The annuity machine
Take the leader in robotic surgery (Intuitive Surgical). Its business is what you'd call an annuity machine: an installed base of more than 11,000 systems worldwide, and every placed system pulls years of instruments, accessories, and service behind it. Roughly 86% of revenue is recurring. The result (as of mid-June 2026): around USD 10bn in revenue and nearly USD 2.9bn in profit for fiscal 2025 — a profitable, growing company.
AI plays a real but incremental role here: digital connectivity, system monitoring, new features like force feedback. It's a giant using AI as an efficiency lever, not a creature that lives on AI.
That doesn't make it risk-free. The stock is down roughly a quarter in 2026 — not because the numbers were bad (they beat expectations), but because a valuation multiple of around 48 times earnings leaves little room when weight-loss drugs shrink certain procedure types and competition starts to scratch at the moat. The annuity machine's risk is a de-rating — a slow erosion of the premium. Not a wipeout.
The option
Now the other end: an AI-native drug-discovery platform (Recursion). The story is impressive — a vast dataset, a supercomputer built with a chip giant, partnerships with big pharma, high-profile backers. But the financials are unsparing (as of June 22, 2026): around USD 75m in revenue against roughly USD 645m in losses for fiscal 2025. The company burns about eight times its revenue and lives off the capital markets, not off products. Approved drugs to date: none.
The share price reflects that: around USD 3, roughly 90% below its 2021 high, near its yearly low. About a third of the float is sold short — a rare concentration of skepticism. At the same time, the average analyst target sits more than 100% above the price, on a "Hold" rating. That is precisely the signature of an option: a high multiplication chance if a clinical proof lands — and a real total-loss risk if it doesn't. Several decisive trial readouts over the next 12 to 18 months make the window concrete.
Side by side
| Dimension | Annuity machine | Option |
|---|---|---|
| What you own | installed base plus recurring revenue | platform plus clinical optionality |
| Revenue / profit | ~USD 10bn / ~USD 2.9bn profit | ~USD 75m / ~USD 645m loss |
| Revenue quality | ~86% recurring, profitable | burns ~8x revenue, lives off capital markets |
| Proof status | 14m procedures, established category | no approved drug, Phase III pending |
| Valuation | ~48x earnings (priced for perfection) | negative P/E (priced on hope) |
| Main risk | de-rating, competition | clinical failure, dilution |
| Loss mode | slow (downward re-rating) | binary (a trial fails) |
| Sentiment | analysts "Buy", moat intact | ~⅓ of float shorted, "Hold" |
The methodical core
Both companies sit under the same label — "AI in healthcare". As investments, they have almost nothing in common. One falls slowly if it falls; the other can multiply or go to zero. One is expensive and proven; the other cheap-looking and unproven. Wanting to be "in the sector" is not yet a decision — it only names the shelf.
So the transferable lesson is: sector is not position. Before you buy, locate yourself on the spectrum between annuity and option — not because one is better than the other, but because they express fundamentally different things and demand fundamentally different stomachs. And if you can't decide which single winner prevails, you have the two clean exits from the earlier pieces: the durable layer of the value chain — or the broad basket that contains both ends anyway.
This is not investment advice. The companies referenced are illustrative examples of two risk profiles, not a recommendation of individual securities. Figures are as of mid-June 2026 and subject to change; this sector is more volatile than most.
FAQ
Can the same sector hold two opposite risk profiles? Yes. Within "AI in healthcare," the range runs from a profitable annuity machine with recurring revenue to an unprofitable option with no approved product. The sector label hides this opposite risk DNA.
What does "annuity machine" mean? A company whose installed base throws off recurring revenue for years — here around 86% of sales. It's profitable; its main risk isn't collapse but a de-rating, if a high valuation multiple comes under pressure.
What does "option" mean for a stock? A profile with a binary payout: a high multiplication chance if a clinical proof lands, and a real total-loss risk if it doesn't. Such companies are often deeply loss-making, burn capital, and depend entirely on future trial success.
Is one better than the other? No. They express different things and demand different risk tolerances. The methodically clean question isn't "which is better," but "where on the spectrum between annuity and option do I want to stand — and can I stomach its particular loss mode?"