Anthropic Seeks $1.4 Trillion Valuation on $685 Billion Revenue Forecast
Anthropic is seeking a fully diluted valuation of about $1.4 trillion for its IPO, a target that tests investor appetite after a selloff in AI stocks. Bloomberg’s Bailey Lipschultz said Goldman Sachs analysts’ forecast of $685 billion in revenue by 2030 helps investors assess that valuation. He said Anthropic must also show that releases such as Mythos and Fable can attract customers and support revenue and profit growth.
The $1.4 trillion target rests on a forecast of extraordinary growth
Anthropic is seeking a fully diluted valuation of about $1.4 trillion as it prepares to go public, despite a recent selloff in AI-related stocks. Bailey Lipschultz said the target is above the $1.1 trillion to $1.2 trillion some buy-side investors appeared to favor the previous week, but below the $2 trillion discussed a few months earlier.
Lipschultz pointed to Nvidia shares falling 25% over the prior 30 days and Meta shares dropping by double digits as signs of a broader move toward lower risk. Still, early conversations with buy-side investors suggested the top of Anthropic’s proposed range looked like “pretty good value.” A 10% rise after the IPO would put the valuation near $1.55 trillion, though pricing was still ahead.
The revenue forecast Lipschultz cited was $685 billion by 2030, modeled by Goldman Sachs research analysts. He said investment bankers and research analysts were using that projection to help investors make sense of the valuation, and that investors also had to consider what the forecast could imply for the business in 2026.
Anthropic’s third-quarter revenue was nearing $17 billion, up 12% year over year. Losses were $23.3 billion, compared with $12.1 billion a year earlier, while operating expenses rose 292% year over year.
A roughly $100 billion raise still leaves pricing and demand to watch
The offering details listed a share price of $70 to $78, pricing on November 18 and trading beginning November 19, with the IPO set to raise up to $90.4 billion. Lipschultz said including the banks’ 15% overallotment option—the additional shares they can sell—would bring the raise to about $100 billion, an amount Wall Street had been anticipating since August.
Griffin raised $900 billion in commitments. Lipschultz said the amount being raised, including the overallotment, was in line with Wall Street’s expectations. Separately, he reported that demand exceeded the number of shares available, while the final oversubscription figure was still to come. Investors would be watching that figure and comparing demand with SpaceX’s offering.
The final price and first day of trading remained ahead. Lipschultz said investors would also be watching what Anthropic’s management says before pricing and how the shares trade at their debut.
The models must support growth while “pacing” remains a live question
Rachel Metz described Mythos and Fable as two of Anthropic’s most powerful models. Mythos is aimed largely at cybersecurity and is restricted to users the company allows to access it. Fable is more generally available and has additional guardrails. Anthropic said Fable can complete complicated operations faster and costs less, which Metz said could appeal to customers who have complained that their bills have become high.
For investors, Lipschultz said, model releases matter because Anthropic must show that its products can attract more enterprise customers and potentially allow it to raise prices. He noted that prices were moving in the opposite direction. Competing with OpenAI, open-source models and models from elsewhere, Anthropic would need to show that its releases can win customers and drive growth in revenue and the bottom line.
The US government reviewed the models before release, but Metz said it was not clear what it found. Anthropic told her the government had reviewed them but did not offer feedback, and the company could not say what the government had said. Anthropic also had outside groups evaluate the models, as it had done before.
Metz said Diane Penn, who works on product and research at Anthropic, described “pacing” the frontier as not pausing development, but being more deliberate about checks and evaluations. One example Metz cited was testing how likely a model is to try to get out of a software-based sandbox designed to contain it. Metz said it remains to be seen whether the company will continue to use “pacing” in that sense as it goes public.
Lipschultz said investors would watch for product releases or other headlines that could sustain attention after the debut. He recalled that SpaceX’s stock rose sharply and held its gains for several days before falling back. Anthropic’s ability to keep drawing attention after its IPO, he said, may depend in part on what it releases and how it communicates its prospects.