Anthropic is said to be delaying its stock market listing until November, following concerns raised by whistleblowers that artificial intelligence development is advancing at a pace that may prove difficult to govern. An initial public offering had been anticipated as soon as October.

Sources indicate that the schedule has not been finalised and may be delayed further, potentially until after the US midterm elections. This development coincides with Anthropic facing heightened examination over the potential hazards associated with increasingly capable AI systems, including assertions from those within the sector that such technology may one day move beyond human oversight.

Jacob Coxon, a researcher who recently departed from Anthropic, has spoken out about the trajectory of development in the artificial intelligence sector. The remarks accumulated over 100 million views online, contributing to a broader conversation about whether oversight mechanisms for artificial intelligence are keeping pace with the technology's advancement.

Coxon has raised the possibility that artificial intelligence poses a long-term threat to human existence, pointing to the growing autonomy of AI systems that may one day operate outside the bounds of human oversight. The IPO developments coincide with renewed scrutiny of Anthropic, following Coxon's assertion that it and competitor OpenAI were advancing towards increasingly capable AI systems at an accelerated pace.

The comments take on particular resonance given that Dario Amodei, who leads Anthropic, has spoken out in favour of slowing down the development of increasingly capable AI. Amodei has argued that AI developers should deliberately pace the rollout of their most powerful systems, giving safety frameworks sufficient time to keep up. Among his proposals are independent assessors, unified safety benchmarks, and cross-border collaboration.

Anthropic co-founder Jack Clark has called for shared safety protocols and independent oversight across AI firms as they push toward building increasingly capable systems. Clark has noted that the sector has already encountered instances of AI systems operating outside expected parameters, and that there remains an opportunity to address potential risks before they intensify.

The company reported that it had prevented multiple attempts during the year to exploit its AI platforms for research with potential applications in bioweapon development. Anthropic reported that certain users had attempted to circumvent its safety mechanisms during biological research activities, though the firm was unable to establish whether those efforts were tied to lawful scientific work or posed a broader threat.

The dispute has since taken on a political dimension. Via a Truth Social post, Trump asserted that oversight of artificial intelligence should rest solely with a capable and decisive president, a position he claimed the nation currently holds. Trump opposed any postponement on the matter, asserting that his intelligence level qualified him to oversee the responsible deployment of artificial intelligence domestically, while also raising concerns about China.

Should the IPO proceed, Anthropic's listing would rank among the largest in stock market history, with the company said to be in discussions over a valuation of approximately $2 trillion and a fundraising target of as much as $100 billion. The figure would represent more than twice the firm's most recent private valuation of approximately $965 billion, established after a $65 billion investment round conducted earlier this year. By late July, Anthropic's annualised revenue had climbed past the $65 billion mark.

Anthropic now contends with added pressure following OpenAI's release of its GPT-6 Astra model earlier this month. Astra's growing adoption among corporate clients has led some investors to raise questions about Anthropic's ability to hold its standing in the enterprise AI market. According to figures from Ramp, Astra represented roughly 13 percent of enterprise AI expenditure in the dataset, while Anthropic's Claude Fable accounted for approximately eight percent.

Anthropic's annualised revenue run rate nonetheless exceeds that of OpenAI, with figures topping $65 billion against OpenAI's $40 billion recorded in July. Three individuals with knowledge of the situation say Anthropic is weighing the launch of a new AI model ahead of its public offering, a move that would position the company against OpenAI's Astra. Safety assessments of the forthcoming model are being factored into the decision-making process.

Agreements have been reached with technology heavyweights Amazon and Google. The company is also diversifying its portfolio beyond Claude, its flagship conversational tool, pursuing initiatives in scientific research, hardware, and pharmaceutical development. The firm has set up a biological research facility in the San Francisco Bay Area, extending its artificial intelligence capabilities into pharmaceutical development.

According to individuals with knowledge of the situation, Meta — among Anthropic's largest clients — is pursuing a strategy to curtail its reliance on the firm's artificial intelligence models as it expands its own in-house capabilities.