The "AI slowdown theory" cannot stop the financing frenzy of model giants! The advent of Astra sparks heated discussion on AGI, OpenAI aims for a $1.2 trillion valuation.

Title context: The "AI slowdown theory" cannot stop the financing frenzy of model giants! The advent of Astra sparks heated discussion on AGI, OpenAI aims for a $1.2 trillion valuation. Text: The developer of ChatGPTthat is, OpenAI, the global leader in AI applicationsis in preliminary talks with potential interested investors for a new round of financing, with a proposed valuation exceeding $1.2 trillion. Media reports citing people familiar with the matter said the talks were initiated by institutional investors, and whether to proceed depends on the timing of a listing, while OpenAI CEO Altman said the company will likely not go public until 2027. This also means that, before formally accepting public-market pricing, investors are still willing to fight for more OpenAI equity. The business logic supporting this interest is that the scope of work that advanced models can handle is expanding, opening a higher growth ceiling for subscription, enterprise services, and API revenue; the support for future valuation depends on how quickly this demand translates into revenue and cash flow. OpenAI considering a new financing round at a valuation that may exceed $1.2 trillion, along with the financing moves of the other two leading AI application playersAnthropic and DeepSeekfurther reflects capital's contest for model platforms. Anthropic announced in May that it completed $65 billion in financing, with a post-money valuation of $965 billion; afterward, media reported on September 11 that its potential IPO may be prepared at a valuation of about $2 trillion and raise up to about $100 billion. If it reaches that fundraising scale, it would exceed the historical record of $86.3 billion set by SpaceX in June. The $2 trillion market figure should be understood more as the potential IPO valuation in the report, not yet an achieved post-financing market value from a new round. As for DeepSeek, China's most cutting-edge large-model developer, media reported on September 9 that its ongoing financing is being negotiated at a pre-money valuation of about $71 billion, while the valuation in its previous first round was about $52 billion. Capital in different markets is making large bets on the growth prospects of AI applications, but the prices still under negotiation need to be verified by final transactions. Pre-IPO capital buildup: OpenAI prepares a new financing round, valuation may exceed $1.2 trillion OpenAI is in preliminary talks with investors about a new financing round that would value the ChatGPT developer at more than $1.2 trillion before an initial public offering. According to media reports, a person familiar with the talks who asked not to be identified said the decision on whether to proceed with the financing will depend on when OpenAI decides to go public. The person said the talks were initiated by investors. Such financing would pave the way for a long-awaited initial public offering. OpenAI CEO Sam Altman previously told Fortune that the listing plan is still moving forward but will not happen this year. At the same time, if the financing is completed at this valuation, OpenAI will surpass its main competitor Anthropic; the latter completed financing in May and was valued at $965 billion after including new investment. The Financial Times reported the talks earlier on Tuesday local time and noted that this financing round would allow long-time OpenAI backers to increase their investment exposure before the company's initial public offering. Anthropic is preparing its own initial public offering and has selected Nasdaq as the listing venue. Bloomberg previously reported that the developer of the Claude chat Siasun Robot&Automation is seeking to raise funds comparable to or even more than SpaceX in an initial public offering that could take place as early as October. SpaceX raised a record $86.3 billion in its June offering. OpenAI CEO Altman said in a recent interview with Fortune that the company will likely not go public until 2027. From the perspective of iterative updates in frontier technology and unit economics, the core reason OpenAI and Anthropic have attracted active institutional investment and produced a high-valuation trajectory lies in converting model capabilities into professional work that can be repeatedly delivered. Pretraining provides foundational capabilities, reinforcement learning and task evaluation improve reasoning and execution, and tool calling, context management, and result verification help models complete longer workflows. OpenAI's recent blockbuster launch of the Astra large model, which ignited heated discussion on "AGI," provides a concrete example: in the delayed simulation evaluation of OSWorld 2.0, OpenAI reported that Astra's task score was 72.6%, with each task taking about 40 minutes; the previous generation GPT-5.6 Sol scored 65.7% and took about 75 minutes. This is evidence of simultaneously improving completion performance and time efficiency under specific evaluation conditions. OpenAI's announced Astra results, which can be called nuclear-level, also include 98% on FrontierMath Level 4 and 99.9% on ARC-AGI-3, showing a leap in capability on specific highly difficult tests. This is also why Nvidia CEO Jensen Huang recently made a heavyweight statement on social media that the arrival of GPT-6 Astra means "the AGI era has arrived." Anthropic's latest blockbuster disclosure of a series of multi-agent research systems also reflects the engineering value of planning, parallel retrieval, and multi-round tool calling. What can be inferred from this is that if these advances continue to reduce rework and manual takeover in customer work, enterprises will have more reason to include AI in their daily budgets; the commercial space of model companies will also expand into high-value tasks such as software development, research, and professional services. The sudden suspension of Pro version subscriptions triggered by Astra simultaneously presents the extremely strong growth in commercial demand related to AI large models and the constraints on computing power supply. According to media reports, OpenAI product lead Tibo said Astra demand is "unprecedented," and the Pro tier puts the greatest pressure on the system; starting September 10, the company suspended new subscriptions and upgrades for the $200-per-month Pro 20x package, while existing subscriptions continue to renew normally. According to an analysis of the general technical mechanism of large-model services, complex agent tasks require repeatedly reading context, calling tools, generating plans, and verifying results: input processing increases computational load, long sessions and high concurrency increase KV cache usage, and the generation stage may also be limited by memory bandwidth. Therefore, expanding available computing power and optimizing scheduling are directly related to how much paid demand the platform can handle and at what cost it can complete tasks. Undoubtedly, the demand pressure brought by Astra strengthens OpenAI's commercial motivation to expand capacity; and the key to supporting a trillion-dollar valuation is continuously improving the delivery efficiency of successful tasks, so that growth in subscription, API, and enterprise revenue ultimately forms sustainable profit and cash flow. When a trillion-dollar valuation collides with a 5% yield, investment around the AI wave increasingly turns toward "delivery capability" At the same time, as the "AI slowdown theory" sweeps the market and the 10-year U.S. Treasury yield surges, U.S. stocks and even global stock markets have begun to compete over the ability to deliver profit growth under the AI boom. AI investment is showing a sharper divergence: capital continues to compete for the long-term growth opportunities of leading model platforms, while public markets are beginning to raise requirements for the speed of AI profit delivery at AI application-related companies, especially software-type listed companies. The "AI slowdown theory"global AI leaders such as Anthropic and OpenAI unanimously called over the weekend for slowing the pace of frontier AI large-model developmentcombined with the surge in long-term U.S. Treasury yields of 10 years and above, is jointly raising the valuation threshold for listed tech stocks. On September 12, Anthropic CEO Dario Amodei called for slowing the pace of frontier capability advancement so that safety measures can keep up, and Altman and others expressed support. On the first stock market trading day after Anthropic CEO Amodei made the heavyweight "slowdown theory" call, that is, September 14, "AI chip superpower" Nvidia's stock price fell about 3.4%, while the global semiconductor bellwetherthe Philadelphia Semiconductor Indexrarely fell sharply by about 6%, highlighting that the market is pricing in risks brought by factors such as the AI slowdown discussion. The market also began to reassess expectations for hyperscale training investment, model release pace, and future computing power procurement growth. On September 14, the Philadelphia Semiconductor Index fell sharply by about 6%; South Korea's KOSPI index, known as the global "AI computing power investment bellwether," fell more than 3% on Monday, and then on September 15 South Korea's KOSPI fell another 0.85%, closing at 6627.26 points, down for the fourth consecutive trading day. On September 15, the 10-year U.S. Treasury yield once again broke through 5%, touching its highest level since 2007. The 10-year U.S. Treasury yield, known as the "anchor of global asset pricing," rose intraday on September 14 to 5.012%, the highest intraday level since 2007, before falling back to 4.960%. Thereafter, as of the close of the U.S. stock market on September 15, the 10-year U.S. Treasury yield stood firmly above 5%, closing near 5.04%, continuing to rise to its highest level since 2007. The "anchor of global asset pricing" is an important reference for long-term U.S. dollar risk-free rates. It affects corporate financing costs and also affects the discount rate used when converting future profits into current value. Energy prices and inflation pressures are pushing rates higher, leaving tech stocks facing both rising funding costs and adjustments to long-term growth expectations. After the 10-year U.S. Treasury yield hit its highest level since 2007, long-term U.S. Treasuries are hovering at high yields, and the "AI slowdown theory" is suppressing market expectations for the expansion of computing power investment and profit growth of chip companies. However, the surge in AI application frequency and user-side computing power demand triggered by Astra also provides a positive signal for the commercialization of AI applications. Against this backdrop, OpenAI is still negotiating financing at a valuation of more than $1.2 trillion, which can be understood as some investors still being optimistic about the long-term space for advanced models to penetrate subscriptions, enterprise services, and professional workflows. The divergence in the capital market is focusing on how quickly AI commercialization can deliver revenue, and whether that revenue can support continuously expanding computing power investment.