Eli Lilly Q2 2026 Earnings: $23.0bn Revenue, +48% — and Why the Guidance Raise Still Implies a Slower Second Half

Eli Lilly reported second-quarter revenue of $23.0 billion, up 48% year-on-year, against a consensus that clustered near $20.5 billion — a beat of about $2.5 billion, or roughly 12%. Non-GAAP earnings per share came in at $8.38, above even the highest published consensus figure for the quarter. Full-year revenue guidance moved up to $85–87 billion from $82–85 billion, and the shares rose more than 5% before the open. But the number that frames the rest of 2026 is not the beat. Subtract the first half from the new guidance range and the company is still pointing at a quarterly run-rate below the quarter it just delivered.

Ahead of this print we argued the interesting content was not the surprise-versus-consensus gap but three mechanisms that could be checked against published evidence: the arithmetic embedded in the guidance range, the first datapoint from an oral GLP-1 launch, and the gross-margin line where pricing concessions land. All three now have answers. Two went in the direction the setup implied. One went the other way.

What actually happened

Lilly released results before the US open on Wednesday 5 August 2026, with the conference call at 10:00 a.m. Eastern, as the company had confirmed on 22 July.

Revenue was $23.0 billion — $22.974 billion precisely — up 48% from $15.6 billion in the same quarter of 2025. Reported earnings per share were $7.94, up 26%; non-GAAP earnings per share $8.38, up 33%. Mounjaro delivered $9.943 billion, growth of 91%, with international sales now marginally ahead of US sales for the product. Zepbound contributed $4.928 billion, up 46%. Foundayo, the oral GLP-1 approved on 1 April and shipping from 6 April, recorded its first revenue line at $98 million.

Chief executive David A. Ricks framed it in the release as continued momentum, noting the company "delivered 48% revenue growth and raised our full-year guidance." The shares rose 5.7% in pre-open trading.

Two items sit below the headline and matter more than it does. Acquired in-process research and development charges were $2.8 billion in the quarter, against $154 million a year earlier. Asset impairment and restructuring charges were $703 million. Both are why the reported and non-GAAP figures diverge, and the first is why the quarter's consensus EPS was never a usable benchmark.

The guidance arithmetic, redone

Before the print, the useful number was a subtraction rather than an estimate: full-year guidance of $82–85 billion, less a Q1 of $19.8 billion, implied $62.2–65.2 billion across Q2–Q4 — an average of roughly $20.7 to $21.7 billion a quarter. The $20.5 billion consensus sat below the bottom of that band, which meant the second half had to accelerate for even the low end of the range to hold.

The quarter resolved that in the most direct way available: $23.0 billion is above the top of the implied average, by about $1.3 billion. The burden of proof the old range imposed has been discharged for one quarter.

Now run the same subtraction on the new numbers.

The raise was real — $3 billion at both ends of the range — and it still leaves a second-half run-rate beneath the quarter just delivered. There are two readings and the guidance range distinguishes between them not at all. One is conservatism: a company that has raised revenue guidance at consecutive quarters is setting a bar it expects to clear rather than one it expects to meet, and a mid-year launch plus a July Medicare pricing change are exactly the sort of inputs a finance function declines to forecast aggressively. The other is a genuine expectation of sequential moderation — realised prices continuing to fall, an unusually strong international quarter for Mounjaro not repeating, or oral competition biting.

What makes this worth stating plainly is that the market has to choose between those two readings without help, and the choice is what a guidance range always leaves unresolved. The same structural point applied to Palantir's quarter: a result describes a period that has ended, a range describes periods that have not, and only the second one can be revised.

Why the EPS "beat" was unreadable — and the guidance range fell anyway

The setup flagged that Lilly's Q2 consensus EPS was the weakest number in the preview, with published figures spanning roughly $6.06 to $7.74 for the same quarter, mostly because providers treat acquired IPR&D charges differently. That is precisely how it played out. The reported non-GAAP figure of $8.38 sat above every one of those estimates; one widely circulated framing described it as a beat of $2.31 against a $6.07 estimate, which is arithmetically true and analytically close to meaningless, because a different provider's number would have produced a different margin of victory on the same morning.

Foundayo's $98m, and the question it did not answer

The oral launch produced a number: $98 million in its first full quarter. Against published expectations that is close to the top of the range — a FactSet-compiled consensus of about $103 million, with at least one large house modelling roughly $40 million. A dispersion that wide on a single product line is itself informative about how little anyone knew, and it means the result cannot fairly be called either a disappointment or a blowout.

What it does not do is answer the mechanism question, and it is worth being explicit rather than implying otherwise. The interesting issue with an oral GLP-1 was never whether it sells; the injectable franchise settled that. It is whether a pill reaches patients who would never have injected — expanding the treated population — or whether it redistributes patients already inside the franchise at a lower price per patient. The revenue line alone is consistent with both, and the release did not break out composition.

The evidence that will settle it is sequential: two or three quarters of oral revenue set against injectable volumes over the same period. If injectable growth holds while oral revenue builds, the population expanded. If oral revenue builds as injectable growth decelerates beyond what pricing explains, it did not. Q2 is datapoint one of a series, and Novo Nordisk's competing oral product launched into the same window, which means the series will not be clean.

Gross margin went the other way

This is the finding that cut against the setup, and it deserves to be reported as such.

Lilly's own Q1 language — revenue growth "primarily driven by volume growth, partially offset by lower realised prices" — pointed at gross margin as the line where price concessions surface unmasked by volume. Most-favoured-nation pricing, $25-a-month commercial coverage, $50-a-month Medicare Part D access and $149 self-pay all land there.

Gross margin was 85.8% of revenue on a reported basis and 86.3% on a non-GAAP basis in Q2 — the latter up 1.3 percentage points year-on-year. It rose.

The mechanism is not mysterious once the number is in hand. Realised price per patient can fall while gross margin percentage rises, because the two are different ratios: margin depends on price against unit cost, and manufacturing cost per unit on a franchise scaling at this rate falls faster than realised price does. Fixed manufacturing capacity spread across 48% more revenue is the whole story. That does not mean the pricing pressure is imaginary — it means volume is currently large enough to overwhelm it. The line to keep watching is what happens to this ratio in a quarter when revenue growth decelerates, because that is when cost absorption stops doing the work.

Tariffs, MFN pricing, and the index channel

Nothing in this quarter changed the tariff position, and it remains the most misread part of the Lilly story. The Section 232 pharmaceutical measures announced on 2 April 2026 carry a 100% headline rate, reduced to 15% for imports from jurisdictions holding US trade agreements — the EU, Japan, Korea and Switzerland among them — and 0% on generics and biosimilars, with duties effective for large companies from 31 July. We covered the structure and its currency consequences in the pharma tariff piece.

Lilly holds an exemption that predates the duty. The November 2025 agreement announced by the White House covers most-favoured-nation pricing for Medicaid and cash-paying patients through the TrumpRx portal, alongside a $27 billion US manufacturing commitment and a three-year reprieve from the tariffs, with the exemptions scheduled to sunset on 20 January 2029. The relevant question for this quarter was therefore never duty cost but how fast MFN pricing flows through realised price — and the gross-margin line above is where that answer lives. For now, it is not visible.

On the index: at roughly $1.02 trillion of market capitalisation against an S&P 500 near $67 trillion, Lilly's weight is about 1.5%, so a 5.7% move is on the order of 0.09 percentage points of the index in isolation. Real, and smaller than the coverage implies. The larger transmission is correlation — Lilly is the reference point for GLP-1 economics, and a result that resets the read on volumes or pricing reprices healthcare names that never reported. For instrument mapping: Lilly is NYSE-listed and not a Nasdaq-100 constituent, so this is an US500 and healthcare-sector event rather than a NAS100 one.

As for currencies, this remains genuinely not a foreign-exchange catalyst. The pharma-tariff FX channel runs through the exporters facing the duty rather than the US manufacturer holding an exemption — a Swiss franc and euro story. What touches USD here is second-order: an earnings season that shifts the equity risk backdrop feeds the risk factor, and the discount rate on long-dated pharmaceutical cash flows is set by the rates factor, not by anything in this release.

What would change the picture from here

Three things, in order of how much they would move the analysis.

The Q3 guidance treatment. The current range implies an H2 quarterly average below Q2's $23.0 billion. If Q3 lands at or above $23 billion and the range is raised again, the conservatism reading is confirmed and the pattern becomes the story. If Q3 lands inside the $21.1–22.1 billion implied band, the moderation reading was right and the deceleration is real.

Foundayo composition — any disclosure or call detail on new-to-class versus switched volume. The revenue line is now known; the population question is untouched.

Gross margin in a slower quarter. 86.3% with revenue up 48% tells you cost absorption is winning. The informative observation is the same ratio when growth halves, because that isolates the pricing effect from the scale effect.

Sources for the figures above: Lilly's second-quarter 2026 results release and first-quarter release; BNN Bloomberg on the guidance raise and product detail; the November 2025 White House fact sheet on the most-favoured-nation agreements; and BioPharma Dive on the Section 232 structure and exemptions. More on how this site reads events like this on the about page.

Educational macro context only — not investment advice.