Anthropic's Poorly Timed Truth-Telling
Plus, venture returns under scrutiny & a rush of deals for space startups
The Week in Short
Anthropic CEO Dario Amodei backs off from harsh criticism of the Trump administration, even as Anthropic is designated a supply chain risk. VC returns — including from megafunds — aren’t measuring up to public markets. Space deals soar on fresh funding rounds. OpenAI launches new models and preps for an IPO while its shopping initiative stalls. Apple breaks with convention and brings out a mass-market laptop. Amazon data centers were taken offline by Iranian drone attacks. An AI startup is rocked with a co-founder lawsuit.
The Main Item
Anthropic Apologizes for Calling It Like It Is
Of course, what Dario Amodei said was true. It was just terrible timing.
On Wednesday, amidst an ongoing fight with the Department of Defense, The Information published a lengthy internal message written by Anthropic’s CEO that explained his view on bitter rival Sam Altman’s quick deal with the Trump administration.
Amodei wrote on Feb. 27:
The real reasons DoW and the Trump admin do not like us is that we haven’t donated to Trump (while OpenAI/Greg [Brockman, OpenAI’s president] have donated a lot), we haven’t given dictator-style praise to Trump (while Sam has), we have supported AI regulation which is against their agenda, we’ve told the truth about a number of AI policy issues (like job displacement), and we’ve actually held our red lines with integrity rather than colluding with them to produce “safety theater” for the benefit of employees (which, I absolutely swear to you, is what literally everyone at DoW, Palantir, our political consultants, etc, assumed was the problem we were trying to solve).
What he’s saying is obviously true.
CEOs have not been cozying up to Trump for their good health. They clearly recognize that it’s the price of doing business in America these days. (In our conversation with Trump loyalist Keith Rabois on YouTube the other week, he had to concede as much.)
Anthropic is perhaps the most visible and successful left-coded technology company in the whole country right now. (It turns out having consistent moral principles is not mutually exclusive with building a good product.)
Even if you’re rooting for Anthropic in its principled fight with the Department of Defense, the harsh wording of the memo was uncomfortable. It leaked at a moment when the conflict seemed like it would find some sort of off-ramp. Amodei’s comments, however honest they might have been, looked like a thumb in the administration’s eye, and their publication appears to have derailed negotiations.
Amodei has a careful dance to do here:
On the one hand, this fight has been a boon for Anthropic’s consumer business. The foundation model company has seen a surge of Claude signups and downloads as Trump-hating liberals rally around the one company that seems to have a backbone. (Anthropic’s Mike Krieger shared that “more than a million people are now signing up for Claude every day.”) And Anthropic has been winning over the hearts and minds of AI researchers, with OpenAI’s VP of Research announcing his defection to Anthropic this week. (It’s funny that Anthropic has created a sort of #DeleteUber moment for rival OpenAI thanks to this brouhaha with the Defense Department. Emil Michael, the former Uber executive turned Department of Defense leader at the center of the conflict, knows about the Delete Uber fiasco better than almost anyone. Sadly, Michael has been on the heel side of both conflicts.)
On the other hand, there’s a difference between a minor dust-up with the executive branch and an all-out war with a self-styled Department of War. Even Anthropic’s sympathizers don’t want to see it make clumsy moves in its fight here. The administration has a lot of different levers to make Anthropic’s life miserable. Even if you can’t come to a win-win deal, you want to give your counterparty an off-ramp to deescalate. The Department of Defense on Thursday said it had formally notified Anthropic that the company had been declared a supply chain risk.
On Thursday, we saw Amodei “apologize directly” for his post. He wrote, “Anthropic did not leak this post nor direct anyone else to do so — it is not in our interest to escalate this situation.”
Now Amodei needs to find a path that allows him to maintain his scruples and his new anti-Trump users while extricating the company from this conflict. A legal fight seems part of that strategy but it would certainly also be in Anthropic’s interest to escape the administration’s laser focus.
Amodei said in his latest memo that the supply-chain designation relies on a narrow statute that would only bar companies from using Anthropic technology specifically for military projects. And the company intends to fight it in court.
Legal analysts say Anthropic has a strong case. The respected legal blog Lawfare broke it all down under the headline “Pentagon’s Anthropic Designation Won’t Survive First Contact With the Legal System.”
Privately, most business leaders are not on board with the Trump administration’s vision of the relationship between the government and the private sector, as Reed Albergotti noted in Semafor. But Amodei is almost completely alone in daring to say so publicly — and even he felt the need to disclaim his sentiments.
If Anthropic can just weather the storm — and America remains a democracy — Anthropic could find itself on much better political footing in a few years. Trump is deeply unpopular. He has dragged America into a poorly-thought-out war (one Anthropic could have been tagged with if it hadn’t made a dramatic last-ditch stand on autonomous targeting). If Democrats take the midterms and someday the White House, Anthropic could end up one of the few companies in good standing when the political winds shift.