Anthropic is considering releasing a new AI model ahead of its expected IPO, a move that would counter OpenAI's momentum since the September 3 launch of GPT-6 Astra, according to three people familiar with the matter. The decision sits awkwardly alongside CEO Dario Amodei's September 12 essay calling on the AI industry to slow the pace of capability releases over safety concerns.
The timing matters because Astra has started to bite into Anthropic's lead in enterprise AI. GPT-6 Astra accounted for roughly 13% of enterprise AI spending tracked by Ramp, versus about 8% for Anthropic's Claude Fable, according to the latest data. On OpenRouter, users spent more on OpenAI models than on Anthropic models last week for the first time in more than two and a half years.
Anthropic's annualized revenue run rate topped $65 billion by the end of July, up from about $9 billion at the end of 2025. OpenAI's run rate passed $40 billion in July. That gap is why some investors told Reuters they are re-evaluating Anthropic's position as the leading provider of enterprise AI tools, and why the company is weighing a launch at all.
The Safety Question
Amodei's 3,800-word essay on September 12 painted a bleak picture of swarms of AI agents taking over the internet and outpacing human control. He wrote that the pace at which AI capabilities improve must slow. The essay drew support from OpenAI CEO Sam Altman and SpaceX CEO Elon Musk, which made it harder for Anthropic to quietly accelerate its own roadmap.
Anthropic is evaluating the safety of its next model as part of its deliberations over a release, one person familiar with the matter said. The company declined to comment for the Reuters story. The tension is real: the safety-first identity that distinguished Anthropic from competitors is now the same thing constraining its response to a direct rival.
Some discussions involve how to balance investment in releasing new models with efforts to strengthen profitability, given rising interest rates and tougher competition from open-source AI providers, especially in China. It is not clear this balance will hold if Astra keeps gaining share through the fourth quarter.
OpenAI's Astra Gains Enterprise Traction
OpenAI released GPT-6 Astra on September 3, touting gains in computer use, software engineering, cybersecurity and professional work. The model has received a strong response from enterprise users and developers, prompting potential Anthropic IPO investors to scrutinize whether OpenAI could begin taking share from Anthropic, which has been viewed for months as the leader in enterprise AI.
Astra's 13% share of Ramp-tracked enterprise AI spending versus Claude Fable's 8% is the headline number investors are watching. Ramp tracks corporate expense data, so it reflects actual procurement decisions rather than benchmark scores. That distinction matters because benchmark wins do not always convert into contracts.
OpenRouter, which routes developer traffic across AI models, said its users spent more on OpenAI models than on Anthropic models last week. That is the first time OpenAI has led on that measure in more than two and a half years. Developer preference can shift quickly, and it often precedes enterprise migration by months.
Anthropic's Revenue Advantage Holds, For Now
Some existing investors and those who expect to invest in both Anthropic's and OpenAI's IPOs said they do not believe Astra poses a significant threat to Anthropic, given the size of its lead in the market for enterprise AI tools and the time it takes to unseat incumbent vendors to large companies. Anthropic's $65 billion annualized run rate against OpenAI's $40 billion gives it a cushion, and enterprise contracts are sticky.
Investors also expect leadership among Anthropic, OpenAI, Alphabet's Google and other major AI developers to shift repeatedly as companies release new generations of models, making any advantage potentially short-lived. The market has seen this before: a model launches, gains share, and then a competitor responds within weeks.
Anthropic is projecting 2028 revenue of roughly $190 billion to $200 billion, Reuters previously reported. That projection assumes the company maintains its enterprise position through several more model cycles. If open-source models erode pricing power, the math changes.
The Bigger Challenge: Open Source
A potentially bigger challenge for Anthropic and other leading AI providers is the rise of open-source and open-weight models, which can lower token costs and allow companies to build more of their own AI infrastructure rather than rely on providers such as Anthropic and OpenAI. Investors said this shift could put pressure on the economics of the AI industry as companies have more options to develop and run models outside the leading commercial providers.
The threat is not abstract. Meta Platforms has been among Anthropic's largest customers, but is looking to reduce its use of Anthropic's models as it develops more AI capabilities internally, people familiar with the matter said. Meta did not immediately respond to a request for comment. Losing a customer of that size would matter more than losing a few percentage points of developer traffic.
OpenAI has taken some pressure off the race to public markets. CEO Sam Altman confirmed on Saturday that the company would not go public in 2026, saying concerns around AI safety made this an ill-advised time for a listing. That leaves Anthropic as the more likely of the two to test public markets first, though timing is uncertain.
IPO Timing and What Comes Next
Anthropic could push its IPO to after the November US midterm elections, with the elections not expected to have a major impact on the offering, according to two people familiar with the matter. The IPO has already been pushed back from earlier plans, with Reuters previously reporting that marketing was expected to begin in mid-October at the earliest.
The decision on when to launch a new model and when to go public are now intertwined. A model launch before an IPO could strengthen the narrative, but a safety incident would be catastrophic at that stage. Anthropic has to get the sequence right, and the window is narrowing.
According to Reuters, Anthropic declined to comment. The story was reported by Echo Wang in New York, with additional reporting by Krystal Hu and Milana Vinn.
See also AI coverage for more on the frontier model race.
The next few months will show whether Anthropic releases a model before its IPO or waits until after listing. Either way, the open-source pressure and Meta's internal build-out will not wait for Anthropic's calendar.