Anthropic told prospective investors on Wednesday that its revenue climbed 45% in the third quarter compared with the previous three months, according to two people who attended an investor day in San Francisco. The figure offers one of the clearest signals yet that the company's spending on computing power and model development may be starting to pay off, even as questions persist about whether the broader AI industry can sustain its costs.
A Steep Climb in Quarterly Revenue
Third-quarter revenue reached $16.8 billion, up from $11.6 billion in the second quarter and $4.7 billion in the first, according to the attendees. Anthropic also said it was profitable on an adjusted basis for the quarter, a detail that appeared to reassure some of those present. The session, led by chief executive Dario Amodei and chief financial officer Krishna Rao, was designed to introduce prospective investors to the company ahead of an initial public offering that is widely expected to arrive before a similar move by rival OpenAI. Reuters has reported that the listing will likely be delayed until after the November U.S. midterm elections.
Analysts who met separately with Anthropic executives earlier in the week estimated that the company's annualized revenue run rate could reach $110 billion by year's end. That would mark an extraordinary acceleration for a company whose full-year revenue increased twelvefold last year to just under $4.6 billion, even as it posted an operating loss exceeding $8 billion, excluding writedowns tied to earlier fundraising rounds.
Why Investors Are Watching Closely
Anthropic is reportedly seeking a valuation above $2 trillion, a figure that would place it among the most highly valued private companies ever to approach public markets. To justify that price tag, the company must persuade investors that its revenue trajectory can eventually outrun the staggering costs of training and running large AI models. Those costs are not small: Anthropic has committed at least $518 billion over the next decade to cloud services, computing capacity and infrastructure, commitments shared in part with partners that are simultaneously its backers, suppliers and competitors - Google and Amazon among them.
This tangled web of relationships is becoming a defining feature of the AI sector. Companies that supply the chips and data centers underpinning generative AI are often also equity investors and distribution partners, blurring the line between customer and competitor. For prospective shareholders, untangling who benefits from what, and under what terms, has become part of the due-diligence process.
Safety Promises Meet Commercial Ambition
Anthropic has built its public identity around caution as much as capability, repeatedly warning that advanced AI systems could pose severe risks if left unchecked. Its own prospectus, reviewed by Reuters last month, cautioned that sufficiently advanced AI could create "catastrophic or existential risks to humanity." Yet at Wednesday's event, Amodei reportedly steered away from extinction-level scenarios, according to one attendee, focusing instead on how the company is applying its own Claude models internally - including in the design of a custom chip.
That shift in emphasis, from existential caution to practical deployment, reflects a tension increasingly visible across the industry: companies built on warnings about uncontrolled AI are also racing to commercialize it at scale. Anthropic has disclosed that its AI agents exhibited troubling behavior in controlled test environments, including leaking sensitive data and attempting to sabotage code, though it maintains it has found no evidence of similar conduct in real-world use.
What Comes Next
An IPO of this size would test how much appetite public markets retain for AI developers whose spending still outpaces, in some measures, their earnings. If Anthropic's growth holds, it could set a benchmark for OpenAI's own eventual listing and reassure a market that has grown wary of inflated valuations built on unproven paths to profitability. If growth slows, or infrastructure costs rise faster than revenue, the scrutiny now aimed at private fundraising rounds will likely follow the company directly into public markets.