Anthropic's alignment lead put the odds of AI wiping out humanity above one in ten.
Then the company quietly opened a biology lab that runs real experiments outside a computer screen.
The timing has insiders asking exactly what Anthropic is racing to build before everything crashes and its stock even trades.
The Same Company That Blocked Bird Flu Research Just Opened a Lab to Do It Themselves
Eric Kauderer-Abrams runs life sciences at Anthropic, and he confirmed the wet lab himself this month.
The lab sits in the Bay Area and runs real biological experiments, not simulations on a screen.
Kauderer-Abrams put it plainly: "We believe that to do biology, the final test is still, and will be for a while, in real lab work. We absolutely are doing that today."
This is the same company that shut down user accounts trying to study how bird flu could spread between mammals.
Company records show five separate misuse attempts between December 2025 and August 2026 tied to dangerous pathogens, including bird flu and viruses from the smallpox family.
In that same window, Anthropic loosened the biological safeguards built into its newest model, the one called Claude Fable 5.
Asked how the company squares building dangerous capability with warning everyone about that same danger, Kauderer-Abrams gave a two-word answer: "We constantly balance."
Investor Chamath Palihapitiya summed up the reaction in one line, needling the company behind "We're All Going To Die" and "Regulate Me Now" for building a wet lab in San Francisco.
Dario Amodei has called AI-assisted bioterrorism a "substantial risk" over the next two to three years, and his company just handed the model doing the touching looser guardrails than the one it replaced.
That is not a hypothetical for the rest of us, a lab full of real pathogens paired with a chatbot running looser safety limits sits several steps closer to your local pharmacy than most people realize.
A Two Trillion Dollar IPO Is Coming Right After the Doom Talk
Anthropic is on pace for more than one hundred billion dollars in annualized revenue by the end of the year.
The company is lining up a public offering that could value it near two trillion dollars, with shares potentially trading before December.
As former Financial Times and Politico journalist Izabella Kaminska wrote in a recent X post that AI industry figures’ whole strategy was to make themselves appear not just “Too Big to Fail” but “Too Dangerous to Fail.”
“I already explained in the Telegraph yesterday that I believe the AI pacing demand is mostly an attempt to hold the global economy hostage. The labs think they are now too big to fail, and are therefore entitled to an antitrust waiver for systemic security purposes,” Kaminska wrote. “However, it’s hard not to interpret the reveal about biolabs as a signal that they’re not just TOO BIG TO FAIL… but also TOO DANGEROUS TO FAIL. Aka handover the money or else.”
Anthropic acquired a stealth biotech startup called Coefficient Bio in April for roughly four hundred million dollars in stock.
The company signed research partnerships with Novo Nordisk, Bristol Myers Squibb, and Genentech, three of the largest drug makers on the planet.
None of that reads like a company bracing for the apocalypse it keeps describing in press releases, it reads like a biotech company building a product line and wrapping it in a doomsday narrative that keeps regulators and headlines pointed somewhere else.
Sam Altman runs the same act at OpenAI, calling for regulation with one hand while shipping new capabilities with the other, so Anthropic did not invent this game, it just added a lab coat to it.
A company selling fear about a technology only it can safely wield gets to help write the rules everyone else has to follow, and that is worth more to Anthropic right now than any warning label ever could be.
The bill comes due when regulators, or the public, finally ask why the loudest doomsayer in the industry needed looser safety rules and a working biology lab before its stock ever traded.
And with major AI-player Oracle – which basically underpins other AI companies – tanking hard on Thursday over a “force majeure” notice, the fear angle is looking more and more like a part of the backup play to suck up Americans’ tax and retirement dollars.
Sources:
- Ben Smith, "Anthropic Warns AI Could Kill Us All, Then Quietly Builds a Wet Lab," RedState, September 23, 2026.
- "AI Doomers at Work: Anthropic Opens 'Wet Lab' for Biology Experiments," Breitbart, September 19, 2026.
- "Anthropic's Wet Lab Raises Questions as AI Safety Warnings Mount," Legal Insurrection, September 2026.