British American Tobacco p.l.c. (NYSE: BTI) — The Cheap One Finally Got Expensive, On Its Own Terms

Report date: 2026-06-27 · Price (2026-06-26): $62.76 · Market cap: ~$135B · Net debt: ~£31.0B (~$42B) · ADR = 1 ordinary share Primary listing: London (LSE: BATS, ~4,650p) · also Johannesburg (JSE: BTI) · Foreign private issuer, files 20-F/6-K, reports in GBP (£)

⚡ Claude’s Take

This block is the author’s own independent opinion and general information only — not investment advice and not a recommendation to buy or sell any security. The analysis that follows takes no position and carries no price target; the only view expressed in this article sits in this block.

Verdict: HOLD / accumulate only on weakness — a genuinely cheap, high-yield, low-volatility tobacco cash machine whose deep-value discount has now been largely arbitraged away. Not a buy at ~13.3x adjusted earnings and a ~5% yield; not a short either. Accumulate toward the low-$50s; this is “fairly valued, on its own terms,” not “still cheap.” Fair-value zone ≈ $55–68 (≈11.5–13.5x FY2026 adjusted EPS of ~$4.95–5.05, supporting a ~4.8–5.8% covered dividend yield, plus a few dollars of under-credited residual ITC-stake optionality). Back up the truck only sub-$52–54 (≈11x / ~6% yield, where it traded as recently as mid-2025); demand a margin of safety, don’t chase $66+.

British American Tobacco is the cheapest, highest-yielding, slowest-growing — and structurally weakest — of the three Western tobacco majors, and for two years that was exactly the trade: a ~9% covered yield and ~6–7x EV/EBITDA on a business everyone had left for dead after the December-2023 ~£27bn write-down of its US cigarette brands. That trade has worked spectacularly. BTI has ~2.6x’d off its 2023 low (~$24 → a $67 peak in May-2026), almost entirely a re-rating of a de-risking story: the US returned to revenue and profit growth, the open-ended Canadian litigation overhang was bounded into a known CAD$32.5bn settlement (implemented August-2025), the FDA menthol ban (a direct threat to Newport) was withdrawn, New Categories crossed into profitability (category contribution +77% to £427m), leverage was repaired to 2.48x, the buyback restarted (£1.3bn for 2026), and the ITC India stake was opportunistically monetized. Every major change of the last two years cut the same way — and the market re-rated BAT from ~6.7x to ~10.6x EV/EBITDA, its richest-ever own multiple (P/S and the composite valuation percentile both sit ~95th–99th of their own decade). The yield compressed from ~9% to ~5%. That compression is the return — and it is mostly spent.

The framing here is re-rated deep-value with genuine low-volatility / dividend-yield momentum (beta 0.20, +38% trailing twelve months, only ~6% off the high) — emphatically not a falling knife. But three things keep me at HOLD rather than chasing. First, the moat is the weakest of the majors: enterprise ROIC is only ~10.6% — barely above its ~7–8% cost of capital — because the 2017 Reynolds acquisition (~$49bn) capitalized the franchise’s value away to the sellers, a verdict the company itself ratified with the 2023 impairment. Second, the smoke-free transition is two-thirds disappointing (Vuse vapour revenue is declining; glo is a distant #3 to PM’s ~76%-share IQOS) and only one-third winning (Velo modern oral, +48%, now US #2). Third, the FY2026 5–8% adjusted-EPS algorithm is promised, not delivered — FY2025 grew just +3.4% constant-currency — and on a re-rated multiple, a growth wobble and a de-rate compound. Versus PM (~21x, the only grower) BAT is still demonstrably cheap and that gap is earned; versus Altria (~13x, US-only) it offers more growth optionality but lower returns and FX noise. You are paid ~5% to wait and the buyback shrinks the float, so the downside is cushioned — but the era of being paid a lot to wait is over.

Conviction: medium. The single piece of evidence that flips me bullish: durable, multi-year delivery of the 5–8% EPS algorithm — specifically Velo sustaining double-digit growth into clear US #2 and Vuse stabilizing on durable illicit-vape enforcement — which would re-rate BAT toward PM’s neighborhood. The single piece that flips me bearish: the FDA’s proposed maximum-nicotine (very-low-nicotine) product standard advancing toward a final rule, which would impair BAT’s US franchise — its largest profit pool and ~58% of group operating profit — more than any peer’s. Tag: “The cheap one finally got expensive — on its own terms.”

📈 Stock Price Action — Five-Year Event Map

Factual price history (ADR, USD). Price moves are FACT (AZI five-year daily series); attributed causes are INTERPRETATION. No price target, no recommendation here — the opportunity judgment lives in Claude’s Take above.

The arc. BTI made a full round-trip-and-then-some. From a five-year low of ~$23.67 (30-Nov-2021), it re-tested the bottom at ~$23.8 on the December-2023 brand write-down, ground sideways through 2024 to $33.62 (end-2024), then doubled-plus to $55.80 (end-2025) and a 60-month high of $67.30 (14-May-2026), settling at $62.76 on 2026-06-26 — ~6.6% off the high, 52-week range $44.45–$67.30. The defining feature: the entire 2024–2026 advance was a re-rating of a de-risking franchise, not a melt-up of a momentum stock — beta is 0.20 and the move was led by dividend-yield compression.

Cycle narrative. (1–2) The 2021–23 slide compounded a structural-decline discount with a rate-driven de-rate of a high-yield instrument. (3) The 6-Dec-2023 write-down is the cycle low and the thesis pivot — BAT conceded that its acquired US brands (Newport, Camel, Pall Mall, Natural American Spirit) no longer merited indefinite-life carrying values, reset the narrative to “smokeless,” and set the bar low enough to clear. (4–5) FY2023/H1-2024 prints stabilized confidence; the sector re-rated; the menthol-ban withdrawal removed a direct Newport tail and supercharged BAT specifically. (6) The big leg — a ~70% advance — priced the delivered inflection: US revenue and profit growth, New Categories turning profitable (+77% contribution), the Canada settlement implemented (overhang → finite drag), leverage falling, the buyback resumed. (7) The FY2025 print was good but already priced — guidance reaffirmed, not raised. (8) The stock has since drifted ~7% off the high.

1. Executive Summary

British American Tobacco is the world’s second-largest listed tobacco-and-nicotine company by value (behind Philip Morris International), selling combustible cigarettes and a growing “New Categories”/smokeless portfolio across ~175 markets. It is defined by a single structural fact that separates it from both peers: BAT is the only major that is both global and heavily US-cigarette-exposed. Its 2017 acquisition of the ~58% of Reynolds American it did not own (~$49bn) brought Newport (the #1 US menthol brand), Camel and Natural American Spirit in-house — and with them, the US now contributes ~45% of revenue and ~58% of adjusted operating profit, a heavier concentration in the fastest-declining major cigarette market than the “global diversification” narrative implies.

FY2025 headline economics: revenue £25,610m (−1.0% reported, +2.1% constant-currency), adjusted profit from operations £11,572m at a 44.0% adjusted operating margin, adjusted diluted EPS 352.1p (+3.4% cc, adjusted for Canada), and a dividend per share of 245.04p (+2.0%). It is one of the most cash-generative businesses in any index — ~83% gross margins, ~45% operating margins, and, normalized for one-off Canadian settlement cash timing, ~£8–9bn of free cash flow funding a ~5% covered dividend and a restarted buyback.

Four findings dominate this report.

First, the GAAP income statement is unusable and must be discarded for adjusted figures. Reported diluted EPS swung from −£6.45 (2023) to £1.38 (2024) to £3.49 (2025) — almost entirely the 2023 ~£27bn US-brand impairment and the 2024–25 Canadian litigation-settlement provision movements, not operating swings. On a clean adjusted basis, BAT has compounded earnings at a dull-but-positive low-single-digit rate. The honest caveat: the new ~£1.4bn/yr brand amortisation that adjusted EPS politely ignores is the accountants’ own confession that BAT’s core US asset is a depleting one with a finite (≤30-year) life.

Second, the moat is real but the weakest of the three majors, and it earns only its cost of capital. Operating margins (44%) are extraordinary, but enterprise ROIC is only ~10.6%, barely above a ~7–8% WACC — because the Reynolds price capitalized the franchise’s value away. Versus PM, BAT is structurally on the wrong side of nearly every growth vector: PM owns IQOS (~76% global heated-tobacco share) versus BAT’s distant-#3 glo; PM owns ZYN (~55–61% US pouch share) versus BAT’s #2 Velo; PM’s smoke-free is 41.5% of revenue at higher-than-combustible margins, versus BAT’s 18.2% dragged by a declining Vuse vapour business. The cheapness versus PM (~13x vs ~21x) is earned.

Third, capital allocation is a tale of two eras. The 2017 Reynolds deal was value-destructive at the enterprise level (ROIC ≈ WACC, ~£40bn+ peak debt, negative tangible equity, eventually written down ~£27bn). But every incremental decision since 2020 has been disciplined: ~£10bn of deleveraging into a comfortable investment-grade 2.48x, a covered progressive dividend, a buyback resumed only inside target leverage and executed at a trough multiple (genuinely accretive, not pro-cyclical), opportunistic ITC monetization recycled into debt and buybacks — and, decisively, the avoidance of the ~$16.5bn+ next-gen M&A catastrophe that befell Altria. The one blemish: there is no return-on-capital metric in the incentive scorecard, conspicuous given that Reynolds is the reason returns are mediocre.

Fourth, and decisively for the investment question, the deep-value discount has been arbitraged away. The re-rating from ~6.7x to ~10.6x EV/EBITDA — to the richest-ever percentile of BAT’s own decade — is the bulk of the realized return, and it leaves a thin margin of safety on a slow-growing, secularly-declining base. The consensus view (“cheap, high-yield, melting, but de-risking”) is broadly correct on the business; the variant question is whether, after a ~2.6x move, the price still compensates for the secular and regulatory risk and the FY2026 acceleration that is promised but not yet delivered. We conclude the easy money in that re-rating has been made. This memo takes no position and sets no price target outside the labeled block above.

2. Business Overview

What BAT is. British American Tobacco p.l.c. (founded 1902, headquartered in London; ~46,000 employees) manufactures and sells combustible cigarettes and smokeless/“New Category” nicotine products across roughly 175 markets. Its combustible brand stable includes Dunhill, Lucky Strike, Rothmans, Pall Mall, Kent, Newport, Camel, and Natural American Spirit; its New Categories are Vuse (vapour), glo (heated tobacco), and Velo (modern oral nicotine pouches), with Grizzly/Kodiak (US moist snuff) in traditional oral. [FACT — FY2025 20-F, SEC CIK 1303523, filed 2026-02-13; ROIC company profile]

Reporting structure — geography first. BAT reports three geographic segments. On a constant-currency basis (FY2025, total cc £26,414m):

[FACT — BAT FY2025 Preliminary Results, bat.com, 2026-02-12; Investegate RNS]

The single most important fact in the mix: the US is ~58% of adjusted operating profit — BAT is more US-cigarette-levered than its geographic footprint suggests, while the region that “should” be the long-run growth engine (APMEA: Indonesia, South Asia) actually shrank profit ~18% in 2025 on excise and regulatory shocks in Bangladesh and Australia. Geographic diversification, here, cuts both ways. [FACT/INTERPRETATION]

Reporting structure — product. By category:

• Combustibles (~75% of group revenue). Total cigarette volume fell −7.9% to ~465bn sticks in 2025; US combustibles volume −7.7% (vs US industry −7.4%). Crucially, US revenue and profit rose for the first time since 2022, which management attributes to commercial execution — though, as with Altria, the larger driver is the exogenous moderation in US cigarette decline as Federal/state enforcement bites into illicit disposable vapes. [FACT US-up; INTERPRETATION on cause]
• New Categories / smokeless (£3,621m revenue, +7.0% cc; category contribution £427m, +77.1%; margin 11.8%, +4.7pts). Three franchises with sharply divergent trajectories — Vuse (declining), glo (treading water), Velo (surging). Smokeless reached 18.2% of group revenue (+70bps) with 34.1m smokeless consumers (+4.7m). [FACT — BAT prelim]
• Traditional oral (Grizzly/Kodiak US snuff + Scandinavian snus) — folded into the smokeless tally.

The ITC stake — a non-operating asset being harvested. BAT historically held ~29% of ITC Ltd (the Indian cigarette/FMCG/hotels conglomerate). It has sold down to ~22.9% (March-2024 block ~£1,577m; May-2025 block ~£1,052m) and separately monetized the demerged ITC Hotels (sold 9% in December-2025 for ~£315m; retained 6.3%). The residual ~23% ITC Ltd stake is worth well over $15bn at market (~16% of BAT’s own market cap) — a large, liquid, non-core asset that is both a deleveraging lever and a genuine sum-of-the-parts input the market arguably under-credits. [FACT — SEC 6-Ks; bat.com press releases]

Recurring nature. Tobacco demand is the most recurring in consumer staples — addicted, daily, automatic repurchase — but BAT’s is a contracting annuity in combustibles (~−8%/yr volume) partially offset by an expanding one in smokeless (+4.7m consumers/yr). Revenue is ~85%+ recurring in character, sold through wholesalers and convenience retail.

§2 read. A global, diversified, 44%-operating-margin cash machine whose profit is ~58% US-cigarette-derived, whose largest “growth” region is shrinking, and whose smokeless transition — now profitable — is led by a category (Vuse vapour) in revenue decline. A harvest-with-a-side-of-transition profile, not a transition story.

3. Industry Dynamics

Structure — a tight, rational global oligopoly (ex-China). Outside the ring-fenced Chinese state monopoly (CNTC, ~44% of world cigarette volume), legal global nicotine is controlled by five players — Philip Morris International, British American Tobacco, Japan Tobacco, Imperial Brands, and Altria (US-only) — plus a handful of national monopolies. Consolidation is essentially complete (BAT/Reynolds 2017, PM/Swedish Match 2022, JT/Gallaher). There is no capacity arms race and, critically, no legal new entrant can build a cigarette brand: advertising bans (US since 1971; EU TPD; plain packaging in Australia/UK/France) plus the FDA’s PMTA premarket-authorization gate fossilize incumbent share. This is the textbook condition for supply discipline and pricing power. [FACT — industry structure]

Combustible volume decline — and BAT’s adverse exposure. US cigarette industry volume fell ~7.4% in 2025; BAT’s global cigarette volume fell −7.9%. This is meaningfully faster than PM’s international combustible decline (~−1.5%), for two structural reasons:

1. The US — BAT’s profit center — is the worst-declining major developed market, running −8% to −10% in recent years (affordability ceilings plus illicit-disposable-vape leakage), only moderating in 2025–26.
2. Emerging markets are now a source of negative surprises for BAT (Bangladesh/Australia excise and regulatory shocks cut APMEA profit −18%), not the stable ballast they are for PM.

BAT sits on the wrong side of the geographic decline gradient — overweight the fastest-declining developed market and exposed to volatile EM fiscal regimes. [FACT — BAT prelim; PM cross-read; INTERPRETATION]

The #1 near-term swing factor — illicit US disposable vapes and enforcement. By public estimate, ~70% of the US e-vapor category is illicit flavored Chinese disposables that evaded PMTA. These both cannibalize Vuse (Vuse vapour revenue −8.6%, volume −12.6% in 2025) and accelerate cigarette decline. The 2025–26 Federal/state enforcement crackdown (the ~$200M FDA allocation, Customs seizures, tariffs) is the single biggest reason US cigarette decline moderated and BAT’s US business “returned to growth.” This is exogenous and reversible — BAT is a beneficiary, not the cause. [FACT/INTERPRETATION]

Regulatory landscape (the tail risks).

• US menthol ban — withdrawn (Jan-2025). A disproportionately large positive for BAT, because Newport (menthol) is its flagship US brand; the withdrawal removed a near-term existential US risk no peer carried to the same degree. [FACT — OMB Unified Agenda]
• US maximum-nicotine (very-low-nicotine, VLN) product standard — proposed Jan-2025. The genuine catastrophic tail: a near-zero-nicotine cigarette mandate. Slow-moving (multi-year rulemaking + certain litigation) but it would impair BAT’s US franchise — its largest profit pool — more than any peer’s. [FACT — Federal Register 2025-00397]
• EU TPD/TED revision (~2028). Proposed EU-level taxation of HTUs, vapour and pouches — a structural overhang on the smoke-free engine across all majors; relevant to BAT’s European (AME) exposure. [FACT — PM report]
• EM excise shocks (Bangladesh, Australia’s record-high excise and resulting illicit market) — already a live drag on APMEA. [FACT]
• Canada (RBH/ITCAN settlement). The CAD$32.5bn industry settlement materially distorts BAT’s reported (not adjusted) profit; see §6/§8. [FACT]

Marathon capital-cycle read. Global tobacco is the archetypal late-stage capital-exit industry: no greenfield capacity, ESG-driven divestment, complete consolidation, asset-light incumbents harvesting an annuity with disciplined pricing on inelastic demand. On the supply side that is textbook bullish — rational players, no price war, real pricing power. But the frame breaks in two ways: (1) demand contracts faster than supply discipline can offset on BAT’s top line — US/global volume −8% overwhelms the usual reinvestment-driven mean-reversion (there is no reinvestment to reverse); and (2) the real “new entrants” are regulatory-arbitrage players — illicit disposables, synthetic-nicotine pouches — deploying capital outside the legal system the framework measures, directly attacking Vuse. The one place legitimate growth-capital flows is smoke-free (glo Hilo, Velo Plus, Vuse Ultra) — but BAT is a share-loser in two of three smoke-free categories, so its capital cycle there is unattractive. [INTERPRETATION — Marathon lens]

§3 Verdict: structurally a BAD-and-slowly-worsening industry, worse for BAT specifically than for PM. The rational-oligopoly pricing economics and formidable legal barriers to entry are intact, and the menthol-ban withdrawal is a real reprieve. But BAT carries the worst combination of the three majors: ~58% of profit in the fastest-declining major market, volatile EM fiscal exposure, a smoke-free leadership position (Vuse) in the one category in revenue decline and most exposed to illicit competition, and distant-#3 status in the most valuable smoke-free category (heated tobacco). A managed decline of the profit pool with a sub-scale, contested smoke-free offset — not a stable annuity, and not PM’s profit-pool-migration story.

4. Competitive Position

Name the moat (Greenwald taxonomy). BAT’s advantage is a stack of three genuine sources, each weaker than PM’s:

1. Intangible / brand + customer captivity (the dominant, real advantage). Decades-old brands with addicted, habitual repurchase — Dunhill, Lucky Strike, Rothmans internationally; Newport, Camel, Natural American Spirit in the US. Addiction is the deepest form of demand captivity, and it shows up financially in 44% operating margins. But the December-2023 ~£27bn write-down — reclassifying the US brands from indefinite-lived to a ≤30-year amortising life — is management’s own admission that these are depreciating assets with a finite terminal value, not perpetual franchises. That is a more honest (and more bearish) accounting stance than Altria has taken on Marlboro, and it quantifies the melting castle. [FACT — FY2023 results / impairment; INTERPRETATION]
2. Economies of scale + distribution/cost advantage. One of two global-scale manufacturers (with PM); fixed manufacturing, distribution, regulatory-science and PMTA-compliance costs spread over ~465bn sticks plus the smokeless book. Real — but PM has greater scale on the growth (smoke-free) side, where scale matters most prospectively. [INTERPRETATION]
3. Regulatory barrier to entry. Advertising bans + the PMTA gate freeze legal share in incumbents’ favor — a moat BAT shares with all majors. The flip side: the same wall is arbitraged by illicit disposables that ignore it, which is exactly what is hollowing out Vuse. [FACT/INTERPRETATION]
4. Switching costs (weak). Only glo has razor/razorblade device lock-in — and glo is a distant #3, so BAT captures little of the single most durable smoke-free advantage (which PM owns through IQOS). Cigarettes and pouches have near-zero switching costs beyond habit. [FACT]

Greenwald share-stability test — MIXED.

• US combustibles: PASS-ish — BAT held/gained US value share in 2025 (volume −7.7% vs industry −7.4%, with positive value-share momentum) — a defensible #2 US position. [FACT]
• Vuse vapour: FAIL on volume — volume −12.6%, revenue −8.6%; Vuse retained US value-share leadership (~51.7%) only because the legal category shrank around it as illicit disposables took unit share. Leading a shrinking legal pool is a weak pass. [FACT]
• glo heated: FAIL — distant #3, ~1% revenue growth, no share leadership anywhere; IQOS owns ~76% globally. [FACT]
• Velo modern oral: GAINING — to #2 US (~10–12% share), +48% revenue, profitable within a year of the Velo Plus relaunch. The one category where BAT is taking share. [FACT]

The ROIC test — the decisive weakness. FY2025 ROE reads ~28.5%, but that is flattered by the write-down-shrunken equity denominator. ROIC is only ~10.6% (2024 ~10.5%, 2022 ~9.2%) — roughly at the cost of capital, weighed down by the ~£49bn Reynolds goodwill/intangibles. [FACT — ROIC.ai, accessed 2026-06-27] This is the cleanest financial signal that BAT’s moat is mediocre at the enterprise level: the operating margins are extraordinary, but the capital deployed to acquire the US business destroyed the returns. Contrast PM (all-in ROIC ~16%, >40% ex-goodwill). A genuine moat that earns only WACC after the price paid for it is a moat whose value accrued largely to the seller (Reynolds shareholders), not BAT’s. [INTERPRETATION — Greenwald: high margins + WACC-level ROIC = advantage capitalized away by acquisition price]

Direct comparison vs PM and MO.

• vs PM: BAT loses on nearly every prospective dimension — PM is the only grower (positive volume, ~14% cc EPS), owns IQOS (~76% heated share) vs glo, ZYN (~55–61% US pouch share) vs Velo, and runs smoke-free at 41.5% of revenue at higher margins vs BAT’s 18.2% dragged by Vuse. PM’s ROIC is mid-teens-to->40%; BAT’s is ~10.6%. The market prices it: PM ~21x forward vs BAT ~13x. The gap is earned. [FACT — PM report]
• vs MO: Closer. BAT carries the same US combustible secular decline as Altria plus lower-margin EM exposure, minus Altria’s pure-play US pricing concentration. BAT’s edge: it actually won in modern oral (Velo #2 and growing) where MO’s on! is losing to ZYN, and it leads global vapour (Vuse) where MO’s NJOY failed and was impaired. BAT’s disadvantage: the ~£27bn write-down and ~10.6% ROIC are uglier than MO’s harvest economics; and BAT carries FX translation noise MO does not. [INTERPRETATION — MO/PM cross-reads]

§4 Verdict: a real but mediocre, weakly-differentiated moat — durable in the abstract Greenwald advantage sense (stable-to-rising US share, 44% margins, no legal entrant), but the cash-flow moat is the weakest of the three majors. BAT passes the margin and US-share-stability tests but fails the ROIC test (~10.6% ≈ WACC) — the clearest evidence the advantage was largely capitalized away by the Reynolds price. It is a distant #3 in the most valuable growth category, a declining leader in vapour, and a clear winner only in modern oral (Velo). The investable framing is deep-value-with-a-WACC-moat, not a quality compounder. The cheapness is the thesis; the moat is not.

5. Growth History and Forward Opportunities

History — flat top line, low-single-digit adjusted EPS. Revenue has gone essentially nowhere in nominal GBP: £25.78bn (2020) → £25.68 (2021) → £27.66 (2022) → £27.28 (2023) → £25.87 (2024) → £25.61 (2025). The 2022 bump was partly FX (a weak GBP inflating translated US revenue); the 2024–25 decline is partly FX reversing (revenue was −1.0% reported but +2.1% cc in 2025). Underneath, the algorithm is price/mix offsetting volume decline: combustible volume −7.9% offset by net pricing, plus New Categories +7% cc. Adjusted diluted EPS has compounded low-single-digit (~+3.4% cc in 2025), with buybacks and a falling share count (2,294m → 2,173m) a modest tailwind. [FACT — ROIC income statement; BAT prelim]

The decomposition that matters. BAT’s growth is ~entirely a value story, not a volume story:

• Combustibles: volume structurally −7 to −8%/yr; revenue roughly flat on ~+8–9% net pricing. A shrinking-but-pricing annuity.
• New Categories (+7% cc, £3,621m): Velo (+48%) is the engine; Vuse (−8.6%) is a drag; glo (+1%) is flat. Net +7% — decelerating from prior double-digit, and below the “low double-digit” 2026 ambition.
• US “return to growth” (+5.5% cc revenue, +5.9% profit): the swing factor — but, per §3, substantially enforcement-driven and therefore of uncertain durability. [OPEN QUESTION: is this structural execution or exogenous decline-moderation?]

Forward opportunities.

• Velo / modern oral — the best organic story BAT owns: US #2, profitable, +48%, in the fastest-growing, highest-margin smoke-free category. Sustained double-digit growth here is the single most credible path to the EPS algorithm. [FACT]
• glo Hilo (premium, faster-heating; launched Sept-2025; ~4.7% of the Japan market within four months) — the response to IQOS, unproven at scale, fighting a ~76%-share incumbent. [FACT — BAT prelim]
• Vuse stabilization on durable enforcement — optionality, not a plan; depends on an exogenous regulatory crackdown holding. [INTERPRETATION]
• US combustible decline-moderation persisting — the largest single profit lever, also the least controllable.
• ITC monetization — not operating growth, but a multi-billion-pound capital-return/deleveraging lever.

FY2026 guidance — the inflection is promised, not delivered. Management reaffirmed (12-Feb-2026) its medium-term algorithm — 3–5% cc revenue, 4–6% cc adjusted operating profit, 5–8% cc adjusted EPS — explicitly framed as deliverable “from 2026” (i.e., 2026 is the first year of the algorithm, not a year already banked). FY2025 landed top-of-its-own-guidance but cc EPS was only +3.4%, below the 5–8% targeted for 2026. The building blocks are real (US growth, New-Cat contribution +77%, Quantum cost savings, Velo momentum); the headline acceleration is forward-promised. [FACT — BAT prelim; INTERPRETATION]

§5 Verdict: LOW-quality, low-rate growth with one genuine bright spot. This is a value-over-volume harvest with a single credible organic growth engine (Velo) and a large pricing annuity in structural volume decline. The forward 5–8% EPS algorithm is plausible but unproven on a +3.4% trailing base; execution risk on the FY2026 step-up is the live debate. Growth is not the reason to own BAT — yield, valuation, and capital return are.

6. Financial Quality

Multi-year summary (GBP, £m unless noted):

[FACT — ROIC.ai (accessed 2026-06-27); BAT FY2025 prelim & 20-F. 2023 operating/EPS lines are impairment-distorted GAAP.]

The one-sentence read. BAT is a fortress cash machine wearing five years of GAAP noise — the headline EPS swing (−£6.45 → £1.38 → £3.49) is almost entirely the 2023 US-brand write-down and the 2024–25 Canadian settlement, not operating deterioration.

The adjusted-EPS bridge. FY2025 reported diluted EPS = 349.1p; adjusted diluted EPS = 352.1p — now only ~3p apart (versus 138p reported vs ~350p adjusted in 2024). Total adjusting items fell to £1,575m (2025) from £9,154m (2024). The 2025 components: ~£1,584m acquired-intangible (US brand) amortisation [the recurring drag]; a net Canada credit (~+£524m: a £708m provision cut less a £184m goodwill impairment); plus restructuring, smaller impairments (Malaysia), and ITC divestment gains/losses (run through adjusting items, so they fund the dividend/buyback but never flatter “adjusted” EPS — directionally conservative, but cherry-pick-able). The add-back list is fully itemised and consistently applied — a green flag for QoE discipline. [FACT — 20-F notes 1/4/24]

The 2023 ~£27bn impairment and its permanent tail. Confirmed: £23,232m trademark amortisation+impairment + £4,614m goodwill impairment in 2023, driving the −£14,367m net loss. Effective 1-Jan-2024 BAT reclassified its previously indefinite-lived US combustible trademarks to a finite ≤30-year life, creating “an expected increase in amortisation expense of £1.4 billion per annum” (20-F note 12). Only Grizzly remains indefinite-lived. QoE flag (AMBER): this ~£1.4–1.6bn/yr amortisation is non-cash and legitimately excluded from adjusted EPS — but it is management’s own codified admission that the core US asset is depleting. We use cash earnings as the denominator, but we do not treat this amortisation as pure accounting noise; it carries a real structural message. [FACT/INTERPRETATION]

The 2024 Canada charge and the 2025 FCF drop are the same event. In 2024 BAT booked a £6,203m charge for the Canadian CCAA settlement (industry total CAD$32.5bn / ~£17.6bn; implemented 29-Aug-2025; BAT funds it via an upfront payment = ITCAN’s cash on hand less a CAD$750m holdback, plus annual payments initially ~85% of ITCAN net income, declining over 18 years). The headline FCF optics: operating cash flow fell £10,125m → £6,342m (−37%), cash conversion to ~63%. But neither year is clean — in 2024 ITCAN was building restricted cash (a ~+£6.0bn working-capital inflow that flattered 2024 OCF); in H2-2025 that cash was paid out into the settlement trust (a ~−£3.9bn outflow that depressed 2025 OCF). BAT’s own LTI plan shows 100.6% cash conversion over the 2023–25 three-year period; normalized ex-Canada FCF is ~£8–9bn. [FACT — 20-F note 24] QoE flag (AMBER): analysts using the reported 63% conversion / ~0.8x single-year dividend cover at face value will misread a one-off cash-timing artifact as a trend.

Balance sheet. Net debt down five years running (£41.7bn 2020 → £31.0bn 2025); 2.48x adjusted net debt/EBITDA, inside the 2.0–2.5x target; floating:fixed 14:86 (low refinancing sensitivity); ratings Moody’s Baa1 / S&P BBB+ / Fitch BBB+, all stable. Intangibles total ~£86.9bn (goodwill £38.9bn + other £48.0bn), exceeding total equity — so tangible book value is deeply negative (~−£17.7/share). This is a benign acquisition-accounting artifact of the 2017 Reynolds deal (identical in character to PM and MO), not distress: the brands generate the cash; the negative TBV is the bookkeeping shadow of goodwill. Discard P/B and ROE; use ROIC and EV/EBITDA. [FACT/INTERPRETATION]

Margins and returns. Gross margin rock-stable at ~83–84%; adjusted operating margin 44.0%; ROIC ~10.6% against a ~7–8% WACC — a modest but real positive spread. Economics are stable, not improving: the combustible base shrinks ~5–8%/yr in volume, offset by price/mix and New Categories, with incremental capital earning ~10–11% (above WACC but unexceptional). [FACT — ROIC.ai]

§6 Verdict: HIGH financial quality on a cash basis, MODERATE on GAAP optics. A genuinely high-quality cash-return vehicle masquerading, on the GAAP income statement, as a volatile mess. ~83% gross / ~45% adjusted operating margins, ~£8–9bn normalized FCF, ROIC comfortably above WACC, five straight years of deleveraging into target, investment-grade and stable. The honest deductions: the engine is a slowly-shrinking combustible base (economics stable, not improving); the ~£1.4bn/yr brand amortisation is a real structural signal, not pure noise; and the 2025 headline FCF/dividend-cover is temporarily ugly on Canada cash timing. Anchor valuation on adjusted diluted EPS (352.1p), ROIC (~10.6%), and normalized ex-Canada FCF (~£8–9bn) — and discard the GAAP volatility.

7. Capital Allocation

Verdict preview: a tale of two eras — one large mistake (Reynolds) being worked off by a disciplined incremental regime.

The Reynolds post-mortem — the central capital-allocation event. In July 2017 BAT paid ~$49bn for the ~58% of Reynolds American it did not own, funded with debt and stock. It is the root of essentially every balance-sheet and return characteristic that defines BAT today: ~£86.9bn of intangibles, persistent negative tangible equity, and — the killer datapoint — enterprise ROIC pinned at ~8–11% across the entire post-Reynolds period, never escaping its cost of capital. The combustible cash flows are real and enormous, but the value was largely capitalized away by the price paid — the Greenwald/Marathon signature of a high-quality asset bought at a price that transferred the surplus to the sellers. The 2023 ~£27bn impairment is the receipt. [FACT/INTERPRETATION — ROIC.ai; FY2023 results]

Dividend. A progressive policy with a long uninterrupted record: 207.6p (2020) → 245.04p (2025), ~+2%/yr lately; cash dividends ~£5,238m in 2025; payout ~67% of adjusted EPS (deliberately reduced from the high-70s% to fund deleveraging). The ADR is quarterly and FX-translated (USD holders bear GBP/USD risk); yield ~5%. On adjusted EPS the dividend is well covered; on normalized ex-Canada FCF (~£8–9bn) it is ~1.6–1.7x covered. [FACT — BAT prelim; ROIC cash flow]

Buyback — the rare value-accretive restart. Buybacks were suspended for years post-Reynolds (deleveraging priority), then restarted in 2025 (£1,173m), with £1.3bn announced for 2026 — the 2025 program topped up £200m specifically from May-2025 ITC proceeds. At ~13x adjusted earnings / ~7.7% FCF yield, repurchases are accretive (every £1.3bn retires ~1.4% of the ~£92bn market cap at a ~9%+ earnings yield, well above the ~5% after-tax cost of the debt being rolled). The sequencing is correct — debt before buyback before M&A — and the timing is genuinely counter-cyclical (buying at a trough multiple), a favorable contrast to the pro-cyclical buyback pattern common across the sector. The constraint is scale, not logic: at ~1.4% of market cap, the per-share uplift is ~1.5%/yr. [FACT/INTERPRETATION]

ITC monetization. Opportunistic harvesting of a non-core, non-controlling financial stake at fair prices (Mar-2024 ~£1,577m; May-2025 ~£1,052m; ITC Hotels Dec-2025 ~£315m), recycled into deleveraging and accretive buybacks rather than left as dead capital. The residual ~23% ITC Ltd stake (>$15bn, ~16% of BAT’s market cap) is a large unmonetized reserve and a real SOTP input. [FACT — bat.com press releases]

M&A discipline since Reynolds — abstinence as the right call. BAT made no large next-gen acquisition; Vuse, glo and Velo were all built organically and New Categories reached profitability in 2025. The decisive peer contrast: Altria destroyed ~$16.5bn+ chasing next-gen via M&A (JUUL ~$12.8bn, Cronos, NJOY), and PM levered up ~$16bn for Swedish Match (a winner, ZYN, but still being deleveraged). BAT made the one giant mistake (Reynolds) and has since avoided the second (next-gen M&A) — disciplined precisely because Reynolds taught the lesson. [FACT/INTERPRETATION — MO/PM cross-reads]

Incentives — the conspicuous gap. The LTIP/STIP scorecard rewards adjusted EPS growth, New Categories net revenue, operating cash-flow conversion (>95% target), and 3-year relative TSR (plus a 10% climate metric). There is no explicit ROCE/ROIC return-on-capital metric — a real alignment weakness given that Reynolds left group ROIC at ~WACC: management is paid on EPS and TSR, neither of which penalizes earning only your cost of capital on an inflated capital base. Mitigant: shareholder support is overwhelming (2025 remuneration policy passed ~98%), and the cash-conversion and relative-TSR measures tie pay to cash discipline and outperformance. [FACT — 20-F remuneration report; AGM results]

Insider read (SEC sweep). As a UK foreign private issuer, BAT directors/PDMRs disclose dealings via LSE RNS, not SEC Form 4 — so US insider data is structurally thin. The lone post-2024 Form 4 (filed 2026-05-29) is a BAT subsidiary purchasing a Charlotte’s Web convertible debenture, not an officer/director buying BAT stock. The 6-K stream is dominated by routine institutional holdings notifications and daily buyback-execution RNS. No discretionary open-market purchase of BAT stock by a named insider surfaced — a neutral-to-mild-negative read (no conviction buying), consistent with the broader coverage pattern. [FACT — EDGAR CIK 1303523]

§7 Verdict: mixed but improving — and better than its closest peer on the dimension that destroys the most value. The 2017 Reynolds deal was value-destructive at the enterprise level and is held accountable by no incentive metric; but every incremental decision since 2020 — deleveraging, a covered progressive dividend, a counter-cyclical accretive buyback, opportunistic ITC monetization, and the avoidance of catastrophic next-gen M&A — has been disciplined and shareholder-aligned. The current regime allocates capital intelligently; the franchise is still digesting the one large mistake that preceded it.

8. Changes and Headwinds — Last Two Years

Major changes (all net de-risking).

• Dec-2023 strategic reset — “Building a Smokeless World.” CEO Tadeu Marroco (since May-2023) took the ~£27bn US-brand impairment, switched those intangibles to finite-life, and reset the narrative around a 50%-smokeless-revenue-by-2035 ambition. The cathartic low that removed the “indefinite US franchise” fiction and set a beatable bar. Strengthens.
• US return to growth (FY2025). US revenue and profit grew for the first time since 2022; Velo Plus posted triple-digit growth to US #2. The source of the 2023 crisis became the source of the recovery — durability is the open question. Strengthens.
• New Categories inflection. Category contribution +77.1% to £427m, margin 11.8%; smokeless now 18.2% of revenue. Crossed from cash-drain to profit contributor. Strengthens.
• Canada CCAA settlement implemented (29-Aug-2025). Converted an open-ended litigation overhang into a known, finite 18-year cash drag — a de-risking even though it costs cash; the source of the +157% reported-EPS optical swing. Strengthens (certainty), modest ongoing FCF headwind.
• ITC monetization (Mar-2024, May-2025, ITC Hotels Dec-2025) — orderly de-gearing using a non-core stake. Strengthens.
• Regulatory tailwind — menthol ban withdrawn (21-Jan-2025); US illicit-vape enforcement crackdown. Strengthens.
• Capital-return restart + Quantum cost program (~£2bn savings 2024–26). Strengthens.

§8 Verdict: NET STRENGTHEN — but the strengthening is now priced. Every major change cut toward de-risking; the thesis is structurally better than two years ago. The catch: the stock has ~2.6x’d off the 2023 low, so the relevant forward question is no longer “is BAT broken?” (answered: no) but “does the 5–8% EPS algorithm hold from a +3.4% base?” The reset is genuine; the cheapness has largely been arbitraged away.

Headwinds.

• US cigarette secular volume decline (~−8%/yr) — the structural core, offset only by net pricing; durability of the US “growth” is the #1 open question. [High likelihood / High impact]
• FDA maximum-nicotine (VLN) proposed rule — existential-if-finalized US tail; not enacted, litigation-laden. [Low-Med / High]
• Illicit disposable vapes hurting Vuse — enforcement is the partial, reversible offset. [High / Med]
• glo losing to IQOS in heated tobacco. [High / Med]
• EM excise/regulatory shocks (Bangladesh, Australia) — already cutting APMEA. [Med / Med]
• EU TPD/TED revision on smokeless/flavours (~2028). [Med / Med]
• FX translation (GBP reporter; USD/EM revenue) — FY25 revenue −1.0% reported vs +2.1% cc is pure translation drag, and ADR holders bear it on price and dividend. [High / Med]
• Canada annual-payment cash drag — known, finite, but a recurring FCF deduction. [High / Low-Med]

9. Risk Analysis (Risk Matrix)

Catastrophic-loss read. A total loss is implausible on any near/medium horizon — BAT is diversified across ~175 markets, investment-grade, generates ~£8–9bn of FCF, and pays a covered ~5% dividend. The realistic downside is a de-rate plus a growth miss compounding (risks 3+4), not impairment of the going concern; the genuine long-tail existential risk is regulatory (risk 2), and it is slow-moving.

10. Valuation Discussion (Embedded Expectations)

Where it trades. At $62.76 (ADR = 1 ordinary share; LSE ~4,650p; GBP/USD ~1.345):

• ~13.3x FY2025 adjusted diluted EPS (352.1p ≈ $4.74 ADR). The AZI GAAP P/E of ~9.3x uses a TTM EPS inflated by the Canada provision credit and ITC gains — discard it; use the adjusted figure.
• ~10.6x EV/EBITDA (EV ~£123.8bn / ~$166bn; EBITDA ~£11.7bn) — up from ~6.3x (2020) / ~7.6x (2021) / ~8.1x (2022): a near-full-turn-per-year re-rating to the top of the decade.
• ~3.6x EV/Sales and ~13x P/FCF on normalized FCF.
• ~5.0% dividend yield — compressed from ~9% at the 2023–24 lows.

Own-history percentiles (the core valuation tell). BAT sits at the richest-ever percentile of its own decade on the clean metrics: P/S ~99th, EV/EBITDA at/near its 8-year high, composite valuation ~95th. The GAAP P/E percentile (also ~99th) is on a noisy denominator — the unambiguous signal is P/S + EV/EBITDA + the composite, all of which say the same thing: the deep-value discount is gone. [FACT — AZI valuation_index; ROIC valuation multiples]

Cross-sectional comps. BAT remains demonstrably cheaper than PM and roughly level with MO:

The PM premium (~60–90%) is earned by PM being the only grower with the de-risked transition; BAT’s discount to PM is rational, not an anomaly. BAT’s discount to its own history, however, has closed. [FACT — peer disclosures; ROIC.ai]

Embedded-expectations math. At ~13.3x adjusted earnings with a ~7–8% cost of equity, the price embeds roughly ~2–3% perpetual real growth in distributable cash — i.e., the market is underwriting that BAT can hold the line (price/mix + Velo + capital return offsetting volume decline) more or less indefinitely, with the FY2026 algorithm partially believed but not fully. What the market is pricing correctly: a durable, high-margin, well-defended cash annuity with a covered ~5% yield and a shrinking float. What it may be pricing incorrectly in either direction: (bull) it gives little explicit credit for the >$15bn residual ITC stake or for a Velo-led re-acceleration toward PM’s multiple; (bear) it gives little weight to the VLN tail, to Vuse’s structural decline, or to the risk that FY2026’s promised acceleration simply doesn’t show — on a re-rated multiple, that disappointment de-rates and misses at once.

Scenario analysis (3-year, illustrative; ADR total return incl. dividends).

• Bear (~−15 to −25%): FY2026 algorithm misses (EPS growth stalls at ~2–3%), Vuse keeps bleeding, a VLN headline or US-enforcement reversal re-accelerates decline; multiple de-rates toward ~9–10x adjusted / ~8x EV/EBITDA. Price → ~$48–53. Yield cushions but doesn’t offset the de-rate.
• Base (~+5 to +20% over 3 yrs, ~mostly the dividend): EPS compounds ~4–5% cc, multiple holds ~12–13x adjusted, ~5% yield collected, buyback retires ~1.5%/yr. Price drifts to ~$66–72 plus ~15% cumulative dividends. A bond-like outcome.
• Bull (~+35 to +50%): Velo sustains double-digit growth to clear US #2, Vuse stabilizes on durable enforcement, the 5–8% EPS algorithm delivers, and BAT partially closes the gap to PM (re-rate to ~15–16x adjusted); ITC monetization surfaces value. Price → ~$82–90 plus dividends.

The asymmetry is roughly symmetric-to-slightly-unfavorable from $62.76: the bull requires further re-rating from an already-rich own multiple, while the bear needs only a growth wobble plus a partial mean-reversion. No price target; the labeled view is in Claude’s Take.

§10 read. A fairly-valued, bond-like total-return profile from here: ~5% yield + ~1.5% buyback + low-single-digit EPS growth ≈ ~high-single-digit base-case total return, with the multiple a one-way risk (little room to expand from the 95th percentile, real room to compress).

11. Variant Perception

Consensus. “BAT is the cheap, high-yield, structurally-weakest major — but it has genuinely de-risked (US growth, Canada settled, leverage repaired, smokeless profitable), and it still trades at a discount to PM, so there’s more re-rating to come.” Broadly correct on the business; the debate is the price.

The strongest bull case. BAT is still the cheapest major (~13x vs PM ~21x) with a covered ~5% yield, a buyback, and a >$15bn ITC stake the market barely credits; Velo is a genuine #2-and-climbing in the best smoke-free category; US decline has structurally moderated on enforcement; the FY2026 algorithm (5–8% EPS) marks an inflection from harvest to modest growth; and a partial re-rate toward PM’s multiple is ~40–50% upside. You are paid to wait while it compounds.

The strongest bear case. The deep-value discount is spent (richest-ever own multiple, yield down from ~9% to ~5%); the moat earns only WACC (~10.6% ROIC) because Reynolds capitalized it away; the smoke-free book is two-thirds losing (Vuse declining, glo a distant #3); the US “growth” is exogenous/enforcement-driven and reversible; the 5–8% EPS algorithm is promised from a +3.4% base; and the VLN tail sits over the largest profit pool. From the 95th valuation percentile, the multiple is a one-way risk.

The 3–5 assumptions that matter most:

1. Durability of US combustible decline-moderation (structural execution vs reversible enforcement). Falsified by a re-acceleration of US volume decline back to −8/−10%.
2. Velo sustaining double-digit growth to clear US #2. Falsified by Velo growth decelerating below ~20% or losing share to ZYN/on!.
3. Delivery of the 5–8% adjusted-EPS algorithm from 2026. Falsified by a sub-4% cc EPS print in FY2026.
4. The multiple holding ~12–13x adjusted (no mean-reversion toward the ~7–9x historical range). Falsified by a de-rate on any growth miss or sector rotation.
5. No VLN final rule. Falsified by the FDA advancing the maximum-nicotine standard toward finalization.

The factor-positioning read (Momentum agent). BTI is a textbook low-volatility, dividend-yield defensive with positive price momentum — beta 0.20, primary factor loading DividendYield (+0.27), Consumer Staples sector, +38% trailing twelve months, only ~6% off its high, with a benign ~14% three-year max drawdown. Risk-adjusted track record is strong and recent (3-year Sharpe ~1.5). This is the empirical signature of a re-rated value name that the market has crowded into for yield and safety — not a falling knife and not (yet) an abandoned value trap. Factor-twins: Imperial Brands (closest), then PM, then a defensive-ADR cohort (Vodafone, Coca-Cola Europacific, Nestlé, Novartis). The positioning read is the evidence behind Claude’s Take’s framing: the crowd is long the yield-and-safety trade here, which is exactly when the margin of safety is thinnest — consensus is offsides on price, not on the business. [FACT — FactorsToday, AZI; INTERPRETATION]

Our variant view. We do not dispute the de-risking; we dispute that the price still pays for the risk. The genuinely contrarian observation is that BAT’s two-year return was almost entirely multiple expansion, not earnings growth (EV/EBITDA ~6.7x → 10.6x while adjusted EPS grew low-single-digit) — so the forward return must come from earnings (the unproven 5–8% algorithm) plus the yield, with the multiple now a headwind, not a tailwind. The market is treating a de-rating that reversed as a compounder that re-rated; those are different, and the distinction is the whole investment question from here.

12. Fact vs. Interpretation Table

13. Open Questions

1. Is the US combustible “return to growth” structural (execution/mix) or exogenous (enforcement-driven decline-moderation)? The single most important determinant of terminal value; needs another 2–4 quarters of US volume data to resolve.
2. What is the precise cadence and present value of the Canada annual payments (85%-of-ITCAN-income, declining over 18 years) as a drag on group FCF?
3. How fast, and at what tax/FDI frictions, is the residual ~23% ITC Ltd stake monetized — and does management return the proceeds or redeploy them?
4. Does Velo sustain double-digit growth into a clear US #2, or does ZYN’s scale and on!'s pricing cap it?
5. Will the FY2026 5–8% EPS algorithm print from a +3.4% FY2025 base, or slip — and how does the ~95th-percentile multiple react?
6. Does the VLN proposed rule advance toward finalization within the planning horizon?

14. What Must Be True

For the bull case to work:

• US combustible decline stays structurally moderated (low-single-digit, not −8/−10%) on durable enforcement; and
• Velo sustains double-digit growth to clear US #2 while New Categories overall re-accelerate to low-double-digit; and
• The 5–8% adjusted-EPS algorithm delivers from 2026, supporting a partial re-rate toward PM’s multiple; and
• The ~5% yield + buyback hold while the float shrinks.
• Falsification test: a single FY2026 (or two consecutive quarters) showing US volume decline re-accelerating toward −8% or cc adjusted-EPS growth below ~4%, or a multiple de-rate below ~10x adjusted on a growth miss — any one breaks the “compounder re-rating” thesis and confirms “de-rate-that-reversed, now fully valued.”

For the bear case to work:

• The multiple mean-reverts from the 95th percentile toward the ~7–9x historical range; and/or
• Vuse keeps bleeding and glo never scales, so the smoke-free offset disappoints; and/or
• A VLN final-rule path or US-enforcement reversal impairs the US franchise’s terminal value; and/or
• FX (a strengthening GBP, or EM currency weakness) erodes ADR value and the translated dividend.
• Falsification test: durable delivery of the 5–8% EPS algorithm with Velo gaining and Vuse stabilizing for four-plus consecutive quarters — which would justify the re-rating and more, and break the “discount is spent” bear.

15. Source Appendix

Primary sources: BAT FY2025 Form 20-F (SEC CIK 1303523, filed 2026-02-13); BAT FY2025 Preliminary Results (bat.com / Investegate RNS, 2026-02-12); BAT 6-Ks (ITC and ITC Hotels block-trade announcements, Mar-2024 / May-2025 / Dec-2025); BAT AGM results 6-K (2025); SEC EDGAR Form 4 (CIK 1303523); ROIC.ai (income statement, balance sheet, profitability, per-share, enterprise value, valuation multiples — accessed 2026-06-27); public market price and valuation data; quantitative factor data; peer-company public disclosures (Philip Morris, Altria, Imperial Brands); FDA/Federal Register (menthol withdrawal; VLN proposed rule); Canadian CCAA settlement public record.

This analysis takes no position and carries no price target. The only view expressed is the clearly-labeled Claude’s Take block at the top — the author’s own independent opinion and general information only, not investment advice.

APPENDIX A — Standard Diligence Questionnaire — British American Tobacco p.l.c. (NYSE: BTI)

Supplemental to the research memo. Report date 2026-06-27. GBP reporting; ADR = 1 ordinary share; USD price $62.76. Fact/Interpretation/Assumption labels applied where it matters.

General

What thoughtful questions have other investors asked about this company?

• Is the 2024–26 doubling a durable re-rating or a melting-ice-cube bounce that has overshot? (The central question — see §10/§11.)
• Is BAT’s US “return to growth” structural execution or a temporary enforcement-driven decline-moderation? [Open Question]
• Why does a 44%-operating-margin business earn only ~10.6% ROIC? (Answer: the 2017 Reynolds price — §4/§7.)
• How much is the residual ~23% ITC Ltd stake (>$15bn) worth in a sum-of-the-parts, and will it be monetized to shareholders?
• Can Velo become a credible US #2 against ZYN, and can Vuse ever stabilize against illicit disposables?
• Is the dividend safe given the 2025 reported FCF/cover optics? (Yes — the optics are a Canada cash-timing artifact; normalized cover ~1.6–1.7x — §6.)

Cyclicality & Earnings Nature

Cyclical high or low? Neither cyclical — secular. Volumes decline structurally (~−8%/yr combustibles); earnings are defended by pricing, not cyclically depressed/elevated. Adjusted EPS is near a normalized level; GAAP EPS is impairment/settlement-distorted (discard). [FACT/INTERPRETATION]

Driven by external environment or internal actions? Both. Internal: pricing, mix, Quantum cost savings, New Categories. External (and currently favorable, but reversible): US illicit-vape enforcement moderating cigarette decline; FX translation; the menthol-ban withdrawal. [FACT]

How stable are revenues? Very stable in character (addicted, recurring demand) but flat-to-declining in level (£25–28bn band; +2.1% cc 2025). [FACT]

Outlook for products/services? Combustibles in managed decline; modern oral (Velo) the one genuine growth vector; vapour (Vuse) declining; heated (glo) sub-scale. [FACT]

How big will this market be — growing or shrinking? Legal nicotine units shrinking in developed markets; value roughly held by pricing; profit pool migrating to smoke-free (which PM, not BAT, captures best). Net: a shrinking-units / flat-value / mix-shifting market. [INTERPRETATION]

Business Quality & Competitive Moat

Industry getting more or less competitive? Legal oligopoly is stable (no new entrants); but illicit competition (disposable vapes, synthetic-nicotine pouches) has intensified and attacks Vuse. [FACT]

How profitable is the business (ROIC, ROE)? ROE ~28.5% (distorted by shrunken equity); ROIC ~10.6% — the honest figure — barely above a ~7–8% WACC. Operating margin 44%; gross margin ~83%. [FACT — ROIC.ai]

How profitable is the industry — competitors, barriers? Extremely profitable; five global majors + China monopoly; barriers (advertising bans, PMTA, brand/addiction) are formidable against legal entrants, porous against illicit ones. [FACT]

Can the business be easily understood? Yes — sell addictive products at high margins, harvest cash, return it. The complications are accounting (impairment/Canada noise) and regulatory, not operational. [INTERPRETATION]

Undermined by foreign low-cost labor? No — brand/regulatory moats, not labor-cost-based. The relevant low-cost threat is illicit Chinese disposable vapes (a regulatory-arbitrage threat to Vuse, not a labor threat). [INTERPRETATION]

Do brands matter? Yes — brand + addiction is the core moat (Newport, Camel, Dunhill, Lucky Strike). But BAT’s own 2023 ~£27bn write-down concedes the US brands are finite-lived, not perpetual. [FACT]

Nature of competition? Share-stable in combustibles (price/mix competition); share-losing to PM in the two most valuable smoke-free categories (IQOS, ZYN); share-gaining with Velo. [FACT]

Customers’ switching costs? Near-zero beyond habit/addiction, except glo’s modest device lock-in (and glo is sub-scale). [FACT]

Financial Condition & Balance Sheet

Assets not fully recognized? The residual ~23% ITC Ltd stake (>$15bn) is carried as an associate/investment but its market value is a real, under-credited SOTP asset. [FACT/ASSUMPTION]

Off-balance-sheet liabilities? The Canada CCAA settlement is now on-balance-sheet (provision); the recurring annual payments (85% of ITCAN income, declining over 18 yrs) are a known future cash drag. Ordinary tobacco litigation otherwise. [FACT]

How conservative is the accounting? Mixed-to-conservative: the 2023 impairment and the finite-life brand reclassification were conservative (and honest); adjusting items are fully itemised and consistent. The one cherry-pick risk: ITC divestment gains run through adjusting items (excluded from adjusted EPS — directionally conservative). [INTERPRETATION]

How CapEx-hungry? Light — capex ~£0.7–0.9bn on ~£25.6bn revenue (~3%); the business is asset-light and cash-generative. [FACT]

Capital Allocation & Management

FCF generation and use; philosophy? ~£8–9bn normalized FCF; priority order: deleverage to 2.0–2.5x (done), progressive dividend (~67% payout), buyback (£1.3bn 2026), no large M&A. Disciplined and shareholder-aligned incrementally. [FACT]

Significant acquisitions recently? None large since Reynolds (2017, ~$49bn — the value-destructive central event). New Categories built organically. [FACT]

Buying back shares? Yes — restarted 2025 (£1,173m), £1.3bn for 2026; counter-cyclical and accretive at ~13x. Share count 2,294m (2020) → 2,173m (2025). [FACT]

Issuing shares to insiders? No material dilution; modest LTIP issuance. [FACT]

Compensation policy / incentives? LTIP/STIP on adjusted EPS growth, New Categories revenue, cash-flow conversion (>95%), 3-yr relative TSR, + 10% climate. No ROIC/return-on-capital metric — a real gap given Reynolds. Say-on-pay support ~98%. [FACT]

Motivations of management? CEO Tadeu Marroco (since 2023) — credibility-rebuilding regime (reset, de-risk, return cash). No insider open-market buying surfaced (UK PDMR disclosure via RNS, not Form 4). [FACT/INTERPRETATION]

Valuation & Market Data

ADR, MLP, or K-1? ADR (1 ADS = 1 ordinary share; quarterly, FX-translated dividend). UK foreign private issuer; no K-1. UK has no dividend withholding tax for US holders — a structural advantage vs many foreign ADRs. [FACT]

Dividend policy? Progressive; 245.04p (2025), +2%/yr; ADR yield ~5.0%; ~67% payout of adjusted EPS; covered ~1.6–1.7x on normalized FCF. [FACT]

How profitable? Very (44% operating margin) on an operating basis; only ~WACC on an invested-capital basis. [FACT]

Net income diverging from cash from operations? Yes, temporarily — 2025 OCF (£6.3bn) was depressed by Canada settlement cash timing vs net income £7.8bn; normalize to ~£8–9bn FCF. GAAP net income is the more distorted of the two (impairment/settlement noise). [FACT/INTERPRETATION]

Risks & Downside

What would cause the stock to decline? A FY2026 EPS miss from a +3.4% base; a de-rate from the 95th valuation percentile; Vuse decline / glo failure; a VLN regulatory headline; US enforcement reversal re-accelerating cigarette decline; a strengthening GBP hitting the ADR. [FACT/INTERPRETATION]

Risk of catastrophic loss? Low — diversified, IG-rated, ~£8–9bn FCF, covered ~5% yield. The genuine long-tail existential risk is a US VLN final rule (slow-moving). [INTERPRETATION]

Chance of total loss? Negligible on any near/medium horizon. [INTERPRETATION]

Recent News & Events

Has the business environment changed recently? Yes, net-positively over two years: US returned to growth, Canada settlement implemented (Aug-2025), menthol ban withdrawn (Jan-2025), New Categories profitable, leverage repaired, buyback restarted, ITC monetized — but the stock has ~2.6x’d, so the de-risking is now priced. [FACT/INTERPRETATION]

Significant acquisitions / accounting changes? No acquisitions; the material accounting change was the 2023 reclassification of US brands to finite-life (~£1.4bn/yr amortisation). [FACT]

Recent changes — markets, facilities, management? glo Hilo launch (Sept-2025); Velo Plus US relaunch; Quantum cost program; CEO Marroco (2023); ongoing ITC/ITC Hotels sell-downs. [FACT]

APPENDIX B — Source Appendix

Report date 2026-06-27. Primary sources first. Internal/aggregator sources labeled. Every non-obvious memo fact is cited inline; this appendix consolidates the source base.

Primary — Company filings & disclosures

• BAT FY2025 Form 20-F — SEC EDGAR CIK 0001303523, filed 2026-02-13. https://www.sec.gov/Archives/edgar/data/1303523/000130352326000017/bti-20251231.htm (annual report, segment/geographic disclosure, notes on impairment (4/12), Canada settlement (24), remuneration report).
• BAT FY2025 Preliminary Results — bat.com, 2026-02-12. https://www.bat.com/media/press-releases/_2026/february/preliminary-results-for-the-year-ended-31-december-2025 ; Investegate RNS mirror. (Revenue, adjusted op. profit/margin, adjusted diluted EPS 352.1p, New Categories contribution, US/AME/APMEA segments, FY2026 guidance.)
• BAT 6-K filings — SEC EDGAR CIK 1303523 (2024–2026): ITC Ltd block trades (Mar-2024 ~£1,577m; May-2025 ~£1,052m); ITC Hotels block trade (Dec-2025 ~£315m, retained 6.3%); AGM voting results (2025, remuneration ~98%); daily “transaction in own shares” (buyback execution); TR-1 holdings notifications.
• BAT prior 20-Fs (FY2021–FY2024) — EDGAR CIK 1303523 (multi-year financials, the 2023 ~£27bn impairment / brand reclassification).
• SEC Form 4 — CIK 1303523, filed 2026-05-29 (BAT subsidiary purchase of Charlotte’s Web convertible debenture — not a BAT-stock insider transaction).

Primary — Regulatory

• FDA menthol cigarette ban — withdrawn (Jan-2025) — OMB Unified Agenda / tobacco-law coverage.
• FDA maximum-nicotine (very-low-nicotine) product standard — proposed — Federal Register 2025-00397, 2025-01-16; comment period to 2025-09-15.
• Canadian CCAA tobacco-litigation settlement (CAD$32.5bn industry; ITCAN/RBH/JTIM) — creditor-approved Dec-2024, sanctioned Mar-2025, implemented 2025-08-29 (public court record / Davies coverage).

Secondary — Industry & trade press

• ECigIntelligence / Tobacco Insider / SnusFriends / ZYLO — Vuse, glo, Velo category share data (corroborating, not primary).
• Bloomberg / Business Today — ITC and ITC Hotels stake-sale coverage.
• CNN / trade press — Dec-2023 ~£25bn write-down coverage.

Quantitative aggregators (cross-check, reconciled to filings)

• ROIC.ai MCP (accessed 2026-06-27) — income statement, balance sheet, profitability ratios (ROIC ~10.6%, ROE 28.5%, margins), per-share data, enterprise value (~£123.8bn), valuation multiples (EV/EBITDA history 6.3x→10.6x). Third-party aggregated; reconciled to the 20-F. GBP reporting.
• AZI valuation_index & news (accessed 2026-06-27) — own-history valuation percentiles (P/S ~99th, composite ~95th); price series (5-yr low $23.67, high $67.30). News feed thin for this foreign issuer.
• FactorsToday (accessed 2026-06-27) — factor loadings (DividendYield +0.27, Consumer Staples), leaderboard (beta 0.20, y1 +38.6%, 3-yr Sharpe ~1.5, maxDD −13.8%), related-stocks (IMBBY closest, then PM, VOD, CCEP, NSRGY, NVS).

Peer disclosures — cross-read

• Altria (NYSE: MO) public filings & investor materials — US menthol/VLN regulatory timeline; illicit-vape enforcement; nicotine-pouch dynamics; next-gen M&A contrast.
• Philip Morris International (NYSE: PM) public filings & investor materials — IQOS ~76% heated share; ZYN; smoke-free crossover economics; cross-sectional valuation; negative-tangible-equity treatment.