AngloGold Ashanti plc (NYSE: AU) — Record Cash, Rented Margins: A High-Cost Gold Senior Re-Rated as if Its Discount No Longer Applied

Independent equity research, for general information only. Report date: 2026-07-03. All figures USD unless noted. AngloGold Ashanti is a UK-domiciled plc and files with the SEC as a foreign private issuer (Form 20-F annual, Form 6-K interim); it reports under IFRS. “AISC” = all-in sustaining cost per ounce, the industry’s headline unit-cost metric. Financials are FY2025 (year ended 31 December 2025) unless stated.

⚡ Claude’s Take

This block is the author’s own independent, subjective opinion and general information only — not investment advice and not a recommendation to buy or sell any security. The body of this article (Sections 1–15) takes no position and carries no price target; only this block does otherwise. Do your own research and consult a licensed adviser before investing.

Verdict: HOLD / accumulate-on-weakness / not-a-short. Medium conviction. Fair-value zone ~$70–90 (roughly 5.5–6.5× forward adjusted EBITDA on a normalized-but-still-strong gold price). Accumulate sub-$70; do not chase above ~$105. Tag: “Record cash, rented margins.”

AngloGold has done almost everything right operationally, and none of it changes what the stock is. Over four years CEO Alberto Calderón has turned a South-African-discounted, high-cost, over-complex miner into a UK plc with a primary NYSE listing, a fortress balance sheet (net cash for the first time in memory), record free cash flow ($2.9bn), the highest dividend yield in the senior peer group (~5.4%), a cleanly-priced all-stock Centamin deal that added a genuine Tier-1 asset (Sukari), and a real Nevada growth option (Arthur). The FY2025 numbers are cash-backed and free of accounting tricks — statutory earnings actually sit below headline earnings, the opposite of a company flattering its print. This is the best-executed self-help story in the senior gold space.

And yet: strip the gold price out and the business barely grew. Managed AISC is $1,751/oz — the highest of any senior producer (Agnico $1,339, Kinross $1,372, Newmont $1,609, Barrick $1,637); two-thirds of production sits in Africa (Tanzania, DRC, Guinea, Ghana, Egypt), the worst jurisdiction mix of the seniors; reserve life is only ~12 years; and FY2025’s entire volume growth was acquired, not mined — organic ounces were flat. What re-rated is the multiple, not the franchise: P/B (5.3×) and P/S (4.3×) are at their richest levels in the company’s recorded history, and EV/EBITDA (~6.6–7.5×) is at the top of its own six-year range versus a mid-cycle average near 5×. The market has quietly erased the cost-and-jurisdiction discount AngloGold should carry against Agnico and Newmont — precisely as spot gold has rolled over roughly a third from its parabolic January-2026 peak. You are being asked to pay a quality multiple for the lowest-quality asset base among the seniors, on peak-cycle margins, at the moment the commodity that produces those margins is correcting. The ~5.4% yield pays you to wait, and the balance sheet and the ongoing gold bull make it a poor short — but the asymmetry from here is unattractive. This is a name to own on weakness in the gold price, not to chase after a 3× re-rate.

Framing: a re-rated, levered-commodity / high-income name — not value (its Value loading is ~zero), not momentum (that loading has decayed to zero as the tape rolled over), not a compounder. It is a gold-price proxy with a dividend and a self-help halo, and it is priced as if the halo were a moat. Conviction: Medium. The single fact that flips me bullish: a durable, structural fall in AISC toward the North-American peer group (Arthur delivering ~$950/oz at scale, or Obuasi/Siguiri/Iduapriem breaking below $1,600) that would earn the re-rating. The single fact that flips me bearish: a sustained gold pullback toward $3,000/oz, which would compress the AISC margin far faster than the multiple would forgive and expose how little organic growth underpins the equity.

📈 Stock Price Action — Five-Year Event Map

Factual price history, not a recommendation. Price moves are Fact; attributed drivers are Interpretation. No price targets or support/resistance levels.

AngloGold’s ADS has round-tripped through an extraordinary gold cycle. From a five-year closing low of ~$10.75 (23 September 2022) the stock rose more than eleven-fold to an all-time closing high of $121.39 (27 February 2026) — intraday $125.48 (2 March 2026) — before correcting to $84.65 (2 July 2026). It now sits ~30% below its all-time high, inside a 52-week range of $42.33–$121.39, up ~95% over twelve months but essentially flat over the trailing six months and hovering right at its 200-day moving average. The overwhelming driver across the whole arc was the US-dollar gold price, which climbed from ~$1,800/oz in 2022 to a record of roughly $5,589/oz in late January 2026 before its own ~⅓ correction — AngloGold is a high-beta expression of that move (see Section 11 and the factor read).

Cycle narrative. (1) Through 2022 AngloGold was a cheap, high-cost, SA-listed miner few wanted to own while real rates rose. (2) The 2023 corporate re-domicile to a London plc with NYSE as the primary listing was designed to shed the “South-Africa discount” and broaden the shareholder base; it coincided with gold bottoming. (3–4) 2024 delivered the operational turnaround and the Centamin deal, but the stock still moved mainly with bullion. (5) 2025 was the payoff year: gold’s break above $3,000 dropped straight through AngloGold’s cost base into record cash flow, and the equity re-rated as well as re-earned. (6) The January-2026 gold blow-off carried AU to $121–125 alongside record annual results. (7) As gold corrected from its parabolic peak, the highest-beta senior gave back roughly a third — leaving it, today, up hugely year-on-year but consolidating well below its high. The move is a Fact; the attribution to the gold price and the specific events is Interpretation, cross-checked against earnings dates, the Centamin timeline, and the gold-price record.

1. Executive Summary

AngloGold Ashanti is the world’s third- or fourth-largest gold producer by output, mining 3.09 million attributable ounces in 2025 from eleven operations on four continents, roughly two-thirds of it in Africa. In September 2023 it re-domiciled from South Africa to a UK plc and moved its primary listing to the NYSE; in November 2024 it acquired Centamin plc in an all-stock deal, adding Egypt’s Tier-1 Sukari mine. FY2025 was a record year on almost every line — revenue $9.89bn (+71%), adjusted EBITDA $6.29bn (+129%), free cash flow $2.91bn (+204%), statutory EPS $5.19 — and the balance sheet flipped to a net-cash position (~$0.9bn adjusted) for the first time in the company’s modern history. The company declared $1.8bn of dividends (357¢/share), a ~5.4% yield and the highest in the senior peer group, under a policy that returns ~50% of free cash flow plus top-ups.

The uncomfortable truth beneath the records is that they are rented from the gold price, not earned from the business. Revenue’s +71% decomposes into a +45% rise in realized gold price ($3,468/oz vs $2,394) multiplied by +16% higher volume that was itself ~99% acquired (the first full year of Sukari); ex-Sukari organic production was flat. Unit costs did not improve — group AISC rose 6% to $1,709/oz, and managed AISC of $1,751/oz is the highest of any senior producer, against Agnico Eagle at $1,339, Kinross $1,372, Newmont $1,609 and Barrick $1,637. Gold mining is a structurally poor, price-taking, depletion-driven industry that destroys capital across the cycle; AngloGold is not an exception to that — it is a leveraged expression of the current up-leg, and among the seniors it is the highest-cost and most geopolitically exposed version of the bet.

What is genuinely better is the balance sheet, the capital discipline and the optionality. Management high-graded the portfolio (selling high-cost South African and Malian assets earlier, pruning Serra Grande, Córrego do Sítio, Doropo and ABC in 2025), executed its one large deal in stock at fair value with no goodwill bloat (avoiding the sector’s classic buy-at-the-top-write-it-off failure), and is funding a credible North-American growth option in the Arthur Gold Project (Nevada, 4.9 Moz initial reserve, ~$954/oz AISC, ~$3.6bn capex, first production early 2030s) that would materially upgrade both the cost curve and the jurisdiction mix if delivered. Earnings quality is high: statutory profit sits below headline earnings, cash conversion is ~1.5×, and there is no one-time gain propping up the print.

The investment tension is entirely in the valuation. The equity has re-rated to the richest P/B (5.3×, ~95th percentile of its own history) and P/S (4.3×, ~94th percentile) it has ever traded at, and EV/EBITDA (~6.6–7.5× depending on the EBITDA definition) sits at the top of its six-year range versus a ~5× mid-cycle average — even as spot gold has corrected ~⅓ from its January-2026 peak. The market has effectively removed the cost-and-jurisdiction discount AngloGold ought to trade at relative to Agnico and Newmont. For the current price to be right, gold must hold near record levels and the multiple must not mean-revert and Arthur must de-risk the franchise — a lot to underwrite on the sector’s weakest asset base. The dividend and the balance sheet make it defensible to hold and unattractive to short; the asymmetry favours accumulating on gold-price weakness rather than chasing strength. No recommendation or price target is made in this body; the labeled Claude’s Take above is the sole exception.

2. Business Overview

What the company does. AngloGold Ashanti is a pure-play primary gold producer: it explores for, develops and mines gold, and sells it into the spot market. By-product credits (silver, sulphuric acid at Sukari, minor base metals) are immaterial to the revenue line — revenue is, to a first approximation, ounces sold × realized gold price. There is no meaningful downstream, no hedge book of consequence (the company sells at spot; it does not run a structural forward-sale program), and no diversification into other commodities. The result is a business whose top line is almost entirely a function of two variables it does not control (the gold price) and partially controls (its own production volume and cost).

Scale and output. In FY2025 the group produced 3.091 Moz of attributable gold, up 16.5% year-on-year, and realized an average price of $3,468/oz (2024: $2,393). That combination lifted revenue from $5,793m to $9,893m (+71%). On a longer view, revenue was remarkably flat for years — $4,595m (2020), $4,029m (2021), $4,501m (2022), $4,582m (2023) — before the step-change of 2024–25 driven by the Centamin acquisition and the gold price. Production of ~3 Moz places AngloGold among the largest four Western-listed gold miners, below Newmont (~6 Moz) and roughly level with Agnico Eagle (~3.45 Moz) and Barrick (~3.26 Moz).

Asset portfolio (the eleven operations). The defining feature of AngloGold’s business is where the ounces come from. By 2025 attributable production the split is roughly 66% Africa, 17% Australia, 16% Americas:

The portfolio is genuinely diversified across eleven assets and four continents — but, as Section 3 argues, diversification of risk is not the same as a cost or quality advantage, and the geographic centre of gravity (Africa) is the highest-risk of the senior group. There is currently no producing US mine (Cripple Creek & Victor was sold in 2015); the United States is a development pipeline (Arthur / North Bullfrog, Nevada) only.

How it makes money, and the leakage. Because revenue is ounces × price, the economics are dominated by the spread between the realized price and AISC. In 2025 that spread was enormous: $3,468 − $1,709 = ~$1,759/oz of AISC margin, roughly half the realized price. But a material share of that margin does not accrue to AngloGold shareholders. Sukari is 50%-owned by the Egyptian state (via EMRA), Ghana and other host governments hold free-carried or profit-share interests, Cerro Vanguardia is 92.5%-held, and Kibali (45%) is equity-accounted rather than consolidated. Non-controlling interests took $538m of group profit (17%) in 2025, host-government payments totalled $2.66bn, and a Sukari profit-share (MRMIA) took a further ~$239m. The headline “group” free-cash-flow and production figures therefore overstate what is attributable to the AngloGold equity — attributable FCF was ~$2.6bn ex-Kibali versus the ~$2.9bn headline. This is a recurring theme throughout this analysis: read AngloGold on an attributable basis, not a headline one.

Recurring vs. non-recurring revenue. None of it is contractually recurring in the SaaS sense; but gold demand is perpetual and the product is fungible and instantly saleable, so “revenue” is better thought of as durable but entirely price-exposed. There is no customer concentration risk (gold is sold into a deep global market), no receivables risk of consequence, and no product obsolescence — the risk is entirely price, cost, volume and jurisdiction.

Verdict. A large, diversified, pure-play gold producer whose revenue is a direct function of the gold price and whose attributable economics are meaningfully diluted by host-government and minority interests. Well-run and clearly-structured, but with no feature of the business model itself that confers pricing power or durable advantage — that question belongs to Section 3.

3. Industry Dynamics

Structure: a textbook bad industry. Gold mining exhibits nearly every attribute of a structurally unattractive industry. The product is a globally fungible commodity priced in a deep, liquid market; no producer, however large, has any pricing power — AngloGold is a price-taker in the purest sense. Costs are subject to industry-wide inflation in labour, energy, reagents, steel and (for African operators) diesel and security. The asset base depletes: every ounce mined must be replaced by exploration or acquisition or the company liquidates itself, and replacement is getting harder and more expensive as the industry’s discovery rate falls and grades decline. Capital intensity is high and lumpy (a single project like Arthur is ~$3.6bn), lead times run a decade, and the geological and political risk of the best remaining orebodies is rising. Against Greenwald’s framework there is no demand-side captivity (no switching costs, no habit, no search costs — gold is gold), no supply/cost advantage that is proprietary (cost position is a function of orebody geology and jurisdiction, not a defensible process), and no economies of scale with customer captivity (scale in mining brings diversification and some procurement leverage, but no share-driven cost moat).

The through-cycle record is damning. Over 2006–2020 the US-dollar gold price roughly tripled, yet the gold-mining equity indices lagged bullion — the industry consistently converted higher prices into higher costs, dilutive acquisitions and write-offs rather than shareholder returns. Published gold-sector valuation primers (a UBS North American Gold Producers primer and related metals-and-mining primers, framework context only) makes the same structural point: gold equities are a “trading sector” that structurally underperform the metal, valuation is done on P/NAV (net asset value at a ~5% real discount rate) rather than P/E because reported earnings are distorted by heavy depreciation and impairments, and a producer’s multiple premium must be earned by demonstrably low cost, long life, low political risk and an unhedged book. This is the lens through which AngloGold must be judged: not “is gold going up?” but “does this company convert the up-cycle into durable per-share value better than its peers and better than simply owning bullion or a royalty company?”

The capital cycle (Marathon lens). The sector currently sits mid-to-late in a favourable capital cycle. Record margins have produced record free cash flow across the seniors, a wave of M&A (an estimated ~$90bn of gold-sector deals in the current cycle, including AngloGold/Centamin), rising exploration budgets and the first serious greenfield commitments in years (Arthur among them). Marathon’s warning is precisely about this phase: high returns attract capital, capital builds supply and buys assets at elevated prices, and returns mean-revert. The discipline of this cycle is better than the disastrous 2010–2012 peak (miners are paying dividends and tendering debt rather than levering into mega-deals), but the direction of travel — more capital chasing a finite set of orebodies at record gold prices — is the classic setup for future disappointment. An investor buying senior gold equities at the top of their own valuation ranges (as AngloGold now trades) is buying late in the capital cycle.

Demand drivers. The current gold bull is monetary, not industrial: central-bank buying has run at ~860–1,045 tonnes/year (led by emerging-market reserve diversification away from the dollar), reinforced by ETF inflows, geopolitical hedging and, in the January-2026 blow-off, momentum. This demand is real but reflexive and sentiment-driven — it can reverse quickly (as the ~⅓ correction from the January peak shows). Jewelry and technology demand are price-elastic and have been crowded out at these levels. The key point for a miner’s valuation: the margin windfall rests on a gold price that is itself extended and volatile.

Regulation and jurisdiction. Mining is intensely regulated at the point of the orebody — permitting, royalties, export rules, local-content and beneficiation requirements, resource-nationalism and profit-share regimes vary by country and change with governments. AngloGold’s footprint concentrates this risk: Ghana (royalty renegotiation ongoing), Tanzania (a history of tax disputes in the sector), DRC (Kibali — chronic political and security risk), Guinea (military government), Egypt (state 50% partner at Sukari), Argentina (capital controls at Cerro Vanguardia). This is materially worse than Agnico’s ~95% Canada/Finland/Australia base.

Verdict: structurally BAD industry, currently in a good moment. Gold mining destroys capital across the full cycle, has no demand moat and a relentless depletion treadmill; it is a price-taker’s game where the metal usually beats the miners. The present moment is unusually favourable — record prices, disciplined capital returns — but that is cyclical, not structural, and the capital cycle is maturing. AngloGold operates in the worst-positioned corner of this industry (highest cost, highest jurisdiction risk) even as it enjoys the up-leg.

4. Competitive Position

The central question: is there a moat? For a gold miner the only advantages that show up in financial outcomes are (i) a low-cost, long-life, low-risk asset base (a cost-curve position that survives the down-cycle), and (ii) superior capital discipline (replacing reserves and allocating cash better than peers). AngloGold has genuinely improved on the second and is the weakest of the seniors on the first.

Cost position — the decisive disadvantage. AngloGold’s managed AISC of $1,751/oz (group $1,709/oz) is the highest of any senior gold producer. The comparison is stark:

Within AngloGold’s own portfolio the dispersion is wide and unflattering: only Sukari (~$1,094) and Kibali (~$1,317) are clearly low-cost, while four material mines — Siguiri, Iduapriem, Sunrise Dam and Obuasi — run at or above ~$2,000/oz. At a $3,468 realized price everything is profitable and the distinction seems academic; at $2,500 gold, a quarter of the portfolio approaches break-even on an all-in basis, and at $2,000 gold much of the African book is uneconomic. A high-cost producer is more levered to the gold price on the way up and more fragile on the way down — the opposite of a moat. AngloGold’s cost position is a competitive disadvantage, and it did not narrow in 2025 (AISC rose 6%; 2026 guidance is higher still at $1,780–1,990 group).

Jurisdiction — the second disadvantage. ~66% of production in Africa (Tanzania, DRC, Guinea, Ghana, Egypt) is the highest political-risk mix of the seniors — the mirror image of Agnico’s Tier-1 concentration. Resource nationalism, royalty renegotiation (Ghana), security (DRC, Guinea/Sahel), profit-share (Egypt) and capital controls (Argentina) are live, recurring risks that both raise the effective cost of capital and cap the multiple. The P/NAV framework in the published sector primers treats jurisdiction risk as an explicit discount to NAV; AngloGold should carry the widest such discount among the seniors.

Reserve life and grade. Attributable reserves rose to 36.5 Moz (from 31.2 Moz), but at 3.09 Moz/year that is only ~11.8 years of reserve life — shorter than Agnico (~16 years) and far shorter than Newmont on a reserve-plus-resource basis. Much of the reserve growth came from the initial 4.9 Moz booking at Arthur, a decade-out development asset, rather than from replacement at the producing mines. Short reserve life on a depleting, high-cost base is the least durable position of the group.

Capital discipline — the one real relative strength. Where AngloGold does differentiate is management quality and capital allocation (see Section 6): the portfolio high-grading, the all-stock and fair-value Centamin deal (no goodwill bloat), the balance-sheet repair to net cash, the sector-leading dividend, and disciplined project selection. This is a better operator and allocator than its own history and than several peers — but “runs the business well” is not a moat; it is table stakes that can be competed away and does not survive a change of management or a cost shock.

Greenwald tests. Market-share stability: fails — AngloGold’s share of global production has drifted with disposals and acquisitions, not held via customer captivity (there is none). ROIC test: FY2025 ROIC of ~27% looks moat-like but is entirely gold-price-flattered — the same asset base produced a GAAP loss (−$235m) as recently as FY2023 and single-digit ROIC in 2021–22. A “moat” that evaporates when the commodity price falls is not a moat. Advantage type: none of Greenwald’s three (proprietary supply/cost, demand captivity, scale-with-captivity) applies.

Verdict: NO durable competitive advantage — and, among the seniors, the weakest asset base. AngloGold is a high-cost, Africa-weighted price-taker with a short reserve life. Its genuine strengths — balance sheet, capital discipline, the Arthur option, the income yield — are real and differentiating within the peer set, but they are management and financial attributes, not a business moat. The company is a well-run bad-industry business, and the market is currently pricing it closer to a good-industry one.

5. Growth History and Forward Opportunities

Historical growth was flat until it was bought. Revenue sat in a $4.0–4.6bn band from 2020 through 2023 before the gold price and Centamin lifted it to $5.8bn (2024) and $9.9bn (2025). But the volume story is the tell. FY2025’s +430koz of production was ~99% acquired: Sukari contributed ~500koz versus ~40koz in the prior stub period, while the rest of the portfolio was essentially flat — ex-Sukari production was ~2,288koz versus ~2,312koz the year before, a slight decline. In other words, the organic mining base is not growing; it is holding roughly steady while individual mines rise and fall (Obuasi and Tropicana up in 2025; Iduapriem −16%, Sunrise Dam −10%, Serra Grande −34%). This is normal for a mature senior — mines deplete and must be sustained — but it means the company’s growth engine is M&A and one large greenfield option, not the existing asset base.

Organic vs. acquired, and quality. The Centamin/Sukari acquisition looks, so far, like a good piece of bought growth: Sukari is a genuine Tier-1, long-life, relatively low-cost (~$1,094 AISC) asset, it was bought in stock at fair value (no goodwill), and management says it recouped roughly a third of the net purchase price in cash in year one. That is high-quality acquired growth by mining standards — the exception that flatters the rule that gold M&A destroys value. Against it, the organic base is low-quality growth: flat volumes at rising unit cost, with the top line moving on price.

The forward pipeline.

• Arthur Gold Project (Nevada) — the one strategic story. Formed from the North Bullfrog / Silicon complex in Nevada, Arthur carries an initial 4.9 Moz Probable reserve, guided AISC of ~$954/oz, ~$3.6bn of capex, targeted >20% returns, and first production in the early 2030s. If delivered, Arthur is transformational for the quality of AngloGold — it adds a large, low-cost, Tier-1-jurisdiction asset in the United States, directly attacking the company’s two structural weaknesses (cost and jurisdiction). It is the plank of Calderón’s explicit campaign to “narrow the rating gap to North American peers.” But it is long-dated, capital-intensive, unproven on cost at scale, and subject to Nevada permitting — optionality, not a certainty, and not a support for today’s multiple.
• Obuasi — the redeveloped Ghanaian legacy mine is ramping but remains high-cost (~$2,000/oz); further improvement is a swing factor for group AISC.
• Sukari / brownfield — continued optimization and exploration upside at the newly-acquired flagship.
• Portfolio simplification — the disposals of Serra Grande, Córrego do Sítio, Doropo and ABC (2025) are negative volume but positive quality: shedding high-cost or non-core ounces to lift the average.

Reserve replacement. The 31.2 → 36.5 Moz reserve increase is encouraging but Arthur-dependent; the producing mines are, at best, replacing depletion. Sustained organic reserve growth at the operating assets is the missing ingredient and an open question (Section 13).

Verdict: LOW-to-MEDIUM quality growth. The organic asset base is flat-to-declining at rising cost; the good growth (Sukari) was bought, and the transformational growth (Arthur) is a decade out and unproven. AngloGold’s near-term “growth” is overwhelmingly a gold-price phenomenon, not a volume or unit-economics one. That is acceptable for a mature senior, but it should not be mistaken for a compounding franchise, and it does not justify a growth multiple.

6. Financial Quality

Margin structure and its driver. The margin expansion is spectacular and almost entirely exogenous. Gross margin ran 22.4% → 35.7% → 49.2% across FY2023–25; operating margin reached 43.2%; IFRS EBITDA margin 56.2% ($5,556m) and the company’s adjusted EBITDA margin ~64% ($6,294m). The FY2024→FY2025 revenue increase of +71% decomposes cleanly into price received +45% × production +16%, with the volume itself ~99% acquired. Crucially, AISC rose 6% over the same period — so the margin expansion is not operating leverage from falling unit costs; it is the gold price outrunning cost inflation. Every incremental dollar of gold price drops almost straight to margin because production and cost were roughly held — which is exactly why the margin, and the equity, would deflate on a gold-price fall.

Cash generation and conversion. This is the genuine bright spot. Operating cash flow was $4,784m against total net income (pre-NCI) of $3,174m — a ~1.5× cash-conversion ratio with no accrual games. Capex of ~$1,449m (capex/sales ~15%, moderate for a producer) left company-defined free cash flow of $2,908m (+204%), a record. The cash is real and largely un-manufactured; the caveat is attribution — after non-controlling interests and host-government profit-shares, attributable FCF is lower (~$2.6bn ex-Kibali). Free-cash-flow yield on the current ~$42bn market cap is ~7% on the headline figure, less on an attributable basis.

Quality of earnings — clean. The single most important QoE finding: there is no one-time gain inflating FY2025. Statutory profit attributable was $2,636m / $5.19 EPS, which is lower than headline earnings of $2,724m / 536¢ — the reverse of a company flattering its print. The gap is net one-time charges carried in statutory earnings (Quebradona impairment $98m, a Serra Grande reversal, disposal losses ~$48m), which the SA/JSE headline measure strips out. The Centamin deal generated only $102m of goodwill, confirming it was struck at ~fair value with no bargain-purchase gain to reverse or intangible to impair later. (For context, FY2023’s −$235m loss was itself driven by non-recurring items — a ~$314m UK-restructuring/re-domicile cost and ~$221m of impairments — not by operating weakness.) Earnings are operational and cash-backed; the risk is that they are cyclical-peak, not that they are fake.

Returns on capital — real but rented. FY2025 ROIC (~27%), ROE (~30%+) and ROA (18.6%) are excellent on their face. But the same asset base generated negative returns in 2023 and high-single-digit ROIC in 2021–22. These returns are a function of a $3,468 gold price against a ~$1,750 cost base; they are gold-price-inflated and non-durable, and should be normalized (toward, say, a $3,000 gold price and rising costs) before any valuation conclusion. This is the crux of the embedded-expectations analysis in Section 10.

Balance sheet — a fortress, newly built. AngloGold ended 2025 with cash of $2,905m against total debt of $2,281m — a net-cash position of ~$838m (company “adjusted net cash” $879m) — a dramatic swing from net debt of ~$1.3bn in 2023 and ~$0.6bn in 2024, funded by the FCF windfall. Liquidity is ~$4.4bn and no debt matures until 2028. Equity attributable to owners is $8,091m (plus $1,825m NCI), giving a book value of ~$16/share. The large rehabilitation and environmental provisions typical of a miner (asset-retirement obligations for mine closure) sit in non-current liabilities and are a real, long-dated call on future cash — worth monitoring but well-covered here. The balance sheet is now among the strongest in the sector and removes financing risk from the thesis entirely.

A note on P/B. The reason the P/B optically screens at the ~95th percentile of AngloGold’s own history (5.3×) is partly an artifact: years of impairments drove retained earnings deeply negative (−$2,148m in 2023, −$1,316m in 2024), and the equity base only turned positive in 2025 (+$76m retained earnings). A small, impairment-rebuilt equity denominator against a re-rated market cap inflates P/B; it is a weaker valuation signal here than P/S or EV/EBITDA (which tell a consistent “richest-in-a-decade” story without the denominator distortion).

Verdict: high-quality, cash-backed earnings and a fortress balance sheet — but the economics are cyclical-peak and gold-price-driven, not durably scale-driven. The business does not get structurally cheaper to run as it grows (AISC rose); it simply earns more when gold is high. The financial condition is excellent; the financial quality of the current run-rate is real but not repeatable if gold falls.

7. Capital Allocation

The most important thing management got right: they didn’t blow it. Gold mining’s defining capital-allocation failure is buying assets at the top of the cycle with debt or over-valued paper and writing them off in the next downturn. AngloGold, under CEO Alberto Calderón (since September 2021), has so far avoided that trap and, by the standards of the sector, allocated capital above average.

M&A — Centamin, done right. The one large deal — the ~$2.5bn acquisition of Centamin plc (Sukari), closed November 2024 — was executed in all stock (~+20% share issuance) at a price that generated only $102m of goodwill, i.e., struck at approximately fair value. It added a genuine Tier-1, relatively low-cost, long-life asset, and management reports it recouped ~⅓ of the net purchase price in cash in the first year at prevailing gold prices. Using stock (rather than debt) preserved the balance sheet and shared cyclical risk with Centamin holders. This is close to a model gold acquisition and stands in favourable contrast to the sector’s history. The ~20% dilution is real and must be remembered when reading per-share growth — but it bought a quality asset, not a write-off in waiting.

Portfolio high-grading. Over Calderón’s tenure AngloGold sold higher-cost and non-core assets (South African deep-level mines and Mali earlier; Serra Grande, Córrego do Sítio, Doropo and the ABC project pruned in 2025), lifting the portfolio’s average quality and simplifying the story. This is disciplined subtraction — rarer and harder than acquisitive growth.

Returns of capital. The dividend framework returns ~50% of annual free cash flow as a base plus top-ups; FY2025’s declared dividends of $1.8bn (357¢/share) represented ~62% of free cash flow and a ~5.4% yield — the highest of the senior group and roughly 5× the ~1% yields at Newmont and Agnico. This is the single most attractive shareholder-return feature of the stock and a deliberate use of the windfall. Notably, AngloGold has not launched a buyback despite reaching net cash — management has said repurchases are “part of the book” but chose deleveraging (the “Project Moet” bond tender, a ~$650m cap on the 2028/2030/2040 notes — debt reduction, not equity repurchase) and dividends instead. Given the equity trades at the richest multiples in its history, not buying back stock at these levels is arguably correct capital allocation — but the absence of any repurchase capacity deployed at lower prices in prior years is a modest missed opportunity.

Capex and the pipeline. Sustaining capex is disciplined (~15% of sales). The large forward call is Arthur (~$3.6bn over the build), which management frames as a >20% return at ~$954/oz AISC. The capital-allocation question for the next several years is whether AngloGold funds Arthur without straining the dividend or the balance sheet, and whether it resists the temptation (classic late-cycle behaviour) to over-pay for further M&A with its now-richly-valued stock.

Incentives. Executive LTI is tied to relative TSR + AISC + production + ESG metrics. This is decent — relative TSR and AISC align management with cost discipline and beating peers — but it is not ideal: it lacks an explicit ROIC or per-share-FCF hurdle, and rewarding production (ounces) carries a mild empire-building tilt (grow ounces even if per-share value doesn’t grow). It is better than a pure-growth scheme and worse than a returns-on-capital one.

Insider behaviour. The SEC Form 3/4 record is clean but uninspiring: routine RSU grants and small director sales, and — notably — no discretionary open-market (code-P) purchases even at record gold prices and record cash flow. The batch of 18 Form 3s filed in March 2026 is an initial-ownership-statement artifact tied to index inclusion, not a signal. The absence of insider buying at the top is unsurprising but means insiders provide no positive conviction signal here; nor is there any red-flag selling.

Verdict: above-sector-average capital allocation — a genuine relative strength. Management high-graded the portfolio, executed its big deal in stock at fair value, repaired the balance sheet to net cash, and returns cash generously via the sector’s best dividend. The gaps — no ROIC-linked incentive, a production tilt in comp, no buyback flexibility deployed at lower prices, and the ever-present late-cycle M&A temptation — keep this from being a top-decile allocator, but it is clearly a bright spot in the thesis and the strongest argument for the equity beyond the gold price itself.

8. Changes and Headwinds — Last Two Years

Strategic and structural changes.

• Re-domicile and re-listing (September 2023). AngloGold moved from a South African to a UK plc structure with the NYSE as primary listing (JSE/A2X/Ghana secondary). The explicit aim was to shed the “South-Africa discount,” widen the US/global investor base and improve index eligibility. Judged by the subsequent re-rating (and index inclusion in 2026), it succeeded — though the gold price makes attribution hard. A one-time ~$314m restructuring cost hit FY2023.
• Centamin / Sukari acquisition (announced Sept-2024, closed Nov-2024). The largest change to the asset base in years: a new Tier-1 flagship, ~+20% shares, and the first full-year contribution in FY2025 (the engine of the +16% volume).
• Portfolio simplification (2025). Disposals/pruning of Serra Grande, Córrego do Sítio, Doropo, ABC and Côte d’Ivoire assets.
• Arthur Gold Project (Nevada) — maiden reserve (2025/26). A 4.9 Moz initial Probable reserve booked; the flagship growth/quality option.
• Balance-sheet transformation. Net debt → net cash within two years on the FCF windfall; the “Project Moet” bond tender (Q2-2026) further reduced gross debt.
• Capital returns. A step-up to the sector-leading dividend ($1.8bn declared for 2025) under the 50%-of-FCF-plus-top-ups policy; index inclusion in 2026 broadened the holder base.

Headwinds and pressures.

• Cost inflation. AISC rose 6% in 2025 and is guided higher in 2026 ($1,780–1,990 group; $1,825–2,050 managed) — labour, energy, reagents and African diesel/security costs are not abating, and the company’s “flat in real terms” framing masks nominal cost creep.
• Gold-price rollover. Spot gold corrected ~⅓ from its ~$5,589 January-2026 peak; the equity fell ~30% from its high in sympathy. The single biggest swing factor is now moving against the 2025 tailwind.
• Jurisdiction developments. Ongoing Ghana royalty renegotiation (unquantified), the standing risks in DRC (Kibali security), Guinea (military government), Egypt (state profit-share) and Argentina (capital controls).
• NCI / host-government leakage. The Sukari 50% state ownership and other minority/profit-share structures mean a growing share of “group” cash flow does not reach the AngloGold equity ($538m NCI in 2025).
• Dilution. The ~20% share issuance for Centamin permanently raised the share count (431m → ~508m weighted); per-share metrics must be read against it.

Verdict: the last two years STRENGTHENED the balance sheet, the portfolio quality and the capital-return profile — genuine, durable improvements — but the thesis’s dominant driver (the gold price) has turned from tailwind to potential headwind, and the structural cost/jurisdiction disadvantages are unchanged and, on cost, worsening. The company is better; the moment is later.

9. Risk Analysis (Risk Matrix)

Chance of a total loss: very low. AngloGold is a diversified, net-cash, cash-generative senior producer; no single event short of a simultaneous gold collapse and multi-asset expropriation threatens the equity’s existence. The realistic downside is a large drawdown (a repeat of the 2013-style ~50–60% peak-to-trough as gold and the multiple both fall), not a permanent capital loss. The balance sheet effectively removes financing/solvency risk.

Net risk read: the risks are dominated by two correlated exposures — the gold price and African cost/jurisdiction — and the equity is positioned at the high-beta, high-cost, high-jurisdiction-risk end of the senior spectrum while trading at the rich end of its valuation history. That is an unfavourable risk configuration for a new buyer at today’s price, and a manageable one for a holder paid ~5.4% to wait.

10. Valuation Discussion (Embedded Expectations)

No price target and no recommendation in this section. The analysis is of embedded expectations and scenarios only.

Where the multiples sit. At $84.65 (2 July 2026), ~505m shares give a market capitalization of ~$42–43bn and, against ~$0.8bn net cash, an enterprise value of ~$42bn. On that basis:

The consistent message across the cycle-immune metrics (EV/Sales, EV/EBITDA) is that AngloGold trades at the richest levels of its own history — the P/E only looks “mid-range” (~45th percentile per AZI) because the earnings denominator is at a cyclical peak inflated by record gold. This is the classic peak-cycle miner signature: cheap on peak earnings, expensive on everything the price can’t flatter.

Embedded-expectations / reverse read. What must be true for ~$42bn EV to be fair value (say, a ~10% required return with no multiple re-rating)? Broadly, AngloGold must sustain something close to its FY2025 ~$2.9bn free-cash-flow run-rate (a ~7% headline FCF yield, less attributable) and the market must keep paying ~6.6–7.5× EBITDA. That in turn requires gold to hold near record levels ($3,400–4,000+/oz realized) against a rising ~$1,800+ AISC — because at a $3,000 realized price the AISC margin compresses from ~$1,760 to ~$1,200/oz (a ~30% margin hit before any volume change), and at $2,500 gold it roughly halves. The equity is therefore underwriting a sustained, near-record gold price on the highest-cost asset base in the group, at the top of its own valuation range.

Scenario analysis (illustrative; gold-price-driven). Because AngloGold is a levered gold proxy, the honest way to frame value is a gold-price grid. Holding production ~flat at ~3 Moz and AISC ~$1,800/oz, and applying a mid-cycle ~5–6× EV/EBITDA (below today’s ~6.6–7.5×, reflecting the structural discount AngloGold should carry):

• Bear — gold reverts to ~$2,800/oz, multiple normalizes to ~5×. AISC margin ~$1,000/oz → attributable EBITDA roughly halves toward ~$3.0–3.5bn; EV ~$16–18bn; equity ~$32–36bn, ~$60–70/share (roughly −20% to −30%). Dividend cut as the 50%-of-FCF policy self-adjusts. This is a plausible mean-reversion, not a crash.
• Base — gold holds ~$3,300–3,600/oz, multiple ~5.5–6×. AISC margin ~$1,500–1,800/oz; attributable EBITDA ~$4.5–5.0bn; EV ~$28–32bn; equity ~$36–42bn, roughly $72–88/share (≈ current to modestly below). The ~5.4% dividend is maintained; the case that the stock is fairly priced for a strong-but-not-record gold environment.
• Bull — gold sustains ≥$4,000/oz (or re-accelerates), multiple holds ~6.5–7×, Arthur de-risks. AISC margin ≥$2,000/oz; attributable EBITDA ~$5.5–6bn+; EV ~$40–48bn; equity ~$50–56bn, ~$100–110+/share (+20–30%), consistent with the Street’s higher targets (Citi $130, Roth $110). Requires the gold bull to resume and no multiple compression and execution.

The distribution is asymmetric to the downside from the current price: the base case sits at roughly today’s level, the bull requires sustained record gold plus no de-rating, and the bear (an ordinary gold pullback) already implies a meaningful drawdown that the ~5.4% yield only partly cushions.

Cross-check vs. peers and the alternative expressions. On EV/EBITDA AngloGold (~6.6–7.5×) trades at a discount to Agnico (~7–10×), roughly in line with Newmont (~5–6×) and above Barrick (~4.2×) — a ranking that is defensible (Agnico’s Tier-1 quality earns its premium; Barrick is cheap for a reason) but which, given AngloGold’s highest cost and worst jurisdiction, arguably still leaves it too close to the middle of the pack rather than at the discount its fundamentals warrant. The royalty/streaming model (Franco-Nevada ~25× EV/EBITDA, Wheaton ~27×) is the structurally superior way to own gold beta (no cost inflation, ~85–88% margins) and is the contrast case, not a comp — the market pays 4–5× AngloGold’s multiple for the absence of exactly the cost and jurisdiction risks AngloGold carries most heavily. On the P/NAV framework favoured by the published sector primers, AngloGold at these prices is likely trading around or above 1.0–1.3× NAV on a normalized gold deck — full, not cheap, for a high-cost operator.

Valuation verdict (no target): the equity embeds a sustained, near-record gold price and no multiple mean-reversion on the sector’s weakest asset base. That is a demanding set of expectations; the risk/reward is balanced-to-unfavourable for a buyer at today’s price and adequate for a holder collecting the sector-best yield. The gold price, not the business, is the variable that determines the outcome.

11. Variant Perception

Consensus view. The sell-side is constructive: AngloGold is seen as a turnaround that worked — self-help under Calderón, a fixed balance sheet, a value-accretive Centamin deal, sector-best income, and a Nevada growth option — deserving of a re-rating toward its North-American peers. Price targets cluster above spot (Citi Buy, $130; Roth Buy, $110), and the stock’s index inclusion and ~95% twelve-month return reinforce the “quality re-rate” narrative. Consensus implicitly assumes gold stays strong and the discount to AEM/NEM continues to narrow.

The factor/positioning read (what the tape is pricing). The FactorsToday model is unusually clear: ~78% of AngloGold’s return variance is explained by factors, overwhelmingly the gold complex — loadings of GoldPrice ~+2.8 and Gold-Miners-industry ~+2.0, with the classic gold inverse to the US dollar (−0.30 to −0.48) and interest rates (−0.28). The only meaningful style tilt is DividendYield (+0.31); it is not Value (~0), not Momentum (~0, having decayed as the tape rolled over), and negative LowVol (−0.34). Equity-market beta is low (0.877) with high positive alpha (+0.455) — i.e., it moves with gold, not the stock market. Its factor-nearest peers are precisely the senior-gold set (AEM, AGI, IAG, the gold-miner ETFs), confirming it trades as an interchangeable gold-beta vehicle with a dividend. The risk-adjusted record underlines the industry point: strong recent Sharpes (y1 ~1.4, y3 ~1.1) sit atop a lifetime Sharpe of ~0.04 and a max drawdown of ~89% — the last few years are the gold-bull exception to a poor long-run history. Today the stock is consolidating ~⅓ below its peak, right at its 200-day average — not a falling knife (still up ~95% y/y, positive alpha), but not trending either. The positioning evidence says: this is a levered commodity bet that has rolled over, dressed as a re-rating story.

The strongest bull case. Gold’s monetary bull market is structural (central-bank de-dollarization, fiscal/geopolitical hedging) and has years to run; at ≥$4,000/oz sustained, AngloGold’s high-cost base becomes a feature (maximum operating leverage), throwing off ~$3bn+ of FCF, funding a rising ~5.4%+ dividend and the Arthur build simultaneously. Arthur then delivers a large, low-cost, US asset that earns the peer re-rating, closing the gap to Agnico/Newmont permanently. In this world the current multiple is not rich — it is early — and the stock compounds with the metal plus the quality upgrade.

The strongest bear case. Gold has already peaked (the January-2026 blow-off to ~$5,589 was the top) and mean-reverts toward $3,000 as real rates, a firmer dollar or simply exhausted momentum reassert. AngloGold’s highest-in-class AISC (rising, not falling) compresses the margin faster than peers; the ~5.4% dividend — a percentage of falling FCF — is cut; and the richest-ever multiple (P/S ~94th, EV/Sales ~94th percentile) mean-reverts toward the ~5× historical EV/EBITDA. Cost-and-jurisdiction reality reasserts the discount the market just erased. The high-cost, Africa-heavy senior gives back the re-rate and the earnings — a 2013-style double compression. Arthur is a decade out and cannot help.

The 3–5 assumptions that matter most.

1. The gold price path (the master variable — everything else is second order).
2. AISC trajectory — does cost inflation keep outrunning “flat in real terms,” or does Obuasi/Sukari optimization and Arthur eventually pull the group cost down?
3. The multiple — does the market keep paying a top-of-range, near-peer multiple for the highest-cost senior, or does the cost/jurisdiction discount return?
4. Arthur execution — on-time, on-budget, at ~$954/oz, or another decade-long capex sink?
5. Jurisdiction shocks — Ghana royalties, DRC/Guinea security, Egypt/Argentina — any of which can impair a chunk of the book.

What would falsify each side. Falsifies the bull: a sustained gold price below ~$3,000/oz, or AISC breaking above ~$2,000/oz group, compressing FCF and forcing a dividend cut. Falsifies the bear: group AISC durably falling toward ~$1,500/oz (Arthur/optimization delivering) and organic reserve replacement turning positive — i.e., the business earning the quality re-rate rather than renting it from gold.

Variant conclusion. Consensus is pricing AngloGold as a quality re-rate; the factor and cost evidence say it is fundamentally a levered, high-cost gold-beta vehicle with a great dividend whose discount has been prematurely erased. The variant perception is that the market is paying a franchise multiple for a rented margin — the risk is not that AngloGold is a bad company (it is a well-run one) but that it is a good operator of a bad-industry, high-cost asset base priced as if it were a good-industry one, late in both the gold and capital cycles.

12. Fact vs. Interpretation Table

13. Open Questions

1. What is a normalized gold price to underwrite? The entire valuation hinges on it, and it is unknowable; the honest answer is a distribution, not a point (Section 10). Where in the $2,800–4,500 range should a disciplined buyer anchor?
2. Can group AISC actually fall, or only rise “less”? 2026 guidance is higher. Is there a credible path (Obuasi ramp, Sukari optimization, high-cost disposals, then Arthur) to a structurally lower group cost — the single thing that would earn the re-rating?
3. How large and permanent is the NCI/host-government leakage, and does it grow? Precise Sukari attributable economics after the EMRA 50% and MRMIA profit-share; the outcome of the Ghana royalty renegotiation.
4. Is AngloGold genuinely unhedged? Management’s framing implies spot exposure (no structural forward-sales), but this should be confirmed from the 20-F — it is central to the “levered gold bet” thesis.
5. Arthur — real economics at scale? ~$954/oz AISC and ~$3.6bn capex are guidance on a decade-out Nevada project; the sector’s history of greenfield cost/time overruns is poor. What is the realistic risked value today?
6. Will management resist late-cycle M&A? With record cash and a richly-valued stock, will AngloGold use paper for another deal — and would it be Centamin-quality or a top-of-cycle mistake?
7. Buyback? Management calls repurchases “part of the book” but has done none. At what price/leverage would they buy, and is the dividend-only policy optimal?
8. Reserve replacement at the producing mines (ex-Arthur) — is organic reserve life stabilizing or shrinking?

14. What Must Be True

For the BULL case to be right (own it and win from here):

• Gold sustains ≥ ~$3,600/oz (ideally re-accelerates toward/above $4,000), keeping the AISC margin near record and FCF at ~$3bn+.
• AISC is contained (group holds ≤ ~$1,900/oz) so cost inflation does not eat the price, and the Obuasi/Sukari/optimization program plus, eventually, Arthur begin to pull the group cost curve down.
• Arthur de-risks on schedule and budget, delivering the promised low-cost US ounces and earning a durable re-rate toward Agnico/Newmont.
• The multiple holds at a top-of-range ~6.5–7× EBITDA (the market keeps treating AngloGold as a quality re-rate, not a high-cost senior).
• The sector-best ~5.4% dividend is maintained or grown, paying holders to compound with the metal.

Bull falsification test: a sustained gold price below ~$3,000/oz, or group AISC breaking above ~$2,000/oz, compressing free cash flow and forcing a dividend cut — any of which breaks the “windfall funds returns and growth” engine.

For the BEAR case to be right (avoid/underweight from here):

• Gold mean-reverts toward ~$2,800–3,000/oz (the January-2026 ~$5,589 print was the cycle top), halving the AISC margin on the highest-cost senior.
• Cost inflation persists (AISC guided higher already), so the margin compresses faster than at lower-cost peers.
• The richest-ever multiple mean-reverts toward the ~5× historical EV/EBITDA as the cost/jurisdiction discount the market erased returns.
• A jurisdiction shock (Ghana royalties, DRC/Guinea security, Egypt/Argentina) impairs a chunk of the attributable book.
• The dividend is cut (a percentage of falling FCF), removing the yield support and the last leg of the bull narrative.

Bear falsification test: group AISC durably falling toward ~$1,500/oz (Arthur/optimization delivering) and organic reserve replacement turning positive — i.e., AngloGold structurally earning the quality re-rate rather than renting it from the gold price. If cost and reserve life genuinely improve, the “rented margin” critique fails.

The synthesis: both cases hinge on the same two variables — the gold price and AISC. The bull needs gold high and cost contained; the bear needs gold lower and cost rising (the current trajectory). Because the stock is priced near the top of its own range on the sector’s weakest asset base, the burden of proof sits with the bull — the base case is roughly “fairly priced for a strong gold environment,” with the asymmetry tilted to the downside on any normalization.

15. Source Appendix

(Full detail in the separate AU_source_appendix.md. Primary sources first.)

Primary — company filings (SEC EDGAR, CIK 0001973832):

• AngloGold Ashanti plc Form 20-F for FY2025 (filed 26 March 2026) — financial statements, ore reserves, cost disclosures, risk factors, remuneration.
• Form 20-F FY2024 and FY2023 — multi-year trend, Centamin purchase accounting, FY23 impairment/restructuring.
• Form 6-K, Q4/FY2025 Earnings Release and Operating Statement (20 February 2026) — production, AISC, price received, FCF, adjusted EBITDA, headline-vs-statutory earnings, dividend.
• Form 6-K, Q1 2026 Operating/Earnings (8 May 2026); Arthur Gold Project 6-K (26 March 2026); Project Moet tender-offer 6-Ks (30 March / 14 April 2026); Technical Report Summaries (26 March 2026).
• Form 3 / Form 4 insider filings (2023–2026); Form SD (conflict minerals).

Primary — quantitative data:

• ROIC.ai (income statement, balance sheet, cash flow, profitability/valuation ratios, enterprise value) — reconciled to the 20-F; used for multi-year trend and multiples.
• AZI valuation percentiles (own-history P/E / P/B / P/S) and 5-year daily price/OHLCV history.
• FactorsToday factor model (loadings, leaderboard, related stocks) — positioning read.

Secondary — industry / peer context:

• Public peer disclosures for the senior gold producers: Newmont (NEM), Agnico Eagle (AEM), Franco-Nevada (FNV), Wheaton (WPM) — peer AISC, multiples, jurisdiction and capital-return comparisons.
• Published sell-side gold-sector primers (e.g., UBS North American Gold Producers primer) — framework/value-chain and P/NAV context only (dated; not current data).
• Gold-price history and central-bank/ETF demand data (public sources); sell-side price-target references (Citi, Roth).

Facts are reconciled to the 20-F where they drive a verdict; interpretations are labeled as such; management commentary is treated as hypothesis and validated against filings and external data.

APPENDIX A — Standard Diligence Questionnaire — AngloGold Ashanti plc (NYSE: AU)

Supplemental to the article. Answers are grounded in the primary filings; Fact / Interpretation / Assumption labels applied where material. Figures FY2025 unless noted.

General

What thoughtful questions have other investors asked about this company? The recurring institutional questions are: (1) Has the re-domicile/re-listing actually closed the “South-Africa discount,” or did gold do all the work? (2) Is the Centamin deal the exception or the start of a late-cycle M&A spree? (3) Can AngloGold’s group AISC — the highest of the seniors — ever structurally fall, or is it a permanent discount driver? (4) How much of “group” cash flow actually reaches the equity after the Sukari 50% state interest, Ghana profit-shares and the Kibali 45% equity-accounting? (5) Is the ~5.4% dividend safe if gold falls? (6) Does Arthur (Nevada) justify a re-rating toward Agnico/Newmont, or is it a decade-out capex risk? These map directly to the article’s cost, jurisdiction, leakage, and valuation sections.

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low? (Fact/Interpretation) A cyclical high. FY2025 EPS of $5.19 rests on a realized gold price of $3,468/oz (+45% y/y) against a ~$1,750 cost base; the same asset base produced a GAAP loss (−$235m) in FY2023. Returns (ROIC ~27%) are gold-price-inflated and non-durable.

Driven by the external environment or internal actions? Overwhelmingly external (the gold price) plus one large internal action (the Centamin acquisition supplying the volume). Organic mining performance was roughly flat. Cost discipline (“flat in real terms”) is a genuine internal effort but AISC still rose 6% nominally.

How stable are revenues? Volume is relatively stable (~3 Moz, mature diversified base); revenue is highly unstable because it is ounces × a volatile gold price. Revenue sat in a $4.0–4.6bn band for years, then doubled in two years on price + M&A.

Outlook for products/services? Gold demand is perpetual and monetary; the product never obsolesces. The price outlook is the swing — currently correcting ~⅓ off a January-2026 record.

How big is this market — growing, shrinking, domestic or international? The gold market is deep, global and effectively infinite for any single producer; AngloGold is a price-taker into it. Mine supply grows slowly (depletion, falling discovery); demand is currently monetary/central-bank-led. AngloGold’s operations are international (Africa/Australia/Americas), sales in USD.

Business Quality & Competitive Moat

Is the industry getting more or less competitive? Structurally competitive and consolidating (a ~$90bn M&A wave). No producer gains pricing power; competition is on the cost curve and for scarce quality orebodies.

How profitable is the business (ROIC, ROE)? (Fact, caveated) FY2025 ROIC ~27%, ROE ~30%+, ROA 18.6% — excellent but gold-price-flattered; normalize toward mid-cycle before drawing conclusions. Through-cycle returns are poor (negative as recently as 2023).

How profitable is the industry — how many competitors, barriers to entry? A structurally poor industry that destroys capital across the cycle; gold equities historically lag bullion. Barriers to entry are geological/permitting (finding and permitting an orebody), not competitive moats. Handful of seniors (Newmont, Agnico, Barrick, AngloGold, Kinross, Gold Fields) plus mid-tiers.

Can the business be easily understood? Yes — ounces × price − AISC, across eleven mines. The complexity is in jurisdiction, NCI/host-government leakage, and reserve accounting.

Can it be undermined by foreign low-cost labour? Not applicable in the offshoring sense; but cost competition is geological/jurisdictional — AngloGold is the high-cost producer, so it is the one most undermined by lower-cost peers on the cost curve.

Do brands matter? No. Gold is fungible; there is no brand premium.

Nature of competition; customers’ switching costs? No customer switching costs (gold is sold into a market, not to captive customers). Competition is for assets and on cost — AngloGold is disadvantaged on the latter.

Financial Condition & Balance Sheet

Assets not fully recognized on the balance sheet? (Interpretation) Yes — mineral reserves/resources and exploration optionality (notably Arthur’s early-stage value) are carried at cost, not fair value; a rising gold price adds economic value not reflected in book. Conversely, book equity is small and impairment-scarred (see P/B note).

Off-balance-sheet liabilities? The large item is mine-closure / rehabilitation (asset-retirement) provisions — long-dated environmental obligations sitting in non-current liabilities; real and material for any miner, well-covered here. Host-government profit-share arrangements (EMRA/Sukari) are contractual claims on cash flow.

How conservative is the accounting? (Fact) Conservative on the current print — statutory earnings ($2,636m) sit below headline earnings ($2,724m); the Centamin deal booked only $102m goodwill (no bargain-purchase gain); cash conversion ~1.5×. AngloGold has historically taken impairments (2022–23) rather than deferred them.

How CapEx-hungry is the business? Moderately — sustaining capex ~15% of sales — but with a large lumpy forward call: Arthur ~$3.6bn. Mining is inherently capital-intensive and depletes, requiring continuous reinvestment just to hold output.

Capital Allocation & Management

How much FCF, and how is it used? FY2025 FCF $2,908m (record). Used for: dividends (~50% of FCF + top-ups; $1.8bn declared), debt reduction (Project Moet bond tender), and reinvestment (sustaining capex + Arthur). No buyback.

Significant acquisitions recently? Yes — Centamin (Sukari), ~$2.5bn all-stock, closed Nov-2024 — well-priced (no goodwill bloat), a genuine Tier-1 addition. The key capital-allocation event of the period.

Buying back shares? No — despite reaching net cash. Management calls buybacks “part of the book” but has chosen dividends + deleveraging. Defensible given the richest-ever multiple, but no repurchase capacity was deployed at lower prices historically.

Issuing large amounts of stock to insiders? No red flag — routine RSU grants; the large issuance was the ~20% for Centamin (an acquisition, not insider enrichment).

Compensation policy of directors/management? (Interpretation) LTI tied to relative TSR + AISC + production + ESG — decent alignment on cost and beating peers, but lacks an explicit ROIC / per-share-FCF hurdle and carries a mild empire-building tilt via the production metric.

Motivations of management? CEO Calderón’s stated campaign is to “narrow the rating gap to North American peers” via balance-sheet repair, portfolio high-grading and the Arthur pipeline — coherent and, so far, well-executed. No discretionary insider buying provides no positive conviction signal.

Valuation & Market Data

Is the stock an ADR, MLP, or K-1 issuer? (Fact) AngloGold is a UK plc whose NYSE-listed security trades as ordinary shares/ADSs; it files as a foreign private issuer (Form 20-F), reports under IFRS, and pays dividends that may carry foreign-withholding considerations for US holders. Not an MLP or K-1 issuer. US investors should confirm ADR/withholding mechanics with their custodian.

Dividend policy? ~50% of annual free cash flow as a base, plus top-ups; FY2025 declared $1.8bn (357¢), ~5.4% yield — the sector’s highest. The policy self-adjusts downward if FCF falls (the yield is not a fixed commitment).

How profitable is the business? See ROIC/ROE above — very profitable at current gold prices, cyclically.

Is net income diverging from cash from operations? (Fact) No adverse divergence — OCF ($4,784m) exceeds pre-NCI net income ($3,174m) by ~1.5×; earnings are cash-backed. The relevant divergence is attributable vs headline cash flow (NCI/host-government leakage), not accrual quality.

Risks & Downside

What would cause the stock to decline? A gold-price decline (the master risk); AISC inflation outrunning price; a multiple de-rating from richest-ever levels; a jurisdiction shock (Ghana/DRC/Guinea/Egypt/Argentina); a dividend cut; an Arthur overrun; a late-cycle M&A misstep.

Risk of a catastrophic loss? Low-probability/high-severity tails exist (tailings-dam failure, fatal safety event, asset expropriation) but are not base-case. The realistic downside is a large drawdown on gold + multiple compression (2013-style), not a permanent impairment.

Chance of a total loss? Very low — diversified, net-cash, cash-generative senior producer; no single event short of simultaneous gold collapse and multi-asset expropriation threatens the equity’s existence.

Recent News & Events

Has the business environment changed recently? (Fact/Interpretation) Yes on two axes: (1) the gold price rolled over ~⅓ from its ~$5,589 January-2026 record, turning the dominant tailwind into a potential headwind; (2) cost guidance moved higher for 2026 (AISC $1,780–1,990 group). Structurally, the balance sheet (net cash) and portfolio quality improved.

Significant acquisitions? Centamin/Sukari (Nov-2024) — covered above. Watch for further late-cycle M&A given record cash and a richly-valued stock.

Change in accounting policies? None material flagged; the re-domicile to a UK plc changed the reporting entity/structure (2023) but not the IFRS basis. A one-time ~$314m restructuring cost hit FY2023.

Recent changes — new markets, facilities, management? Re-domicile + NYSE primary listing (2023); Centamin/Sukari and index inclusion (2024–26); Arthur maiden reserve (2025/26); Project Moet bond tender (2026); ongoing portfolio simplification (Serra Grande, CdS, Doropo, ABC, Côte d’Ivoire disposals).

APPENDIX B — Source Appendix — AngloGold Ashanti plc (NYSE: AU)

Primary sources before secondary; recent before stale. Management commentary is treated as hypothesis and validated against filings and external data.

1. Company filings — SEC EDGAR (CIK 0001973832)

AngloGold Ashanti plc files as a foreign private issuer (Form 20-F annual, Form 6-K interim/events), reporting under IFRS.

• Form 20-F, FY2025 (filed 26 Mar 2026) — au-20251231.htm. Audited IFRS financial statements; ore reserves & resources; per-mine production and AISC; risk factors; remuneration; NCI/host-government arrangements. Primary source for financial statements, reserves, cost structure, jurisdiction.
• Form 20-F, FY2024 (filed 15 Apr 2025) and FY2023 (filed 25 Apr 2024) — multi-year trend; Centamin purchase accounting; FY2023 restructuring/impairment.
• Form 6-K — Q4/FY2025 Earnings Release & Operating Statement (20 Feb 2026) — agaq4dec2025earningsrelease.htm, agaq4december2025operating.htm. Production 3,091 koz; AISC $1,709/oz group; price received $3,468/oz; FCF $2,908m; adjusted EBITDA $6,294m; adjusted net cash $879m; dividends $1.8bn / 357¢; headline earnings $2,724m / 536¢ vs statutory $2,636m / 519¢. FY2026 guidance.
• Form 6-K — FY2025 earnings-call transcript / investor presentation (20 Feb 2026) — management commentary on capital returns, Sukari integration, Arthur, cost outlook (AngloGold reports semi-annually; no Q1-2026 call).
• Form 6-K — Arthur Gold Project (26 Mar 2026) — arthurgoldproject.htm; maiden 4.9 Moz reserve, ~$954/oz AISC, ~$3.6bn capex, first production early 2030s.
• Form 6-K — Project Moet tender-offer launch & pricing (30 Mar / 14 Apr 2026) — bond tender (~$650m cap on 2028/2030/2040 notes); deleveraging, not a buyback.
• Form 6-K — Technical Report Summaries (26 Mar 2026); annual suite of reports (26 Mar 2026).
• Form F-4 / Form 425 (2024) — Centamin plc acquisition (all-stock).
• Form 3 / Form 4 (2023–2026) — insider transactions; routine RSU grants and small director sales; no code-P open-market purchases; March-2026 Form 3 batch tied to index inclusion.
• Form SD (27 May 2026) — conflict-minerals disclosure.

2. Quantitative data sources (reconciled to filings)

• ROIC.ai — multi-year income statement, balance sheet, cash-flow statement; profitability/valuation ratios; enterprise value. Used for trend and multiples; reconciled to the 20-F (note: ROIC’s per-share book-value field was glitched and was not relied upon; a large “extraordinary item” field is a mapping artifact, not a real one-timer — see Section 6).
• AZI valuation percentiles — own-history P/E (45.6th), P/B (94.7th), P/S (93.6th) as of 2 Jul 2026; and 5-year daily price/OHLCV history (adjusted + unadjusted, EMAs, beta, alpha).
• FactorsToday factor model — stock loadings (GoldPrice +2.8, Gold-Miners +2.0, DividendYield +0.31; R² ~0.74–0.78), leaderboard (risk-adjusted returns/drawdowns), related stocks (AEM/AGI/IAG/gold ETFs). Positioning read only; third-party statistical estimates.

3. Peer / industry context (secondary)

• Public peer disclosures for the senior gold producers: Newmont (NEM), Agnico Eagle (AEM), Franco-Nevada (FNV), Wheaton Precious Metals (WPM) — peer AISC, EV/EBITDA, P/NAV, jurisdiction and capital-return comparisons.
• Published sell-side gold-sector primers — e.g., UBS North American Gold Producers primer; metals & mining / global mining primers. Framework, value-chain and P/NAV methodology context only (dated ~2008–2010; not current data).
• Gold-price history and demand data — public sources for the 5-year gold-price path (~$1,800/oz 2022 → ~$5,589/oz Jan-2026 record → correction), central-bank buying (~860–1,045 t/yr) and ETF flows; sell-side price-target references (Citi Buy $130; Roth Buy $110), public financial media.

4. Key figures quick-reference (FY2025 unless noted)