Dario Amodei says AI could end cancer and Alzheimer's in a decade

Felix Laurent 11 Nov 2026 4 min read

Big tech

Dario Amodei says AI could make cancer, Alzheimer's, heart disease and mental illness "relics of the past" in less than a decade. The Anthropic CEO made the claim at the end of the company's IPO roadshow video, published on November 10. A clip of the line spread on X a day later.

The timing gives the claim extra weight. Anthropic plans to price its shares on November 18, a week after the clip went viral. Investors are deciding whether to buy into a company valued at close to $1 trillion.

What Dario Amodei said in the roadshow video

The roadshow video runs about 43 minutes and is posted on Anthropic's investor site. Amodei's prediction comes near the very end, at about the 40-minute mark. He names "cancer, Alzheimer's, heart disease, mental illness" as the targets. "These will be relics of the past in, I think, less than a decade," he says. The quote comes from transcripts by Constellation Research and CellCog.

He then ties that promise to safety. Amodei argues that the company which manages AI's risks well will also earn society's trust and a "license to operate" from it. His last words invite investors to "join us on the next phase of this journey" with Anthropic.

The disease forecast is Amodei's personal estimate. He offers no study, trial or timeline behind it, and the line does not appear in the prospectus. The founders' letter on Anthropic's IPO site is more careful. It says AI could "bring cures for major diseases within reach."

The same video gives a much smaller figure for medicine. Another speaker says Anthropic expects AI to cut the time needed for a clinical trial by 20% to 30%. That gap between the two numbers explains much of the reaction.

Where Amodei's prediction comes from

The idea is not new for Amodei. In October 2024 he published an essay called Machines of Loving Grace. In it he argued that powerful AI could compress 50 to 100 years of biology research into 5 to 10 years. He wrote that such AI could arrive as early as 2026.

In that essay he listed most infectious diseases, most cancers and Alzheimer's as targets. He also hedged heavily. Progress would depend on lab experiments, clinical trials and regulators, which AI cannot simply skip.

Anthropic has started building in that direction. It sells Claude to drug makers and research labs, and Bristol Myers Squibb appears as a customer in the roadshow video. We looked at what those tools have delivered so far in our piece on Anthropic Claude in drug development.

The difference this time is the setting. A prediction in a personal essay is one thing. The same prediction in a video made to sell shares drew a much sharper response.

How people reacted to the claim

Critics pounced within hours. Larry Dignan of Constellation Research used the word "absurd" in the heading on the claim. He wrote that "there's a certain amount of hubris with making those predictions in the first place" at all.

AI skeptic Gary Marcus went further on X and called Amodei a "scam artist" in a post. Writer Jesse Singal called it a "really terrible idea" to make such predictions if Amodei wants people to take AI seriously. Others focused on the timing, a week before the IPO prices.

Supporters answered that the goal is worth chasing. Futurum Group CEO Daniel Newman wrote that people "should absolutely be cheering for AI to solve all of these" on X. Analyst Ben Bajarin asked why anyone would not encourage the pursuit if it is even remotely possible.

Kimmonismus, an AI commentator on X, shared the clip with a note of support. He argued that the claim is more than IPO hype after the release of Claude Fable 5.2. He also pointed to a Bloomberg report that the IPO is already oversubscribed several times. We could not see that report ourselves.

Enthusiasm aside, the claim runs into how medicine actually moves.

Why a decade is a hard target

A new drug usually takes 10 to 15 years from discovery to approval. Most candidates fail in clinical trials, and trials cannot run faster than patients' diseases progress. Even a perfect drug designed by AI today would still need years of human testing.

Alzheimer's shows the problem. Decades of research have produced only a few drugs that slow the disease modestly, and none that stops it. Cancer is hundreds of different diseases with different biology. Mental illness is even harder to define in a way a single cure could address.

AI can speed up parts of that pipeline. It can screen molecules, design proteins, read papers and plan experiments. Anthropic's own 20% to 30% estimate for clinical trials is a realistic picture of what that looks like today.

The final question is what investors should take from it.

Will AI cure cancer and Alzheimer's within a decade

Almost certainly not in the sense of making them disappear. Amodei's line is a long-term vision dressed as a forecast. He framed it as his own opinion with the words "I think" in the middle. The line stayed out of the prospectus, the document Anthropic can be held to.

AI is likely to make some treatments arrive sooner, and that could still be worth a great deal. For investors in ANT, the useful numbers are elsewhere. Anthropic reported $17 billion in third-quarter revenue and targets an adjusted operating margin above 35% in the long run. The disease forecast is a sales pitch, and buyers should treat it as one.

FAQ

Here are short answers to the questions readers ask most about the claim.

What did Dario Amodei say about cancer?

In Anthropic's IPO roadshow video, Amodei said cancer, Alzheimer's, heart disease and mental illness could be "relics of the past" in less than a decade. He presented it as his own estimate.

Is the claim in Anthropic's IPO prospectus?

No. The prospectus makes no forecast about curing diseases. The founders' letter only says AI could bring cures for major diseases "within reach."

When does Anthropic go public?

Anthropic plans to price its IPO on November 18 and start trading on Nasdaq under the ticker ANT on November 19, according to CNBC. The price range was set at $70 to $78 a share.