Anthropic advisers say slower model pace would not necessarily hurt finances

The people making the claim that a slower release pace is free are the ones advising the sale. Advisers believe even a slower pace of new model releases "would not necessarily have a major effect on Anthropic's financial prospects," people familiar with the IPO planning told the Wall Street Journal. Their reasoning: the revenue behind the run rate comes from Claude models already deployed ^[7][8]. Set that against the sequence: roughly $9 billion at the end of 2025 ^[17], about $14 billion in February and past $30 billion by April ^[18], past $47 billion in May ^[19], $65 billion in July on Bloomberg's confirmation ^[20], past $100 billion on Friday ^[1]. More than tenfold in a calendar year by the Times's count, and about eleven times on the two endpoints ^[21][1]. Thirty-five billion of that arrived in the last two months, about $17.5 billion of added annualized run rate a month ^[6]. The account does not link those steps to specific model releases.

The May round is the cleanest price anyone has actually paid: $65 billion in, at $965 billion post-money, against a run rate that had just passed $47 billion ^[19]. Call it 20 times ^[2]. On Friday's figure the same private mark is 9.7 times, and the money that went in bought about 6.7 percent of the company ^[2][3]. Investors now put the annualized rate at between $100 billion and $120 billion ^[22]. The floor of that forecast is where the company already is ^[5].

Pacing would cost something, though not in the revenue line. Amodei asked for third-party evaluators embedded inside frontier labs with employee-level access, voluntary coordination on safety standards among AI companies in democratic countries, and a push for coordination with authoritarian governments too ^[4]. The first of those puts outside staff inside the perimeter where a listed company keeps material non-public information. Elon Musk, whose SpaceX sells Anthropic compute capacity, posted "Dario is right" ^[16].

"I actually think that, given everything happening with safety, right now would be an ill-advised moment to go public, and we don't feel pressure on that," Altman told Fortune ^[14]. He also wrote on X that "I agree with Dario that we need to pace the frontier" ^[15].

In my view the essay costs Anthropic nothing in the quarter it is about to put in front of buyers. The postponement is the evidence: a company worried a pacing pledge would dent bookings does not volunteer another quarter of bookings before pricing. Two ways to be wrong. If the third quarter shows growth off that $17.5 billion a month, then $100 billion annualized was one moment on a curve and release cadence does matter ^[6]. The other is disclosure. Evan Hubinger, who leads alignment science at Anthropic, wrote that the company "does not yet have a plan to solve alignment for superintelligence and are not clearly on track to," and he put his own estimate of human extinction within the decade above 10 percent ^[12][11].