–1.9x book and 6–11x earnings every single year — a textbook under-followed value name.

2. A violent, mostly-earned re-rate. The stock is up ~7x in five years and ~155% in ten months, re-rating to 4.2x book (its richest multiple ever) and ~24x trailing / ~19–20x forward earnings. Roughly half the move is genuine earnings/book growth; roughly half is multiple expansion as the market “discovered” the name and the RJ O’Brien acquisition (closed July 2025; $942M; makes StoneX the largest non-bank US FCM) re-cast it as a consolidating franchise leader.

3. An earnings level that is running hot. FY26 H1 diluted EPS of $3.87 is +81% YoY, but it stacks durable RJO accretion on top of a clear cyclical windfall — record OTC-derivatives volumes on the US–Iran conflict and an all-time precious-metals quarter. Management flagged moderation into Q3; the highest-margin segment is already mean-reverting; insiders (founder, CEO, CFO, CAO) sold into the high with no open-market buys.

The moat is narrow and operational, not structural. StoneX’s edge is real but composite: genuine scale in regulatory/clearing infrastructure (a barrier that drove banks out and let StoneX roll up the survivors — a Marathon-favorable capital cycle), plus relationship switching-costs in commercial hedging, plus scope economics from breadth. It is not an Interactive-Brokers-style structural cost moat; share gains come substantially by acquisition, and per-trade economics are thin. The ~20% ROE rests on continued volatility, serial-integration execution and a favorable rate backdrop rather than pricing power.

Net: a high-quality, well-run, diversified intermediary that has migrated from obviously-cheap to fully-priced. The body below argues both sides; the valuation section frames what the $139 price embeds (continuation of a ~20%+ ROE and a ~$7 EPS base) and what would have to break for it to be wrong.

2. Business Overview

StoneX operates a vertically-integrated financial intermediary organized into four client-facing segments plus a large unallocated Corporate cost center. The unifying logic is a “network” or “ecosystem”: one balance sheet, one set of regulatory licenses, clearing memberships and technology rails, sold across many products to clients the bulge-bracket banks find too small or too operationally messy to serve. The company connects clients to ~40 derivatives exchanges, the major global securities exchanges, ~185 FX markets, and makes markets in ~18,000 OTC securities and a vast range of OTC commodity derivatives.

How it actually makes money (FY25 10-K). Gross “Total revenues” of $132,378.2M are deceptive — $128,251.3M of that is the cost of physical commodities sold (almost entirely precious metals moving through the Commercial segment), a pass-through with razor-thin margin. The figures that matter:

• Operating revenues: $4,126.9M (total revenues less physical-commodity cost) — up ~20% YoY and ~5.7x in five years.
• Net operating revenues: $2,052.8M (operating revenues less transaction-based clearing/introducing-broker fees paid away and interest paid to clients) — the closest analog to a “net revenue” line.
• Net income: $305.9M; diluted EPS $4.07 (FY25, pre-split).

Revenue is a blend of (a) commissions and clearing fees on listed futures/options and securities, (b) bid/ask spreads on OTC derivatives, FX, fixed income and physical commodities, © physical-trading margins (dominantly precious metals), (d) payment spreads on cross-border transactions, and (e) interest income earned on client cash and the firm’s fixed-income inventory. Gross interest income was ~$1,734M in FY25, though net of interest paid to clients and on financing it is a far smaller — and partially hedged — contributor.

Segments (FY25, $M):

The top two segments produce ~76% of segment income; Payments is the smallest but the highest-quality (most recurring, ~58% segment-income margin); Retail is the highest-margin but the most cyclical (segment income down ~23% YoY on a TTM basis as FX/CFD “capture” normalized).

Recurring vs. transactional. This is predominantly a transactional, volatility-levered business, not a subscription or spread-banking annuity. Management explicitly frames “volatility is a key driver.” The genuinely recurring pieces — Payments spreads, custody/clearing fees, and the protected portion of client-float interest — are a minority of net operating revenue. The majority scales with client volumes × volatility, which is why FY26 earnings spiked with the markets and why the durability question dominates the thesis. Verdict: a broad, diversified, genuinely global intermediary with real operating scale — but one whose revenue base is more cyclical and transactional than its “ecosystem” branding implies.

3. Industry Dynamics

StoneX occupies the mid-market seam of the financial-intermediation value chain: below the bulge-bracket banks (Goldman, JPMorgan, Morgan Stanley) that exited large swaths of commodity intermediation, FCM clearing, physical trading and correspondent payments after Dodd-Frank/Volcker, and above the fragmented universe of regional brokers and single-product specialists. This positioning is the single most important industry fact, and it is Marathon-favorable: the global financial crisis and the post-2010 regulatory wave drove capacity out of these businesses — banks shed FCM units, exited physical commodities, and de-banked correspondent relationships — leaving a supply-side vacuum that a disciplined consolidator could fill cheaply. StoneX has spent fifteen years rolling up that vacated capacity.

Profit pools and structure. The underlying activities are, individually, commoditized and cyclical: futures clearing is a low-margin, scale-driven utility; FX and CFD spreads compress over time; physical metals trading is thin-margin volume; payments spreads erode with competition and transparency. What makes the aggregate attractive is twofold. First, regulatory and operational complexity is a real barrier to entry — net-capital rules, client-money segregation regimes across dozens of jurisdictions, exchange memberships, clearing relationships, and the technology and compliance overhead to run them are expensive and slow to build, which is precisely why banks chose to exit rather than optimize, and why new entrants are rare. Second, diversification across uncorrelated revenue drivers (commodity vol, equity vol, rates, FX, payments) smooths what would otherwise be violently cyclical single-product economics.

Competitive landscape by segment. Commercial/FCM: Marex (the closest public twin, also a roll-up of bank-retreat assets), ADM Investor Services, producer-affiliated FCMs, and the consolidated remnants of ED&F Man/INTL’s old peer set — post-RJO, StoneX is the #1 non-bank US FCM by client segregated funds (~$7.4B required segregated + ~$6.3B foreign secured at 9/30/25). Institutional: regional broker-dealers, Virtu and Citadel Securities in market-making, bank fixed-income desks. Retail FX/CFD: IG Group, CMC Markets, Plus500, and Interactive Brokers (Forex.com/City Index are StoneX’s own brands). Payments: Corpay (Cambridge), Convera, and the correspondent-banking networks.

Regulation. SEC, FINRA, CFTC, NFA, the UK FCA, and dozens of local regulators. The regulatory burden is simultaneously the moat (barrier to entry) and the risk (capital requirements, segregation rules, and episodic enforcement). Verdict: a structurally mediocre set of businesses being assembled into a structurally better diversified whole, riding a favorable, multi-year capital cycle of bank retreat. The industry is “good” for the surviving consolidators specifically because it is hard and unglamorous — but it remains cyclical and volatility-dependent, and the favorable capital cycle is now mature, with Marex and others competing for the same roll-up assets.

4. Competitive Position

The central question: does StoneX have a durable moat, and if so, of what Greenwald type? The honest answer is a narrow, composite, operationally-grounded moat — not a wide structural one.

What is real:

• Economies of scale in regulatory/clearing infrastructure. Running self-clearing across futures, securities and FX in dozens of jurisdictions has high fixed costs that StoneX spreads across a large and growing client/volume base. This is a genuine supply-side cost advantage relative to a sub-scale would-be entrant, and it is why the roll-up math works — each acquired book of business plugs into existing rails at low marginal cost. It is durable because it is expensive and slow to replicate and because regulation deters new entry.
• Relationship switching costs in commercial hedging. A mid-sized agricultural processor or energy end-user that relies on StoneX for risk-management advice, structuring, execution, clearing and sometimes financing does not re-tender that relationship casually. These are moderate, real switching costs — but they are relationship-based and people-mediated, not contractual lock-in, and the firm’s ~5,400 employees are the moat as much as the technology is.
• Scope/breadth. The ability to serve a client across commodities, FX, securities and payments from one counterparty has genuine value and creates cross-sell optionality.

What is oversold:

• The “network/ecosystem” claim is scope economics plus a flywheel, not a true network effect. A new client does not make the platform more valuable to existing clients the way a marketplace or exchange does. Management’s framing implies Metcalfe-style network value that the economics do not support.
• There is no Interactive-Brokers-style structural cost wedge. IBKR (see a prior published analysis, 19-Jun-2026) is a software-and-automation compounder with a ~77% pre-tax margin, ~43% ROIC, and ~30%+ organic account growth — a lowest-cost-producer moat that earns its returns without acquisitions. StoneX is the inverse construction: a people-and-relationship-driven, acquisition-built diversified roll-up whose growth is substantially bought. Both earn high ROEs, but IBKR’s is structural and self-reinforcing while StoneX’s is operational, volatility-dependent and integration-dependent.

The Greenwald share-stability test — does StoneX hold or gain share without competing returns away? — passes, but largely by acquisition rather than by organic capture, which is a weaker form of the test. The ROIC test is muddied by the balance sheet (operational float dwarfs invested capital), so ROE is the better lens: ~20%+ through cycle is strong evidence of a moat, but the moat’s durability is contingent on three things continuing — market volatility, successful serial integration, and a non-hostile rate environment — none of which is a pricing-power moat in the classic sense.

Verdict: a real but narrow composite moat (scale-in-infrastructure + relationship switching costs + scope), best described as “operational excellence plus roll-up scale with thin per-trade economics.” It supports durable high-teens-to-low-20s ROE, but it is not the wide, self-widening structural moat the 4.2x-book multiple implicitly underwrites. Investors are paying a structural-compounder multiple for an operational-compounder business.

5. Growth History and Forward Opportunities

History. The growth record is genuinely impressive and is the strongest pillar of the bull case. Operating revenue grew from ~$1.3B (FY20) to $4,126.9M (FY25); net income from $169.6M to $305.9M; book value per share from $9.01 (FY20) to ~$24 (FY25). Across this run ROE held ~18–26% — the acquisitions did not dilute returns, which is the rare and important tell that this roll-up created value rather than destroying it (contrast the typical Marathon cautionary tale where serial acquirers’ returns mean-revert to the cost of capital).

Growth has been roughly half organic, half acquired, and the two reinforce: acquisitions add client books and capabilities that then cross-sell into the existing network. The major deals — Gain Capital (2020, retail FX/Forex.com + City Index), the Cargill petroleum business, CDI, JB Drax, Octo Finances, Benchmark (research/IB), and the transformational RJ O’Brien (2025) — progressively broadened the franchise from a niche commodities FCM into a four-pillar global intermediary.

Forward opportunities:

• RJ O’Brien integration and cross-sell. RJO roughly doubles StoneX’s listed-futures client base and makes it the #1 non-bank US FCM. The under-appreciated lever is cross-sell: RJO’s clients were largely listed-only; StoneX can sell them OTC derivatives, FX, physical and payments. Management repeatedly cites this but has declined to quantify it — a hypothesis, not yet a number.
• Payments expansion. The highest-quality segment, with a long runway in cross-border B2B, NGO and bank flows to under-served corridors (180+ countries, 140+ currencies). This is the piece most deserving of a premium multiple.
• Continued bank-retreat consolidation. The capital cycle still offers roll-up targets, though competition (Marex) is rising.
• Geographic and capability build-out: the Birmingham (UK) hub, Madrid/Paris/APAC expansion, digital-asset and bullion initiatives, and platform/technology investment.

Quality of growth. Mixed-to-good. The durable portion — RJO scale, payments, client/volume secular growth, market-share gains in market-making — is high-quality. The cyclical portion — volatility-driven transactional capture and rate-driven interest income — is lower-quality and is currently inflated. The FY26 surge (H1 EPS +81%) is the acute illustration: real franchise growth layered with a volatility windfall. Verdict: high-quality growth engine (proven ~20% compounding without ROE erosion), but the FY26 growth rate is flattered by cyclical factors that will not repeat at the same magnitude. Underwrite the trend, not the spike.

6. Financial Quality

Earnings power and trajectory. Net income compounded $169.6M → $305.9M (FY20→FY25); FY26 is inflecting hard, with H1 net income running ~80% ahead of the prior-year half and Q2-FY26 setting a record at $174.3M. Diluted EPS by quarter (pre-split): $1.17 → $0.97 → $0.85 → $1.09 (FY25) → $1.73 → $2.14 (Q2-FY26). TTM diluted EPS is ~$5.80.

Margins and returns. On the net operating revenue base, net margins are healthy and improving with scale (operating leverage is real — Corporate overhead grows slower than revenue, and the Retail segment showed +40% segment-income growth on +15% net-operating-revenue growth in a strong quarter). ROE is the right return metric (ROIC is distorted by the enormous operational float on the balance sheet): ROE ran 34.8% (FY20) → 18.5% → 26.3% → 23.6% → 20.7% → 19.8% (FY25), and is annualizing toward the mid-20s in FY26. The consistency of ~20%+ ROE through a decade of acquisitions is the single best quantitative evidence of franchise quality.

Quality of earnings — clean, with two caveats.

• Clean: SBC is modest (~$49M, ~16% of net income); there are no large non-cash gains flattering FY25; the EPS is GAAP, not a heavily-adjusted construct. Capex is tiny (~$65M), consistent with an asset-light intermediary.
• Caveat 1 — cash flow is uninformative. Operating cash flow swings wildly ($2.1B, −$0.2B, −$24M, $0.5B, $4.4B across FY21–25) because it is dominated by working-capital movements in client balances and inventory — these are not free cash flow and the ROIC-derived “FCF” figures are meaningless for this business. Read net income, not cash flow.
• Caveat 2 — net income diverges from cash from operations for the same reason; this is structural to a broker-dealer and not a red flag, but it means standard FCF-yield analysis does not apply.

Balance sheet. Total assets $45.3B (FY25), almost entirely operational: client receivables, securities borrowed/purchased under resale, physical inventory, and deposits with clearing organizations, funded by client payables and securities-financing liabilities. The $21.8B of “debt” on aggregator screens is ~95% operational financing float, not corporate leverage. Actual holding-company debt is ~$1.16B of senior secured notes (7.875% due 2031, 6.875% due 2032) plus revolvers — modest, roughly 1–1.5x EBITDA, investment-grade-style. Equity is $2,377M (FY25), lightly encumbered by intangibles (~$736M goodwill+intangibles, mostly newly added by RJO; tangible book is modestly below reported book). The firm is well-capitalized against regulatory requirements with comfortable excess net capital. Verdict: economics genuinely improve with scale (operating leverage + ~20% ROE), earnings quality is clean if you use the right metric (net income, not cash flow), and the balance sheet is far more conservative than screens suggest. The financial profile supports the “quality” half of the thesis — the debate is entirely about the price paid and the cyclicality of the current earnings level.

7. Capital Allocation

Capital allocation is above-average and is a genuine support for the bull case — with one structural feature (100% retention, zero direct shareholder yield) and one behavioral flag (insider selling) to weigh against it.

M&A — disciplined, value-creative, in-sector consolidation. The defining capital-allocation fact is the ~10-deal, six-year roll-up that did not erode returns — ROE held ~20% throughout while book and earnings nearly doubled. The capstone, RJ O’Brien, was announced 14-Apr-2025 and closed 31-Jul-2025 for $942.0M total consideration ($651.9M cash + $300.1M stock via 3.09M treasury shares − a $10.0M receivable), plus ~$125.7M of assumed RJO subsidiary debt. The cash leg was funded with a $625M 6.875% senior secured notes issuance — no new primary equity dilution. Purchase accounting: ~$357M tangible net assets, ~$410.6M intangibles (mostly client relationships), and $174.4M goodwill — the first time the historically asset-light balance sheet carried meaningful goodwill (total goodwill $61M → $298M). The deal is strategically coherent (makes StoneX the #1 non-bank US FCM, deepens the clearing-scale moat, brings cross-sell optionality) and was financed conservatively. Integration is reportedly “on track” for substantial completion within FY26, with synergies scaling toward a ~$50M run-rate by ~FY27. This is the rare roll-up that passes the Marathon test — high returns that did not mean-revert as capital was deployed.

Returns to shareholders — none, by design. StoneX has never paid a dividend and buybacks are dormant (no repurchases in FY24 or FY25; a token 2.25M-share authorization runs through FY26). It is a net issuer, not a returner — diluted share count rose ~16% over five years (64.7M → 75.2M) from acquisition stock and equity comp, and the recent jump toward ~81.5M diluted reflects RJO shares plus comp (there are no convertible notes). The logic is defensible — reinvesting at ~20% ROE beats paying out — but it means total return depends entirely on per-share book/earnings compounding, and the ~16% dilution is a real drag investors should net against the headline growth.

Incentive alignment — above-average, with capital-discipline governors. Unusually for this coverage set, the comp plan uses genuine capital-return metrics: the annual bonus references adjusted ROE, cost control, adjusted-EBITDA growth and share price, and the LTIP references adjusted ROE, book value per share, tangible BVPS, debt-to-equity, TSR and EPS. That is a meaningfully better-aligned structure than the revenue-or-EBITDA-only plans common among peers. Founder Sean O’Connor owns ~6.27% (~3.28M shares); insiders and directors together ~11.8% — substantial skin in the game. The mild knock: the ROE target (~15.6%) sits below the ~20% actually achieved, so the bar is not especially stretching. Leadership transitioned in late 2024 — O’Connor (the 22-year CEO and M&A architect) to Executive Vice-Chairman, Philip Smith to CEO, with Charles Lyon as President and a CFO transition underway.

Insider behavior — mildly-to-moderately negative. Across the trailing Form-4 record there are zero open-market purchases (code P) at any price. The dominant pattern is exercising deep-in-the-money options and selling into strength. In April–June 2026, near the all-time high, named insiders sold heavily: founder O’Connor ~450K shares (including ~150K discretionary, non-10b5-1 sales at $131–135 — right at the high), the CFO ~190K, director Maurer ~140K, plus the CEO, President and CAO. Many sales are 10b5-1-planned (diversification, not signal), but the combination of zero buys and the founder’s discretionary sales at the top is a valuation tell: those closest to the business are treating 4.2x book / 24x earnings as fair-to-full. O’Connor still holds 6.27%, so this is trimming, not exit. Verdict: management has allocated capital intelligently (disciplined, ROE-preserving consolidation, conservative financing, well-aligned comp), which underwrites the franchise quality — but the absence of any shareholder yield, the steady dilution, and the insider selling into the high all argue the value creation is being shared with the market at a full price, not handed to new buyers at a discount.

8. Changes and Headwinds — Last Two Years

Strategic and structural changes:

• RJ O’Brien acquisition (announced Apr-2025, closed Jul-2025) — the largest deal in company history; six total acquisitions closed in FY25, plus Benchmark (research/market-making).
• Leadership transition — Sean O’Connor → Executive Vice-Chairman; Philip Smith → CEO; President and CFO changes. A generational handoff from the founder/architect; execution risk and a less-tested capital-allocator at the top are now live considerations.
• 3-for-2 stock split announced 26-May-2026 — cosmetic, but signals management’s read of retail demand and confidence; not yet effective as of this writing (current price/share count are pre-split).
• Litigation cleared — several legacy matters (BTIG, option-seller, GAIN patent) resolved in Q2-FY26, a modest legal-expense tailwind.
• Build-out — Birmingham (UK) operating hub, Madrid/Paris/APAC expansion, bullion/digital-asset and payments-technology (Xpay) initiatives.

Headwinds / tailwinds-turning:

• The FY26 earnings surge is partly cyclical. Q2-FY26’s record stacked RJO accretion on a US–Iran-conflict volatility spike (OTC derivatives +98% YoY) and an all-time precious-metals quarter. Management explicitly flagged “moderation coming into April” (Q3). The highest-margin Retail/FX segment already swung from −34% to +15% YoY net-operating-revenue as capture mean-reverted — the clearest tell that part of the surge is a spike.
• Rate sensitivity (quantified, and smaller than feared). Disclosed sensitivity to a 100bp move is ~$47.6M net income / ~$0.58 EPS annualized (~10% of TTM earnings per 100bp), actively hedged down via a ~$1.8B SOFR swap book (~3.38%, ~2yr). A realistic 50–75bp easing cycle trims ~5–8% of earnings — material but not catastrophic. Unlike IBKR (57% of net revenue is NII), for StoneX volatility normalization, not rate cuts, is the bigger cyclical risk.
• Valuation/positioning risk — the multiple has re-rated to its richest-ever level just as earnings run hot; the two can compress together.

Verdict: the changes are net franchise-strengthening (RJO scale, payments, litigation cleared), but they are now fully reflected in both the earnings and the multiple. The headwinds — cyclical earnings normalization, modest rate drag, founder handoff, and a stretched multiple — are the reasons the risk/reward at $139 is balanced rather than asymmetric to the upside.

9. Risk Analysis

Catastrophic-loss assessment: low but non-zero. The genuine tail risks for an intermediary of this type are (a) a large counterparty/clearing default or a physical-commodity bad-debt event that overwhelms risk controls (cf. historical FCM blow-ups), and (b) a funding/liquidity squeeze in an extreme market dislocation. Both are mitigated by diversification, excess net capital, and a hedged book, but they are the scenarios that would impair the equity rather than merely compress the multiple. Total permanent loss is improbable; a 30–50% drawdown on combined earnings normalization + multiple compression is a realistic, non-tail risk.

10. Valuation Discussion (Embedded Expectations)

Where the multiple sits. At $139 (pre-split), StoneX trades at ~24x trailing / ~19–20x forward earnings, 4.2x book value, and ~10x EV/EBITDA (on EBITDA, treating the operational balance sheet correctly). The own-history percentiles are emphatic: P/B 99.98th, P/S 99.98th (distorted by gross transactional revenue — ignore), composite 97.4th, P/E 92.4th. The clean read is P/B: 4.19x is the richest valuation in the company’s public history by a wide margin — every year from FY19 to FY24 ended between 1.7x and 1.9x, and even the FY25 year-end was 2.82x.

What the price embeds. Decompose 4.2x book at ~20% ROE: the implied normalized earnings yield is ROE ÷ (P/B) = 20% ÷ 4.19 ≈ 4.8%, i.e., ~21x normalized earnings. For a business compounding book at ~18–20% with a clean ~20% ROE, ~21x is defensible but not cheap — it is roughly a market multiple for a faster-than-market, higher-ROE compounder, leaving little margin of safety. The price is underwriting two things simultaneously: (1) that the ~$7 FY26 EPS base is durable rather than a volatility peak, and (2) that ~20%+ ROE and ~18–20% book growth continue. If both hold, the stock compounds with book and the multiple need not expand further. If either breaks — earnings normalize toward a ~$5.0–5.5 base or ROE reverts to mid-teens — the 4.2x multiple is exposed.

Scenario analysis (pre-split, ~76M shares):

At $139 the stock sits in the upper half of the base case — fairly valued on a continuation scenario, with asymmetric downside if the bear’s earnings-normalization-plus-de-rate plays out (the two compress together, which is why the bear band is ~30–50% below spot). The peer cross-check supports “full, not cheap”: the boutique broker-dealer twins (Piper Sandler, Perella Weinberg, Evercore, Jefferies) and the closest operating comp (Marex) do not trade at 4x book; StoneX’s premium is earned by its higher ROE and diversification but is now priced in.

Embedded-expectations conclusion: the market is correctly pricing StoneX as a quality compounder rather than a cheap broker — the re-rate is largely right — but at $139 it has moved from pricing the business to pricing the cycle, and is underwriting a continuation of an earnings level that management itself has signaled will moderate. No price target; the above are scenario illustrations, not forecasts.

11. Variant Perception

Consensus. Sell-side coverage is thin and recently turned constructive (Jefferies, William Blair on the calls; the rest is Zacks-tier “Strong Buy” momentum commentary). The emerging consensus narrative: an under-followed compounder finally being discovered, with RJO making it the #1 FCM and a diversified ecosystem deserving a re-rate — essentially a momentum-and-quality story with the multiple expansion treated as catch-up to fair value.

Strongest bull case. StoneX is a proven ~20%-ROE compounder that has nearly doubled earnings and book in five years without eroding returns through a ~10-deal roll-up — a genuine rarity. RJO is a transformational franchise step-up (#1 non-bank US FCM) with un-quantified cross-sell optionality on top. The business is more diversified, more global, and structurally larger than at any point in its history, and Payments offers a high-quality recurring-growth runway. At ~20x forward earnings for a 20% compounder, it is not expensive if the compounding continues — and the market historically under-valued exactly this kind of unglamorous quality.

Strongest bear case. This is a volatility-levered, thin-per-trade-economics intermediary at a volatility peak, re-rated to 4x book — the richest multiple in its history — just as the founder/architect steps back and insiders sell into the high with zero buys. The FY26 surge is partly a one-off (Iran-conflict OTC vol, metals super-quarter) that management has already flagged will moderate; the highest-margin segment is mean-reverting; rate cuts shave ~5–8%; and a re-rate-plus-earnings-normalization can compress the stock 30–50% without anything “going wrong” operationally. You are paying a structural-compounder multiple for an operational-compounder business at a cyclical high.

The 3–5 assumptions that matter most, with falsification tests:

1. The FY26 earnings base is durable (~$7) rather than a spike toward ~$5.5. Falsified by: two consecutive quarters of QoQ net-operating-revenue decline as the metals/OTC-vol windfall fades.
2. ROE stays ~20%+ as market volatility normalizes. Falsified by: ROE drifting to mid-teens over the next 3–4 quarters.
3. RJO cross-sell delivers measurable revenue synergy. Falsified by: management continuing to decline to quantify it, or integration slipping past FY26.
4. The 4.2x-book multiple holds or compounds with book. Falsified by: de-rating toward 2.5–3x book even on stable earnings (the “discovery trade unwinds”).
5. The founder handoff preserves capital-allocation discipline. Falsified by: a pricier, lower-ROE acquisition or a strategic drift under the new CEO.

Factor-positioning read (subordinate to the thesis). The tape confirms the framing: StoneX is at an all-time high (rs_peak ~0), beta ~1.04, with elite risk-adjusted history (y3 +72%/yr, Sharpe ~2.0), but it does not load on the Momentum factor — its loadings are Market (+0.90), DividendYield (+0.45), SmallSize (+0.44), CreditRisk (+0.27) and Financials/Broker-Dealers. This is a small-cap, high-beta financial whose price has risen on earnings, not a crowded factor-momentum trade and not a falling knife. The positioning evidence says consensus is not obviously offsides in either direction — which is itself the point: the easy mispricing (cheap, ignored) has been arbitraged away, and what remains is a fairly-priced quality name where the edge is entry price, not direction. The variant perception is therefore modest: not “the market is wrong,” but “the market has fully caught up, and is now extrapolating a cyclically-elevated earnings level.”

12. Fact vs. Interpretation

13. Open Questions

1. How large is the RJO cross-sell opportunity, quantitatively? Management repeatedly cites it but declines to size it — the single biggest swing factor in the durable-growth case.
2. What is the normalized (mid-cycle-volatility) earnings power? Is the base ~$5.5 or ~$7? Two to three more quarters across a calmer tape are needed to know.
3. How much of FY26-Q2 was the Iran-conflict spike specifically, and what does a “normal” precious-metals/OTC-derivatives quarter look like post-RJO?
4. Will the new CEO preserve capital-allocation discipline, or pursue a pricier, ROE-dilutive deal now that the currency (4x book stock) is strong?
5. Why no shareholder returns at all (no dividend, dormant buyback) — and would management ever lean into the buyback if the stock de-rated, given the founder’s discretionary selling at the high?
6. How would the balance sheet behave in a severe market dislocation — is the financing/float genuinely robust, or is there hidden funding fragility?

14. What Must Be True

For the bull case (stock compounds from $139):

• The FY26 earnings base proves durable (~$7 EPS holds, not a spike toward ~$5.5), with RJO accretion and secular volume/share growth offsetting the fading volatility windfall.
• ROE stays ~20%+ as market volatility normalizes; RJO cross-sell becomes a quantifiable, delivering revenue lever.
• The 4.2x-book multiple holds while book compounds ~18–20%/yr, so the stock tracks book higher.
• Falsification test: two consecutive quarters of QoQ net-operating-revenue decline, or ROE drifting to mid-teens, within the next 3–4 quarters — would prove FY26 was a cyclical peak and the multiple un-earned.

For the bear case (stock de-rates 30–50%):

• FY26 is revealed as a volatility/rate-flattered peak; net operating revenue rolls over for two-plus quarters as metals/OTC-derivatives capture normalizes and the Retail segment keeps reverting.
• Rate cuts trim ~5–8%, and the multiple compresses from 4.2x toward 2.5–3x book as the “discovery trade” matures and growth decelerates.
• Falsification test: earnings grow sequentially through a calmer-volatility tape (proving the base is structural, not cyclical), and RJO cross-sell is quantified and delivering — would refute the cyclical-peak thesis and validate the re-rate.

The elegance of this setup is that both falsification tests key off the same observable — the sequential trajectory of net operating revenue across the next two to three quarters as market volatility normalizes. That is the number to watch.

A source appendix follows below.

APPENDIX A — Standard Diligence Questionnaire — StoneX Group Inc. (NASDAQ: SNEX)

Supplemental to the analysis above. Fact / Interpretation / Assumption labeled where it matters.

General

What thoughtful questions have other investors asked about this company? The recurring questions, consistent with the thin but constructive coverage: (1) How much of the recent earnings surge is durable vs. a volatility/rate windfall? — the central debate. (2) Can a serial acquirer keep ~20% ROE as deals get larger? — answered “yes so far,” but RJO is the biggest test. (3) Does the “ecosystem” actually create cross-sell value, or is it just a collection of brokerages? (4) Why does a 20%-ROE compounder trade at 1.5x book? — historically the value question; now inverted to “is 4.2x book too far?” (5) Read operating revenue, not the $132B gross transactional revenue — the most common rookie error on the name.

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low? Interpretation: a cyclical high. FY26 H1 EPS is +81% YoY, stacking durable RJO accretion on a volatility windfall (Iran-conflict OTC derivatives +98% YoY, record metals quarter). Management flagged “moderation.” Normalized earnings power is likely below the FY26 run-rate.

Driven by external environment or internal action? Both. Internal: the RJO acquisition and a decade of franchise-building (durable). External: elevated market volatility and still-high short rates boosting transactional capture and client-float interest (cyclical). The split is roughly half-half.

How stable are revenues? Low-to-moderate stability. Predominantly transactional and volatility-dependent; the genuinely recurring pieces (Payments spreads, custody/clearing fees, protected client-float interest) are a minority of net operating revenue. Diversification across uncorrelated drivers smooths but does not eliminate cyclicality.

Outlook for products/services? Positive secular backdrop (bank retreat → consolidation; cross-border payments growth; market-share gains in market-making), but the rate of growth is currently inflated by the cycle.

How big is the market — growing, shrinking, domestic, international? Large and global (clients in 180+ countries; ~40 exchanges; 185 FX markets). Growing in payments and in consolidated FCM share; mature/competitive in FX-CFD and physical metals. Predominantly international in reach.

Business Quality & Competitive Moat

Is the industry getting more or less competitive? Interpretation: the favorable capital cycle (bank retreat) is maturing — Marex and other survivors now compete for the same roll-up assets, and FX/CFD/payments spreads structurally erode. Still a high-barrier industry (regulation, capital, clearing infrastructure), but competition for growth is rising.

How profitable is the business (ROIC, ROE)? ROE ~20% FY25 (range 18–26% through cycle), annualizing toward mid-20s in FY26. ROIC is not meaningful — the operational float dominates invested capital; use ROE. Returns are strong and have been durable across the roll-up.

How profitable is the industry — competitors, barriers? Individually commoditized/thin-margin activities; the aggregate is attractive only for scaled survivors. Barriers (net-capital rules, segregation regimes, clearing memberships, technology/compliance overhead) are real and deter entry — which is the moat.

Can the business be easily understood? Moderately difficult. The gross-vs-operating-revenue distinction, the operational balance sheet (the “$21.8B debt” is float, not leverage), and the segment interplay require care. Net income is the clean earnings metric; cash flow is uninformative.

Undermined by foreign low-cost labor? No — regulated financial intermediation, not labor-arbitrage-exposed.

Do brands matter? Moderately — Forex.com/City Index (retail), StoneX (institutional/commercial) and RJ O’Brien (FCM) carry recognition, but the moat is infrastructure and relationships, not brand.

Nature of competition / switching costs? Competition is on execution quality, breadth, balance-sheet capacity, and price. Switching costs are moderate and relationship-based in commercial hedging; low in retail FX and physical metals.

Financial Condition & Balance Sheet

Assets not fully recognized on the balance sheet? Interpretation: the regulatory licenses, clearing memberships, and client relationships are under-recognized economically (only partly capitalized as intangibles via acquisitions). The network/scale is an off-balance-sheet economic asset.

Off-balance-sheet liabilities? Standard for a broker-dealer (client guarantees, clearing obligations, derivative exposures) — disclosed; managed via risk controls and excess net capital. No unusual off-balance-sheet leverage identified.

How conservative is the accounting? Reasonably conservative. GAAP EPS (not a heavily-adjusted construct); modest SBC (~16% of NI); no large non-cash gains flattering results; goodwill only recently meaningful (RJO). Cash-flow volatility is structural (client balances), not an accounting red flag.

How CapEx-hungry? Asset-light — capex ~$65M (FY25), a small fraction of earnings. Growth capital goes into working capital (client float) and acquisitions, not PP&E.

Capital Allocation & Management

How much FCF does it generate, and how is it used? Conventional FCF is not measurable (cash flow dominated by client-balance swings). Economic earnings (~$306M FY25, rising) are 100% retained and reinvested into working capital and acquisitions at ~20% ROE. No dividend; buyback dormant.

Significant acquisitions recently? Yes — RJ O’Brien ($942M, closed Jul-2025), the largest in company history, plus five other FY25 deals (incl. Benchmark). Disciplined and conservatively financed ($625M notes, no primary equity).

Buying back shares? No (dormant; token authorization). The company is a net issuer (~16% dilution over five years from deal stock + comp).

Issuing large amounts of stock to insiders? Modest equity comp (~$49M SBC); a 3.09M-share treasury issuance funded part of RJO. Not egregious, but dilution is a real per-share drag.

Compensation policy of directors/management? Above-average alignment: bonus on adjusted ROE/cost control/EBITDA growth/share price; LTIP on adjusted ROE/BVPS/TBVPS/debt-to-equity/TSR/EPS. Founder O’Connor owns ~6.27%; insiders ~11.8%. Mild knock: ROE target (~15.6%) below achieved (~20%).

Motivations of management? Builders/compounders (the O’Connor-era roll-up created real value). Caveat: zero open-market buys and discretionary insider selling at the ATH suggest they view current value as fair-to-full.

Valuation & Market Data

ADR, MLP, or K-1 issuer? No — a US C-corporation (NASDAQ: SNEX), standard 1099 treatment.

Dividend policy? None. 100% retention.

How profitable is the business? ~20% ROE, ~$306M net income FY25 (rising sharply in FY26). High-quality returns on the right metric (ROE), distorted on the wrong ones (ROIC, margins on gross revenue).

Is net income diverging from cash from operations? Yes — structurally, because operating cash flow is dominated by client-balance working-capital swings. This is normal for a broker-dealer; use net income, not cash flow, and do not read the divergence as a quality flag.

Risks & Downside

What would cause the stock to decline? Earnings normalization as volatility/rates fade; multiple compression from the richest-ever 4.2x book; a disappointing quarter revealing FY26 as a peak; rate cuts; an integration stumble or a pricier ROE-dilutive deal; a regulatory/credit event.

Risk of catastrophic loss? Low but non-zero — a large counterparty/clearing default, physical-commodity bad-debt event, or funding squeeze in an extreme dislocation are the tail scenarios that would impair equity (vs. merely compress the multiple). Mitigated by diversification and excess net capital.

Chance of total loss? Very low — diversified, profitable, well-capitalized, investment-grade-style holdco leverage. The realistic downside is a 30–50% drawdown (earnings + multiple compression), not a wipeout.

Recent News & Events

Has the business environment changed recently? Yes — RJO close (largest-ever deal), a leadership transition (O’Connor → Exec Vice-Chairman; Smith → CEO), a 3-for-2 split announcement (May-2026), legacy litigation cleared, and a volatility-driven record Q2-FY26.

Significant acquisitions? RJ O’Brien + five other FY25 deals + Benchmark.

Change in accounting policies? None material identified; goodwill now meaningful post-RJO.

Recent changes — new markets, facilities, management? Birmingham (UK) hub, Madrid/Paris/APAC build-out, bullion/digital-asset and payments-technology initiatives; new CEO and CFO transition.

APPENDIX B — Source Appendix — StoneX Group Inc. (NASDAQ: SNEX)

Primary sources prioritized. All figures in the analysis reconcile to these. Accessed 2026-06-21 unless noted.

Primary — SEC filings (CIK 0000913760)

• StoneX Group Inc. Form 10-K, FY2025 (period ended 30-Sep-2025) — segment note (Commercial / Institutional / Self-Directed-Retail / Payments / Corporate operating revenue, segment income, net operating revenue); income statement (total revenue $132,378.2M; physical-commodity cost $128,251.3M; operating revenue $4,126.9M; net operating revenue $2,052.8M; net income $305.9M); balance sheet; interest income/expense detail; regulatory net-capital and client-segregation disclosures. Mirrored locally at output/SNEX/sources/10-K/.
• StoneX Group Inc. Form 10-Q, Q2-FY2026 (period ended 31-Mar-2026) — quarterly income statement (record net income $174.3M; diluted EPS $2.14); rate-sensitivity disclosure (~$0.58 EPS / 100bp); client-float and segment detail. output/SNEX/sources/10-Q/.
• Form 10-Q, Q1-FY2026 (ended 31-Dec-2025) — diluted EPS $1.73; segment trajectory.
• Forms 8-K (2025–2026) — RJ O’Brien acquisition announcement (14-Apr-2025) and closing (31-Jul-2025); $625M 6.875% senior secured notes due 2032 issuance; 3-for-2 stock split announcement (26-May-2026); earnings releases. output/SNEX/sources/8-K/.
• DEF 14A proxy statement (FY2025) — executive compensation metrics (annual bonus: adjusted ROE / cost control / adjusted-EBITDA growth / share price; LTIP: adjusted ROE / BVPS / TBVPS / debt-to-equity / TSR / EPS); insider/director ownership (Sean O’Connor ~6.27%; insiders+directors ~11.8%); leadership transition (O’Connor → Executive Vice-Chairman; Philip Smith → CEO). output/SNEX/sources/DEF_14A/.
• Forms 3/4/5 (2025–2026) — insider-transaction corpus: open-market sales by O’Connor (~450K shares incl. ~150K discretionary at $131–135), CFO (~190K), director Maurer (~140K), CEO Smith, President Lyon, CAO Schroeder; zero open-market purchases (code P) in the trailing period. Listed in output/SNEX/sources/filing_index_SNEX.txt.

Primary — Company disclosures

• StoneX FY2026-Q2 earnings call transcript (May-2026) and Q1-FY2026 / Q4-FY2025 transcripts — management framing of RJO integration (“on track,” ~$50M synergy run-rate by ~FY27), FY26 surge drivers (RJO + Iran-conflict OTC-derivatives volatility + record precious-metals quarter), “moderation coming into April” comment, interest-income/rate sensitivity, client-float balances. Source: ROIC.ai earnings-call tools (list_earnings_calls, get_earnings_call_transcript).
• StoneX investor relations — segment definitions, RJO deal materials, corporate/strategy disclosures.

Quantitative data sources

• ROIC.ai (third-party aggregator; reconciled to filings) — income statement, balance sheet, cash flow (FY2019–2025 + quarterly FY2025–26); profitability ratios (ROE, margins); enterprise value; valuation multiples (P/E, P/B, EV/EBITDA by year); per-share data (BVPS, TBVPS, EPS).
• AZI / azitrading.com — daily split/dividend-adjusted price CSV (full history; current $139.01 at 18-Jun-2026; 52-week range $54.63–$139.01; 200-EMA ~$85.5; beta ~1.04); valuation_index own-history percentiles (P/E 92.4th, P/B 99.98th, P/S 99.98th, composite 97.4th; TTM EPS $5.80; BVPS ~$33).
• FactorsToday (third-party factor model) — factor loadings (Market +0.90, DividendYield +0.45, SmallSize +0.44, CreditRisk +0.27, Financials/Broker-Dealers; no Momentum-factor load; R² ~0.32); leaderboard (y1 +143%, y3 +72%/yr, y5 +48%/yr; Sharpe y3 ~2.0; lifetime maxDD −86.6%); stock-info (alpha +0.53, rs_peak ~0); related/factor-similar peers (PIPR, PWP, EVR, JEF, OWL).

Comparables referenced

• Interactive Brokers (NASDAQ: IBKR) — referenced for the brokerage-moat contrast (structural lowest-cost-producer / software-automation model vs. StoneX’s acquisition-built relationship/scale model; IBKR ~57% NII vs. StoneX’s predominantly transactional mix).
• Boutique broker-dealer / capital-markets comps for valuation context: Piper Sandler (PIPR), Perella Weinberg (PWP), Evercore (EVR), Jefferies (JEF), Marex (closest operating twin).

Methodology notes

• Read operating revenue ($4,126.9M), not gross transactional revenue ($132,378.2M): ~97% of “total revenue” is physical-commodity cost pass-through.
• Use ROE, not ROIC: the operational balance-sheet float (~$45B assets) dwarfs invested capital and renders ROIC/leverage screens meaningless; the “$21.8B debt” is ~95% operational financing, not corporate leverage (actual holdco debt ~$1.16B).
• Valuation is pre-split: the 3-for-2 split (announced 26-May-2026) is not yet effective; price $139, ~76M shares, TTM EPS $5.80, BVPS ~$33, P/B 4.19x are internally consistent on a pre-split basis.
• Cash flow is uninformative; use net income — operating cash flow is dominated by client-balance working-capital swings.
• Third-party aggregated data (ROIC.ai, AZI, FactorsToday) treated as cross-checks; SEC filings are primary and authoritative for all material figures.