Operating efficiency "leaves others in the dust"! OpenAI's computing power profit margin soars to 70%, supporting a valuation of $83 billion.

According to media reports, OpenAI's efficiency in artificial intelligence (AI) operations has continued to improve, with its computing power profit margin reaching 70% in October, significantly higher than the 52% at the end of last year, and double the around 35% in January 2024. This is an internal metric measuring the revenue share after deducting the cost of running models for paid users for enterprise and consumer-level products. This indicates that despite the high computing power costs preventing profitability, OpenAI is achieving higher revenue returns in every dollar invested in server operation costs to support ChatGPT subscription services and sell model usage rights to enterprise clients. According to an insider, the improvement in OpenAI's computing power profit margin can be attributed to two main factors: the continuous decrease in annual computing power leasing costs and the optimization and adjustment of AI models by the company to increase operational efficiency. In addition, OpenAI has also introduced higher-priced subscription service packages, resulting in increased revenue from certain customer groups. In contrast, media analysis shows that OpenAI's competitor Anthropic had a computing power profit margin of about -90% last year. Analysis suggests that Anthropic is expected to raise this profit margin to about 53% by the end of this year, and their most optimistic forecast indicates that this metric may reach 68% next year. However, overall, Anthropic's forecast data suggests that if the costs of running AI models for non-paying users and model training costs are taken into account, the company's comprehensive server operating efficiency will surpass that of OpenAI. Most people use the free version of ChatGPT, while Anthropic's chatbot Siasun Robot & Automation has a much smaller user base. Therefore, in the future, OpenAI will need to monetize these users through advertising placement or shopping rebates to make up for this efficiency gap. At the same time, OpenAI is also aggressively promoting its enterprise versions and paid software functions for industries such as financial services and education, competing with Alphabet Inc. Class C (GOOGL.US) and Anthropic in these areas. It is worth noting that at a time when OpenAI is in negotiations for a significant amount of financing, this good news about its computing power profit margin may help alleviate some investors' concerns about the company's profitability prospects. It is reported that OpenAI is planning to raise up to $100 billion in a new round of financing, which is currently in its early stages. If OpenAI can successfully raise the full target amount of funds, its company valuation could reach as high as $830 billion. Reports suggest that OpenAI plans to complete the latest round of financing as soon as the end of the first quarter of next year. Losing the Lead! OpenAI Shifts Strategic Focus Earlier this month, OpenAI CEO Sam Altman issued a "code red" internally, calling for the suspension of several sideline projects, including the Sora video generation project, for about eight weeks, to focus entirely on improving ChatGPT in response to increasing competition. This decision highlights a deeper ideological divide within OpenAI between "pursuing broad consumer appeal" and "achieving breakthrough research goals." Altman believes that in order to ensure OpenAI's survival, the company must prioritize user satisfaction above its original goal of achieving general artificial intelligence (AGI). The era of OpenAI's dominance in the field of artificial intelligence seems to be a thing of the past. Alphabet Inc. Class C (GOOGL.US) released its latest, self-proclaimed "smartest" AI model Gemini 3 last month. In early tests, Gemini 3 has surpassed ChatGPT, especially when compared to the GPT-5.1 model recently released by OpenAI. Gemini 3 has ascended to the top of the global AI model LMArena rankings with a historical high score of 1501, and achieved a 37.5% score in the Humanity's Last Exam benchmark test measuring general reasoning ability, surpassing the previous record of 31.64% held by GPT-5 Pro. Additionally, it has demonstrated a Ph.D.-level performance in multiple academic-grade benchmark tests, scoring 91.9% in the GPQA Diamond test, setting a new record of 23.4% in the MathArena Apex benchmark test in the field of mathematics, and achieving a score of 72.1% for factual accuracy in the SimpleQA Verified test, all significantly surpassing GPT-5.1. Altman internally warned employees that Alphabet Inc. Class C's strong return to the AI field could bring "short-term economic pressure" to OpenAI, and the company's CFO Sarah Friar also admitted to investors last month that growth in ChatGPT was slowing down. Another closely watched AI startup, Anthropic, has formed a value of $350 billion "strategic partnership" with OpenAI's main partners Microsoft Corporation (MSFT.US) and chip giant NVIDIA Corporation (NVDA.US). The tilting of key roles in the supply chain towards competitors signifies a potential weakening of OpenAI's advantage in computing resources, while Anthropic's research and commercialization capabilities will be significantly enhanced. For Altman, a particularly concerning signal is that the user gap is continuing to narrow. Alphabet Inc. Class C claims that its Gemini application has 650 million monthly active users. In contrast, Altman stated in October that ChatGPT was enjoying 800 million weekly active users. Although the statistical methods differ, the growth momentum of the competitors is evident. These events have made OpenAI, which rose to fame with ChatGPT nearly three years ago, appear more vulnerable in its industry dominance than ever before. With Alphabet Inc. Class C making measurable progress in model technology and Anthropic receiving support from giants, the pressure on OpenAI is intensifying. Some analysts predict that Alphabet Inc. Class C and Anthropic will continue to weaken OpenAI's lead. JonesTrading's Chief Market Strategist Mike O'Rourke said, "Given the scale and industry position of Alphabet Inc. Class C and its first-mover advantage in the search field, Gemini may grab market share, leading OpenAI and other companies to fall behind." AI field veteran scientist Gary Marcus has a sharper view, stating, "OpenAI has basically squandered its once-leading technological advantage. Alphabet Inc. Class C has caught up." Alphabet Inc. Class C's brief lead in app downloads highlights a key issue - in the AI product market, attracting users to try and cultivating usage habits are two different battles, with the latter being much more difficult. This AI competition is shifting from acquiring new users to competing for user time, which determines who can truly transform AI products into daily habits. Altman believes that the real competitors are not only Alphabet Inc. Class C, but also Apple Inc. (AAPL.US). As AI integration in consumer devices becomes increasingly critical, this is especially apparent. OpenAI has attracted a significant number of talents from Apple Inc., demonstrating its strategic layout in hardware capabilities, which may play a key role in future AI applications. Therefore, in the context of Alphabet Inc. Class C, Apple Inc., and other competitors rapidly advancing their AI technologies, the decision by OpenAI to prioritize improving ChatGPT over other projects is of great significance. This signifies that the company hopes to maintain competitiveness by improving ChatGPT and solidify market share in the intensifying competition of large models, in order to achieve more sustainable development. Slowing Commercialization Pace Puts Financial Sustainability under Pressure In order to defend its position in ChatGPT, Altman has decided to slow down other non-core projects and focus all efforts on ChatGPT. Among the suspended projects, the most notable is the advertising business, which has broad commercial prospects. Along with the advertising business, two other innovative projects were also put on hold: one is the AI Agent project that automates complex tasks such as shopping and health management, and the other is the Pulse feature that generates personalized news reports for users every day. These explorations for the future must make way for the current battle for survival. Slowing down commercialization means that it will take longer for OpenAI to achieve profitability. Despite never being profitable, OpenAI's valuation has skyrocketed like a rocket. In the eyes of the capital markets, a high valuation must be supported by matching and predictable revenue data. Financial data for the first half of 2025 shows that OpenAI's revenue reached $43 billion, a more than 200% increase year-on-year, impressing investors such as SoftBank and NVIDIA Corporation with its growth potential. However, at the same time, its net loss for the same period reached $13.5 billion, significantly higher than the $3.1 billion from the previous year. It has been disclosed that about 70% of OpenAI's "annual recurring revenue" comes from ChatGPT subscription fees. As a platform with over 800 million monthly active users globally, ChatGPT boasts a massive pool of traffic that any internet company would envy, but only 5% of these users are willing to pay for the service. The subscription fees paid by these monthly active users, which amount to $20 per month, may cover some operating expenses, but for the "gold-devouring beast" that is artificial intelligence, the cash flow from subscription fees is like a drop in the bucket. From hardware costs, exorbitant electricity bills, to the maintenance and human resource costs of research and development, each one is significant. Therefore, if OpenAI fails to present a clear and credible profit model, proving that it can not only change the world but also make a profit, this huge valuation "bubble" could burst at any time. The profit gap is at the core of the controversial "AI bubble theory" that has been so much talked about in the past. What is more concerning is that if OpenAI fails to present a credible profit model, a series of massive investments centered around the company may not yield a sufficient return, potentially leading to the bursting of the AI bubble. It is worth mentioning that HSBC Bank previously released a research report warning that even under the most optimistic growth assumptions, OpenAI would still not be able to achieve profitability by 2030. According to the bank's calculations, from 2025 to 2030, OpenAI's cloud and computing power costs will amount to a staggering $792 billion. By 2033, their total computing power commitment costs will soar to $1.4 trillion, with data center leasing alone accounting for a bill of $620 billion. The report predicts that by 2030, OpenAI's cumulative free cash flow will still be negative, with a funding gap of $207 billion, which must be filled through additional debt, equity financing, or more aggressive revenue-generating means. Against the backdrop of difficulty in monetizing through advertising and acquiring debt, the company's business model sustainability is facing challenges.