Pfizer

When we drew up a shortlist of possibly-cheap companies, we said openly that Pfizer looked the cheapest and that our prior was that it is probably a value trap. We also said we would test that rather than confirm it. This report is that test, and the answer is more interesting than either caricature.

Pfizer sells patented medicines. That single sentence contains the whole problem, because a patent is a wasting asset with an expiry date printed on it. A brand does not expire — Coca-Cola's recipe is not scheduled to enter the public domain. A drug patent is. Which means this company must, every decade, replace a large share of its own revenue or shrink.

And Pfizer has told the market exactly how large that share is. From its own investor presentation of 16 December 2025, in its own words:

"Anticipate ~$17 billion in annual revenue impact from LOE products by 2030 (vs 2025)." — Pfizer investor presentation, slide 15. That is company guidance, not an analyst's estimate.

Against 2025 revenue of $62.6 billion, that is 27% of the business. Pfizer has told its owners it expects to lose more than one dollar in every four it earned last year, within five years. Everything else in this report — the pipeline, the acquisitions, the dividend — is an argument about whether that hole gets filled.

★ That assumption is a lever worth knowing about. Pfizer's $17 billion is a net figure — it already assumes vaccines hold up and some products erode slowly. If the exposed base erodes 75% rather than 58%, the impact is closer to $22 billion, and the hole in Part VI widens by a third.

★ Pfizer has roughly two more years of near-normal cash generation before the step. That is why management refuses to buy back a single share, and it is why "the dividend is fine today" and "the dividend is fine in 2029" are two completely different questions.

Look at which products are carrying Pfizer's current growth. In the first half of 2026, Eliquis grew 14% and Vyndaqel 6% — the two largest items on the cliff list. And Eliquis grew primarily on higher US price, not volume.

A product taking price in the last twenty-four months before generic entry is borrowing revenue from its own future. It flatters the results now and steepens the drop later. ★ So the better 2026 and 2027 look, the harder 2028 lands — which is precisely the optical illusion that makes this sector so dangerous for investors who buy on trailing numbers.

In December 2023 Pfizer completed the largest acquisition of its recent history: Seagen, for $43.4 billion net of cash, funded substantially by a $31 billion bond issue. The stated aim was to lead the field in antibody-drug conjugates — a technology that attaches a chemotherapy payload to an antibody that seeks out cancer cells — and to double the oncology pipeline.

Here is what that $43 billion is generating today.

★ Pfizer paid $43 billion for a portfolio now generating about $3.9 billion a year — roughly eleven times revenue — of which only one product is growing meaningfully. The oldest and once-flagship product is shrinking 18%.

The sigvotatug failure deserves a paragraph of its own, because of how it failed. It was designed to target a protein found on about 90% of solid tumours — the clearest embodiment of the whole "platform" thesis. From Pfizer's own results release in August:

"In the exploratory analysis, no clear IB6 expression-response relationship was observed." — Pfizer, second-quarter 2026 results

Translated: the drug did not work better where its own target was more abundant. When a targeted therapy shows no relationship between effect and target expression, what is in question is not the dose or the trial design — it is the targeting hypothesis itself.

★ And that is why we read this as a pattern rather than bad luck. Three separate programmes from this acquisition have now been written down or killed. Pfizer bought antibody-drug-conjugate leadership and has so far demonstrated antibody-drug-conjugate fallibility.

The fair defence, and it is real. Padcev is genuinely excellent — growing 30%, and approved in July 2026 as the first platinum-free regimen in its setting, which materially expands who can take it. It could plausibly reach $5–6 billion alone. Around twenty Phase 3 oncology programmes remain. Elsewhere the science is delivering: Lorbrena's seven-year data showed 55% of patients alive without progression against 3% on the older drug. This is not a company that cannot do research. ★ It is a company that has repeatedly paid too much for it.

⚠️ And the pattern extends beyond Seagen. Global Blood Therapeutics cost $5.4 billion in 2022; its drug Oxbryta was withdrawn from the market and the remaining value written off in the second quarter of 2026. That is a total loss on a $5.4 billion acquisition.

We have published analyses of both Novo Nordisk and Eli Lilly, so readers deserve to know exactly where Pfizer stands in the largest new market in pharmaceutical history. The short answer is: late, and it cost a great deal to get there.

Two failures first. Lotiglipron was discontinued in 2023 over liver enzymes. Then danuglipron — the once-daily oral pill that was Pfizer's whole obesity hope — was discontinued on 14 April 2025 after a trial participant showed signs of drug-induced liver injury. That was the second oral obesity failure in under two years, and it left Pfizer with essentially nothing.

Then it bought its way back in, and the manner matters.

★ Pfizer paid roughly double its own opening bid, in a public auction, against a competitor with its own reasons to be desperate, for a Phase 2 asset — four months after writing off its own obesity programme. Winning a bidding war is not the same thing as making a good investment.

That said, the asset itself is genuinely good, and better than the bear case allows. Berobenatide is a long-acting injectable that supports weekly and monthly dosing. Phase 2b data presented in June 2026 showed 15.9% weight loss at 32 weeks with no plateau, low gastrointestinal side effects and a very small injection volume. Ten Phase 3 studies begin this year.

The honest read: berobenatide is credible and differentiated — monthly dosing genuinely matters — but Pfizer is three to four years behind two competitors who will by then hold entrenched formulary positions, manufacturing scale and physician habit. The efficacy is Wegovy-class, not CagriSema-class.

★ And for the question that actually matters in this report — does obesity help fill the 2030 hole? — the answer is essentially no. A 2029 or 2030 launch produces immaterial revenue inside the window where the patent cliff bites. It is a 2032 solution arriving to fix a 2028 problem. That timing mismatch is the thing most bullish write-ups quietly skip.

Implied revenue in 2030: roughly $55–56 billion, against $62.6 billion in 2025. A business smaller in 2030 than it was in 2025 — seven years after the largest cash windfall in the history of the industry and some $60 billion of acquisitions.

★ And one silence is worth as much as any number. Pfizer's own framing is that growth towards the end of the decade will come from a maturing pipeline, completed deals and new launches. That is a claim about direction. Pfizer has conspicuously declined to publish a 2030 revenue target. For a company that publishes a precise figure for the damage, declining to publish one for the recovery is itself information.

★★ The pipeline is real, materially better than the market's caricature of it, and insufficient. Pfizer does not collapse. It also does not grow. It is not a melting ice cube and it is not a compounder — which is exactly why "cheap or trap?" is a genuinely hard question here rather than an obvious one.

Our house rule is that any dividend-paying company gets its payout examined properly. At a 6.4% yield this is the reason most people would own Pfizer, so it deserves the most careful section in the report — including two corrections to things we ourselves said while researching it.

The 226% is not a cash fact. Roughly $4.7 billion a year of amortisation of acquired intangibles — about 83 cents a share — is a non-cash charge that never touches the cash flow, plus $4.3 billion of one-off write-offs in a single quarter. Add back the amortisation alone and reported earnings are already about $1.59 before touching the impairments.

⚠️ The same artefact produced a second error, this time in our own data feed. It shows net debt at 4.9 times earnings — which, for a company facing a 27% revenue cliff, would be a balance sheet in trouble and a dividend on the block. Pfizer's own figure, on adjusted earnings, is about 2.7 times. That is investment grade and manageable. ★ We flag this because a reader running the same screen we did would reach a materially wrong conclusion about solvency.

While researching this we told our own reader that the dividend was "covered 1.12 times" by cash. That figure is for the trailing twelve months. On the fiscal year accounts it is worse, and we would rather correct it than leave it standing.

★★ In two of the last three years Pfizer paid out more cash in dividends than it generated in free cash flow. And the largest product on its patent cliff has not gone generic yet.

★ The dividend has also been frozen. It rose from 42 to 43 cents in January 2025 — a token 2.4% — and has sat at 43 cents for six consecutive quarters since. We have now seen this exact pattern at Diageo, at Innovative Industrial Properties and at Deere: the freeze always comes first, and almost nobody writes about it because "unchanged" is not a headline. ⚠️ Note also that Pfizer has cut this dividend before, in 2009, to fund the Wyeth acquisition.

So: cheap, or a trap? We promised to answer that, and the honest answer requires defining the difference.

A cheap stock is one whose price has fallen further than its earning power. A value trap is one whose price has fallen because its earning power is falling, and will keep falling. The test is not the multiple; it is the direction of the denominator.

★★ Pfizer trades at nine times earnings that consensus expects to be 24% lower in three years. That is the definition of the trap — a low multiple on a declining number. But it is not the whole answer, because the decline is finite, disclosed, and already scheduled.

Here is where we come out. This is not a classic value trap, because a classic trap denies its problem. Pfizer published the size of its own cliff, to the billion, in an investor deck. Nor is it cheap in the way a cyclical at a trough is cheap, because there is no cycle here that turns — a patent that expires does not come back.

It is a third thing: a business being priced, roughly correctly, for a managed decline in its current portfolio and an uncertain replacement. At nine times earnings with a 7.2% cash yield, you are being paid something for that. Whether you are being paid enough depends almost entirely on whether the $7 billion gap in Part VI is $7 billion or $12 billion — and nobody, including Pfizer, currently knows.