Anthropic is weighing the release of a new AI model as it seeks to keep pace with OpenAI following the launch of GPT-6 Astra, according to Reuters citing three sources. The possible rollout comes as Anthropic prepares for an anticipated IPO and amid growing competitive pressure from its rival.
The move would come despite CEO Dario Amodei recently urging the AI industry to rein in the speed at which new capabilities are developed and released, citing concerns over AI safety, as per them.
“We must slow the pace at which we improve the capabilities of AI models,” Amodei wrote in a 3,800-word essay published on September 12.
In the essay, Amodei warned of a potentially bleak future in which large numbers of AI agents could dominate the internet and develop faster than humans can maintain control over them. His concerns received backing from OpenAI CEO Sam Altman and SpaceX CEO Elon Musk.
OpenAI’s GPT-6 Astra, which was launched earlier this month, has seen growing adoption among businesses. This has raised concerns among some investors, who told Reuters they are reassessing Anthropic’s standing as a leading provider of AI tools for enterprise customers.
As part of its discussions about whether to launch its next model, Anthropic is also assessing its safety, according to the report.
People familiar with the company’s thinking stated Anthropic is weighing how much to spend on bringing new models to market while also improving profitability. Higher interest rates have made investors increasingly concerned about when companies are likely to begin generating profits.
The spending debate reflects the mounting pressure on major AI companies to demonstrate sustainable cash generation sooner. The pressure is being driven by higher borrowing costs as well as intensifying competition from open-source AI models, particularly those developed in China.
Anthropic declined to provide a comment, the report noted.
A possible release of the new model could put Anthropic’s ability to protect its position in the enterprise AI market against OpenAI to the test, while also preserving the safety-focused approach that has set the company apart from its rivals.
OpenAI launched GPT-6 Astra on September 3, highlighting improvements in areas including computer operation, software development, cybersecurity and other professional tasks. The model has drawn a strong response from businesses and developers, leading potential investors in Anthropic’s IPO to assess whether OpenAI could start eroding its rival’s market share after Anthropic was widely regarded as the enterprise AI leader for much of the past several months.
Data from corporate expense management platform Ramp showed Astra represented roughly 13% of enterprise AI spending, compared with around 8% for Anthropic’s Claude Fable.
OpenAI has also gained an edge on OpenRouter, a popular platform that directs developer requests across different AI models. OpenRouter said its users spent more on OpenAI models than Anthropic’s last week, marking the first time OpenAI had surpassed Anthropic on the metric in more than two and a half years.
Existing investors and prospective investors planning to participate in both Anthropic’s and OpenAI’s IPOs said they do not view Astra as an immediate major threat to Anthropic. They pointed to Anthropic’s sizeable lead in enterprise AI and the difficulty of persuading large businesses to switch away from established technology providers.
Anthropic’s annualized revenue run rate exceeded $65 billion by the end of July, rising sharply from around $9 billion at the end of 2025. The company is also forecasting revenue of approximately $190 billion to $200 billion in 2028, Reuters has previously reported. OpenAI’s annualized revenue run rate surpassed $40 billion in July.
Investors also expect the competitive pecking order among Anthropic, OpenAI, Alphabet’s Google and other leading AI developers to change frequently as new generations of models are launched. That means any company’s lead could prove temporary.
Garvit Bhirani is a journalist based in Gurugram. He is a Deputy Chief Content Producer at LiveMint, where he covers national and international news stories, focusing on accuracy and compelling storytelling for readers. <br><br> With a total of six years of experience in journalism, he has previously worked with Vaco Binary Semantics for Google, taking on the role of news curation lead, and reported from the field on health, education, and agriculture stories for 101reporters and News9. He has also served as a content editor for entertainment and news media organisations. <br><br> Garvit holds bachelor’s and master’s degrees in journalism and mass communication from Guru Gobind Singh Indraprastha University and Gurugram University, respectively. During college days, he joined India’s only non-profit student journalism network, where he anchored daily news updates and produced his own weekly show called ‘Data Fix’. <br><br> He was selected for the YES Foundation Media for Social Change Fellowship in Delhi, the Talking Data to the Fourth Pillar residential workshop, and the VOICE Fellowship in Pune. <br><br> He holds certificates in COVID-19-verification reporting, data journalism, food & agriculture, tech policy, media literacy and countering misinformation, and tackling election disinformation courses from Thomson Foundation, IndiaSpend, The Dialogue, US Mission in India, and AFP. <br><br> He can be reached on <a href="https://www.linkedin.com/in/garvit-bhirani">LinkedIn</a> or on <a href="https://x.com/GarvitBhirani">@garvitbhirani</a> on X
Catch all the Business News, Market News, Breaking News Events and Latest News Updates on Live Mint. Download The Mint News App to get Daily Market Updates.
Oops! Looks like you have exceeded the limit to bookmark the image. Remove some to bookmark this image.
Facts Only
* Anthropic is considering the release of a new AI model.
* OpenAI launched GPT-6 Astra on September 3.
* Anthropic CEO Dario Amodei published an essay on September 12 urging a slower pace of AI capability development.
* GPT-6 Astra represents approximately 13% of enterprise AI spending, while Anthropic’s Claude Fable represents approximately 8%.
* OpenAI expenditures on OpenRouter surpassed Anthropic expenditures last week for the first time in over two and a half years.
* Anthropic's annualized revenue run rate was over $65 billion by the end of July.
* OpenAI's annualized revenue run rate surpassed $40 billion in July.
* Anthropic forecasts revenue between $190 billion and $200 billion in 2028.
* Anthropic is preparing for an initial public offering (IPO).
* OpenAI is preparing for an initial public offering (IPO).
Executive Summary
Anthropic is currently balancing a strategic tension between its stated commitment to AI safety and the commercial necessity of competing with OpenAI. Following the September 3 launch of GPT-6 Astra, which has seen strong enterprise adoption and captured a larger share of spending via platforms like Ramp and OpenRouter, Anthropic is weighing the release of its own new model. This internal debate occurs despite CEO Dario Amodei's public advocacy for slowing the pace of AI development to prevent loss of human control over AI agents.
The pressure to launch is amplified by preparations for an IPO and investor concerns regarding profitability amid higher interest rates and competition from open-source models, particularly from China. While some investors worry about OpenAI eroding Anthropic's enterprise market share, others maintain that Anthropic’s existing lead and the high friction of switching providers protect its position. The competitive landscape remains fluid, with revenue run rates for both companies in the tens of billions and leadership likely to shift with each new model generation.
Full Take
The strongest version of this narrative depicts a clash between ideological safety commitments and the relentless gravity of venture-backed capitalism. It presents Anthropic as a company attempting to maintain a "safety-first" brand identity while operating within a market structure that penalizes hesitation.
The narrative relies heavily on a perceived tension between CEO Dario Amodei’s philosophical warnings and the company's corporate actions. By juxtaposing a 3,800-word essay on safety with the pragmatic pressures of an IPO and revenue targets, the situation is framed as a conflict of interest. However, the primary evidence for "market erosion" is based on specific snapshots from third-party platforms (Ramp and OpenRouter), which are then used to justify the necessity of a new product launch.
Patterns detected: none
The root cause is the "AI Arms Race" paradigm, where the fear of obsolescence overrides the precautionary principle. This echoes the historical pattern of "regulatory capture" or "safety-washing," where companies advocate for caution publicly while accelerating development privately to ensure they are the ones defining the safety standards.
The implication is a steady erosion of human agency; when the driver of innovation is investor ROI and competitive parity rather than societal benefit or risk mitigation, safety becomes a marketing differentiator rather than a hard constraint.
Bridge Questions:
1. If "safety" is a primary corporate value, what specific, measurable thresholds must a model meet before release, regardless of a competitor's launch?
2. To what extent do revenue run rates in the billions actually correlate with the long-term utility or stability of the AI tools provided?
Counterstrike Scan: A coordinated campaign to destabilize Anthropic's valuation would highlight the hypocrisy between its CEO's safety rhetoric and its commercial aggression to trigger investor doubt. The current content does not match this pattern; it reports the tension as a business reality rather than a moral failing.