openai

King Charles III of the United Kingdom will meet with leaders of major artificial intelligence companies in Scotland to discuss how to develop and deploy AI in a way that benefits society. Participants include leaders from Nvidia, Google, DeepMind, OpenAI, and Anthropic, as well as government officials.The purpose of this meeting is to explore how to leverage AI technology for the benefit of humanity, enhance community cohesion, and improve lives.

Anthropic, OpenAI, and Google are discussing the establishment of an AI industry standards organization. Anthropic CEO Dario Amodei has called for AI companies to coordinate on technical testing and audits. OpenAI founder Sam Altman expressed support for the establishment of testing and auditing organizations but believes that large AI labs should establish industry standards organizations on their own rather than relying on the U.S. government.

OpenAI CEO Sam Altman stated that the company will not go public this year, with the IPO potentially delayed until 2027. In an interview with Fortune, Altman mentioned that given the current circumstances surrounding safety issues, going public now is not a wise move, and the company feels no pressure in this regard.He added that the company has a lot of work to do, including meeting safety and alignment requirements, as well as promoting collaboration between the industry and the government. This statement comes as Anthropic CEO Dario Amodei publicly called for a slowdown in the AI race on Saturday, with both Altman and Elon Musk quickly expressing their agreement.

According to a report by Decrypt, OpenAI's latest flagship model GPT-6 Astra has seen a surge of user complaints on the X platform about its performance "getting dumber" just a week after its release. An anonymous developer, synthwavedd, posted that "Astra feels noticeably dumber today," joking that it was a "post-release lobotomy." Previously praising the model, developer Pranjal Paliwal stated after reviewing the code Astra wrote for him: "We don't have AGI; what we're encountering is a performance regression."Developer Pankaj Kumar listed the symptoms: faster responses with poorer quality, and he suspects OpenAI has lowered the "juice value" (the computational power invested by the model before answering). Salio and researcher Md Ismail Sojal compared Astra on the release day with the current version using the same prompts, both yielding worse results. T3Chat founder Theo believes Astra is simply less stable than Claude Fable, with users beginning to showcase poor results after the honeymoon period ended. There are also opinions pointing out that OpenAI's previous flagship GPT-5.6 Sol experienced a similar cycle in July, when OpenAI executive Tibo Sottiaux denied intentionally weakening the model but admitted the company had been experimenting with "reasoning effort" settings.OpenAI has not released a statement regarding Astra similar to that of Sol.

According to CNBC, Y Combinator CEO Garry Tan stated at the YC annual Demo Day that regarding the distillation of OpenAI and Anthropic models by Chinese companies (including DeepSeek, Moonshot AI, MiniMax, etc.), he chose to "do nothing" and suggested that regulators focus instead on maintaining a commercial balance between open-weight models and cutting-edge models.Tan also holds a reserved attitude towards the doomsday theory of AI, believing that cybersecurity is currently the most urgent risk, while the employment impact and economic transformation caused by AI will take decades to truly permeate society.

OpenAI CEO Sam Altman stated that he is considering slowing down the development pace of cutting-edge artificial intelligence and hopes other AI companies will follow suit.According to informed sources, Altman mentioned the possibility of adjusting the AI development pace during a company meeting, collaborating with other AI labs. Recently, concerns among AI company employees about the risks of advanced AI systems have heightened internal anxiety.

OpenAI launched ChatGPT for Financial Services on September 10, 2026, which is a ChatGPT Work experience tailored for financial services, combining built-in financial data with the reasoning capabilities of GPT-6 Astra to assist teams in conducting research, financial modeling, and customizing client materials. The product was shaped through design collaborations with Morgan Stanley and Evercore.The product includes data from providers such as Daloopa, PitchBook, LSEG News, and Crunchbase, covering information such as earnings call transcripts, financial statements, company fundamentals, and private companies, indexed and hosted by OpenAI with support for granular citations. The product natively offers models like GPT-6 Astra, supporting enterprise security and governance controls, allowing administrators to publish Excel, Word, and PowerPoint templates. Data is not used for training models by default and supports encryption, SSO, SCIM, and more. The product initially focuses on investment banking and equity research, available to eligible financial institutions.

OpenAI announced the launch of the Agents API.

Cryptocurrency exchange OKX announced the launch of Pre-IPO perpetual contracts related to OpenAI and Anthropic in Europe. Eligible traders can go long or short on the valuations of the two AI companies with up to 10 times leverage, but will not receive actual shares of the two companies. This product bets on the valuation changes of the two companies before their IPOs, providing a new speculative avenue for retail investors who find it difficult to directly buy shares of private companies.OKX is not the first platform to launch such products. Hyperliquid has offered markets related to OpenAI and Anthropic through HIP-3, and Binance also provides Pre-IPO perpetual contracts for the two companies. OKX stated that such products are not substitutes for the shares themselves but can meet the ongoing market demand for exposure to large tech companies that have not been publicly listed for a long time.Meanwhile, OKX has also launched 100 tokenized stocks and ETFs, including Nvidia, Alphabet, Palantir, as well as SPY and QQQ, allowing European traders to trade traditional markets around the clock. These tokens track the prices of the underlying securities but do not grant holders actual shares or voting rights, and some tokens can be withdrawn to self-custody wallets. OKX Europe CEO Erald Ghoos stated that the demand for its derivatives business continues to grow, with trading volume for X-Perps in the European market quadrupling since the transition period of MiCA ended in July.

According to Fox Business, the Trump administration is collaborating with OpenAI to use artificial intelligence tools.

User synthwavedd stated that they have learned that the Hodge Conjecture is very close to being verified by OpenAI, and that either OpenAI or Anthropic is also close to solving the Behr and Swinnerton-Dyer Conjecture. The user mentioned that the intensity of the behind-the-scenes competition for the "next" is a description they have never heard before.The user also claimed that if the above situation is true, it could mean that three Millennium Prize Problems might be solved within a month, but there are currently many rumors, and this matter may not be true.

Leifeng Network · Whale Rhino has learned that OpenAI is secretly developing a screenless AI hardware product in Shenzhen. Industry insiders say that the project's confidentiality level is extremely high.Engineers came to Shenzhen for a business trip without even bringing their phones. The product is planned to be launched in 2027.

David Sacks, the chair of the President's Council of Advisors on Science and Technology, stated that Anthropic's initial public offering must be paused until the allegations against the "whistleblower" are investigated.The "whistleblower" refers to Jacob Cox, a former researcher at Anthropic. He previously conducted pre-training research at OpenAI and joined Anthropic in July 2026, but resigned in September after only about 4 months, publicly warning on X that many people inside leading AI companies genuinely believe that this technology could lead to human extinction within a decade. Both companies are irresponsibly racing towards self-improving superintelligence, essentially gambling with the lives of all humanity.When he left, his equity had not yet vested, as companies typically require 6 months for vesting to begin, thus he forfeited potential substantial stock gains. This occurred at a time when Anthropic had secretly submitted its IPO application, with a valuation approaching one trillion dollars.

The research institution Sage Road Research released an executive summary of "The AI Trade," stating that since the beginning of this year, the Magnificent Seven has underperformed the Russell 3000 index by about 8 percentage points and the MSCI ACWI by nearly 9 percentage points. In July, the CBOE NDX volatility index relative to the VIX reached its highest point since the internet bubble, and after entering a correction, the Nasdaq index saw a 5% rebound over four days. As of the writing of the report, AI company stock prices have dropped 20% from their 52-week highs in June.Companies are struggling to achieve returns on investment amid soaring AI costs, with Uber, Amazon, Meta, and Walmart implementing restrictions on employee AI usage. Model homogenization limits pricing power, and Chinese open-source models have become a cheap alternative to OpenAI and Anthropic. AI capital expenditures have exceeded expectations, with the consensus for 2026 rising from $527 billion at the end of 2025 to about $800 billion by mid-year. Capital expenditures for hyperscale cloud providers in 2027 are expected to account for 3% of U.S. GDP, more than double the peak of 1.2% during the late 1990s telecom fiber construction. Allianz Research calculates that there is nearly a 46% growth gap between AI investment and sales, worse than the 32% during the 2001 telecom bubble.As of June, hyperscale cloud providers and related entities like Nvidia issued $225 billion in bonds, a year-on-year increase of 973.7%. The off-balance-sheet liabilities of tech giants have increased eightfold over four years to $1.65 trillion.

On September 9, Ara Kharazian, Chief Economist of the corporate spending management platform Ramp, released the Ramp AI Index for September 2026. This index tracks the AI usage of U.S. enterprises based on Ramp's spending data. In August, Anthropic held a share of 43.8% among U.S. enterprises paying for subscriptions or tokens, an increase of 0.34 percentage points month-over-month; OpenAI accounted for 39.8%, with a month-over-month increase of 0.09 percentage points. Overall, the adoption of AI continues to grow, but the growth rate is slowing, with technology sectors such as information, finance, and professional services leading.The median monthly AI spending per capita for the top 1% of enterprises decreased by 9.7%, from $7,976 to $7,205. The effective price per million tokens dropped by 41% from the peak of $1.15 in March 2026 to $0.68. The usage is driven by cheaper standard models, with cutting-edge models (Opus, Fable, Sol) accounting for 45% of the token share. The adoption of open-source models remains limited, representing 6.4% of AI spending enterprises and 3.6% of all enterprises.

OpenAI is pushing for mandatory national-level artificial intelligence safety requirements.

According to the New York Times, OpenAI expects its computing power expenditure to reach $750 billion by 2030.

According to a report by the Financial Associated Press, the Chief Financial Officer of OpenAI stated that they are considering reforming the charging model, testing a new pricing model based on performance, gradually moving away from the usage-based billing method, and competing with lower-priced open-source models.

Documents obtained by The Intercept through a lawsuit under the Freedom of Information Act show that the U.S. Department of Defense once requested OpenAI to provide a version of artificial intelligence specifically tailored for the military, with a "very low refusal rate" for commands. This statement appeared in the updated contract for the 2025 prototype agreement "P00003," which is a two-year agreement worth up to $200 million, covering logistics, intelligence decision-making, and operational uses. OpenAI, Google, xAI, and Anthropic all agreed to develop such militarized prototypes in 2025.OpenAI spokesperson Nate Evans stated that the company never agreed to contract terms that included a "very low refusal rate," and the relevant documents were early drafts from the Department of Defense, with the final executed contract not containing that statement. Department of Defense spokesperson Jacob Bliss also indicated that the phrase did not appear in any current contracts. The Intercept pointed out that the documents obtained were not marked as drafts; a Department of Justice lawyer representing the Pentagon once confirmed that the document was a signed executed version, but later requested to disregard that confirmation and stated that further investigation was needed.The latest contract signed on February 27 allows OpenAI services to be deployed on military classified networks, with similar delivery terms completely redacted. OpenAI stated that it has drawn a red line regarding autonomous killing and surveillance of Americans, but the contract text allows the government to determine any use as legal.