Anthropic announced on Wednesday that Claude, its AI model, spent 21 hours and roughly 210 million tokens searching DNA databases before flagging an unusual enzyme system that the company has named ART—a find it bills as a new molecular machine hidden in virus DNA. CRISPR co-inventor Feng Zhang reviewed the resulting pre-print and called the discovery “genuinely intriguing.” The findings were published as a pre-print, which has not yet been peer-reviewed. Pre-prints are a staple of modern science—rapid-release papers that circulate before a journal sends them to outside referees.

Even so, CEO Dario Amodei acknowledged in Anthropic’s press release that the system’s “precise function, biotechnological utility (if any), or level of significance is not yet clear.”

In other words: Claude found something. Nobody yet knows exactly what it does, whether it is useful, or whether it matters. Anthropic announced it anyway.

What the Model Found

The system turned up inside bacteriophages—the viruses that infect bacteria, and among the most abundant biological entities on Earth—sitting next to a repeating stretch of DNA that resembles a CRISPR array. CRISPR is the gene-editing tool that turned an obscure bacterial defense mechanism into a multi-billion-dollar medical industry, and it has since underpinned approved gene-editing therapies and a rapidly growing biotech sector.

Anthropic named its find ART, short for array-associated reverse transcriptases. Reverse transcriptases are enzymes that copy RNA into DNA, a trick several viruses and bacteria use to defend themselves.

Claude combed through more than 200,000 of them, narrowed the pile to 3,500 unusual candidates, and then to 20 judged worth closer study. The entire search took 21 hours, involved roughly 950 AI agents working in parallel, and consumed about 210 million tokens—the chunks of text an AI model reads and writes as it thinks.

“At minimum it is work I would have been proud to do as a PhD student,” Amodei said.

Critics Push Back

Not everyone accepted that comparison. Ravid Shwartz Ziv argued on X that the announcement oversold the achievement:

The sad thing is that Dario knows better. He was a PhD student. He knows the significance level of this result. He knows that if he had walked into Bill’s office (his advisor) with “we found an interesting system, but we still don’t know what it does” and said he was ready to…

— Ravid Shwartz Ziv (@ziv_ravid) September 23, 2026

Tyler Lockton broke down what the announcement actually described:

In plain English: what this was, and whether it was “bad.”You pointed at a tweet by Min Choi. That tweet was hyping an Anthropic announcement: their AI, Claude, helped spot a weird stretch of DNA in bacteriophages (viruses that infect bacteria). Next to a known kind of enzyme sat…

— Tyler Lockton (@Tlockty) September 23, 2026

Voices of Support

Anthropic also had plenty of onlookers cheering from the sidelines. Per the company’s release, Feng Zhang, the MIT and Broad Institute researcher who helped pioneer CRISPR, reviewed the pre-print and called it “genuinely intriguing.”

“This is an exciting example of how AI agents can contribute to biological discovery,” Zhang added. His remark lands in a field where machines already have a track record: DeepMind’s AlphaFold, whose protein-structure predictions earned a share of the 2024 Nobel Prize in Chemistry, showed that AI could crack biological problems that had resisted researchers for decades.

Owen Lewis, who writes the Techno-Optimist Substack, argued that the process behind the discovery may prove even more consequential than the discovery itself.

“The discovery itself is fascinating, but the process may be even more important. Claude did not perform the laboratory experiments (though it probably soon could),” Lewis wrote. “Human scientists reviewed its hypotheses, selected the most promising candidates, and carried out all of the wet lab work.”

According to Lewis’s post, the result is the first published by Anthropic’s new molecular biology (wet) lab, where scientists are using Claude to explore large biological datasets, search the scientific literature, and generate hypotheses.

Well that didn't take long. p(abundance) is increasing. Anthropic has published the first result from its new molecular biology (wet) lab, where its scientists are using Claude to explore large biological datasets, search the scientific literature, generate hypotheses, and…

— Owen Lewis (@is_OwenLewis) September 23, 2026

Markets Move Early

Wall Street did not wait for the caveats. Shares of gene-editing companies including CRISPR Therapeutics, Intellia Therapeutics, and Beam Therapeutics—firms whose pipelines are built on CRISPR-family editing tools—dipped on the news. The pullback came even though Anthropic itself emphasized the uncertainty around the find.

A Familiar Pattern

This is not the first Anthropic result to arrive with asterisks. In July, the company said Claude Mythos had cracked post-quantum cryptography that had stumped humans for years, as Decrypt previously reported. That claim held up—but only after hundreds of hours of human verification, and only on a signature scheme and a reduced-round cipher variant that nobody has ever deployed. A real result, heavily qualified—a pattern that, critics argue, is repeating with ART.

What is actually confirmed at this stage: an odd-looking protein pattern, one prominent outside scientist calling it “intriguing,” and no timeline for determining whether any of it matters. If ART eventually proves useful, it remains years away from touching an actual patient—and nobody at Anthropic is claiming otherwise.

Anthropic says laboratory work is already underway. The company has not said how long that work will take. Until that work produces results and the pre-print goes through peer review, ART remains exactly what Amodei described: a system whose significance is not yet clear.