ELI LILLY & Co

Lilly's earnings are strong on paper, but thin equity and missing cash data make its future hard to estimate.

Eli Lilly discovers, develops, makes and sells human prescription medicines, and reports one business segment. Revenue was $65.2B and net income was $20.6B in the latest year, with much of the sales coming from a few cardiometabolic products. Our tests show profits in almost every year measured, with 1 loss year across 15 years of filings. Return on equity averaged 71%, but equity is thin at $26.5B of book value, so that average overstates the economics. The shares trade at 54× earnings and 42× book value, and we cannot compute owner earnings to test whether that price is backed by cash.

Eli Lilly discovers, develops, makes and sells human prescription medicines, and reports one business segment. Revenue was $65.2B and net income was $20.6B in the latest year. A large share of sales comes from a few cardiometabolic products, sold mainly through U.S. wholesalers and, in some markets, direct channels such as LillyDirect.

• Volume growth in cardiometabolic medicines, led by Mounjaro and Zepbound
• Patent and data protection that allows premium prices until exclusivity ends
• Launches of new medicines and new indications from the pipeline
• Gross margin on product sales, which rose in 2025
• Sales outside the U.S., mainly from Mounjaro

Inside the circle? Patent expiry, government price setting and pipeline outcomes can each cut revenue quickly, and the filings say Lilly cannot reliably estimate the cost or timing of its development projects. Owner earnings cannot be computed from our data, and the return on equity figure is distorted by thin equity, so the economics cannot be tested properly. A careful owner could describe the business today but would struggle to estimate its earnings ten years out.

Lilly's protection rests mainly on patents and data exclusivity for its medicines, with trademarks and the Lilly name adding some further protection. The filings say a branded medicine can lose its protection quickly, and that loss of exclusivity often leads to a severe and rapid decline in revenues. Net margin averaged 20% and return on equity averaged 71%, but equity is thin (book value is $26.5B), so these ratios overstate the economics. We cannot cross-check returns on capital, because operating income and invested capital are missing from our data. On the evidence we have, the moat is narrow: it lasts as long as a few key products keep their exclusivity and payers and governments allow prices to hold.

Management is spending heavily on factories and on buying research programmes, and it is also returning cash to owners through buybacks and a rising dividend. Capital spending rose in 2025 and the filing says it will be meaningfully higher in the near term. Total debt also grew, which means the company is funding part of its growth with borrowing. The share count trend of -1.8% a year points to net shrinking, which is a positive sign for owners. The filing is frank about what it cannot forecast, but owner earnings cannot be computed from our data, so we cannot test whether buybacks and acquisitions are made below intrinsic value.

• Capital spending rose sharply, and the filing expects it to stay high for several years. Cash spent on plants is cash that cannot go to owners in the short term.
• Buybacks and a rising dividend return cash, but total debt rose in the same year. Our net cash reading is -$33.6B and debt to equity is 1.5×.
• Acquired research charges appear in both 2024 and 2025, so part of the pipeline is bought rather than built. The filing does not say how much of these deals has paid off, so we cannot apply the $1 retained test to them directly.
• The filing says the company cannot reliably estimate the cost or timing of its development projects. That candour is useful, but it also limits how far a buyer can estimate future cash.

• MixedDemonstrated, consistent earning powerMixed

  One loss year across 15 years of filings.

  Why this test
• MixedHigh returns on equity without leverageMixed

  Return on equity averaged 71%, reaching 15% in 9 of the years measured; the weakest year was -1.6%. The ratio is unreliable: an average above 60% usually reflects very little equity rather than extraordinary economics.

Discount rate: 10-year Treasury (5.1%) plus a 3% hurdle. End-of-day prices from free public sources, may be delayed or missing.

Buffett's tests start with whether you can understand a business and estimate its future earnings. Lilly's reported profits show the demonstrated, consistent earning power he asks for, with 1 loss year in 15 years. The moat is narrower than the profits suggest. The filing warns that loss of patent protection can cause a "rapid and severe decline in revenues", and most of the protection is time-limited. Pricing power is hard to judge. The filing describes government price setting and rebates, and says 2025 volume growth was "partially offset by lower realized prices". Owner earnings, the cash test Buffett relies on, cannot be computed from our data, and the margin of safety test therefore cannot be run. The 10-year Treasury yield of 5.1% also matters, because a higher rate lowers what future cash is worth today. Our research status is hard to predict. The filing says it "cannot reliably estimate the nature, timing, and costs of the efforts necessary to complete the development" of its research projects, so the long-range earnings picture stays uncertain.

• Depreciation and capital spending figures in future filings

  These are needed to compute owner earnings, the cash test that our current data cannot support.

• Capital spending against operating cash flow

  The filing says capital spending will be meaningfully higher in the near term, so check whether operating cash keeps pace.

• Government-set prices for Jardiance, Trulicity and Verzenio

  The filing names Jardiance for 2026 and Trulicity and Verzenio for 2028, and says more products are likely to be selected.

• Regulatory decisions on orforglipron and the share of revenue from Mounjaro and Zepbound

  The filing says near-term results depend on approval timing, and these two products make up a large share of revenue.

• Share count and buyback activity against the price paid

  Buybacks only create value for owners when the shares are bought below intrinsic value, which we cannot yet estimate.

Questions the filings couldn't answer

• The filing text gives a total debt figure that differs from the figure in our data. Which figure is right, and does the difference change the debt test?
• What depreciation and capital spending figures are needed to compute owner earnings, and how do they compare with reported profit?
• Does the price of the shares sit below a conservative estimate of value, once the cash the business can return is estimated?