A Wake-Up Call: Nvidia’s Reappraisal of OpenAI—Why Microsoft Should Be the Most Concerned

Nvidia, OpenAI, and Microsoft form an AI collaboration chain, which is showing signs of loosening due to differences in "execution" and "business discipline." On Wednesday, Bloomberg technology writer Parmy Olson stated in her latest column that Jensen Huang’s hesitation towards OpenAI is not just a shift in investment direction, but more like a reminder to Microsoft that possessing exclusive models and intellectual property does not automatically translate into product advantage; market competition is shifting from “whose model is stronger” to “who can deliver better.” Nvidia CEO Jensen Huang recently told industry insiders that his previously announced $100 billion investment in OpenAI is “not binding,” and privately criticized OpenAI for lacking business discipline. Huang denied dissatisfaction with OpenAI and told reporters last Saturday, “We’ll invest a lot of money,” but he also left room for commitment. Parmy Olson believes that this investment, originally linked to infrastructure expansion, is now more likely to appear in OpenAI’s current fundraising process as “tens of billions of dollars” and may happen before a potential IPO. The report also says OpenAI is discussing about $100 billion in funding with Nvidia, Microsoft, and Amazon, which is separate from the previously proposed infrastructure deal. The more direct market impact falls on Microsoft. Although Microsoft, through a restructuring agreement announced last October, secured exclusive access to OpenAI’s intellectual property and models until 2032, and will integrate Copilot with OpenAI’s latest GPT-5 model in August 2025, Copilot’s product experience continues to be criticized by users and lags behind competitors in functionality, highlighting the contradiction of “having top-tier models but struggling to realize the advantage.” Jensen Huang’s “Buyer’s Remorse” In Parmy Olson’s view, the AI industry has long been numb to massive bets, so Huang’s caution regarding his commitment is especially notable. The Wall Street Journal mentioned that Huang not only emphasized the non-binding nature of the investment but also privately questioned OpenAI’s business and execution discipline, injecting uncertainty into OpenAI’s fundraising narrative and partner confidence. For investors, this signal has two layers: First, the capital and computing power plans around OpenAI may still be rapidly changing; second, even one of the core industry partners is re-evaluating exposure to risk and return paths. Parmy Olson pointed out that Sam Altman’s management style continues to cause unease, including the dramatic firing and return at the end of 2023, and a series of complex, high-value deal arrangements. She wrote that these arrangements have saddled OpenAI with up to $1.4 trillion in computing resource commitments, which is about 100 times OpenAI’s projected 2025 revenue. On the product side, OpenAI’s progress is equally rushed. Parmy Olson cited that OpenAI tried to create a developer marketplace with the GPT Store and custom GPTs, but “stalled” due to a lack of clear strategy. Between “continuously leading model capabilities” and “commercialization and organizational volatility,” OpenAI’s partners need to price in uncertainty. Microsoft’s Execution Dilemma Microsoft’s capital returns are extremely impressive. Parmy Olson wrote that Microsoft’s early $13 billion investment in OpenAI equates to about 27% equity, valuing that stake at around $135 billion—more than tenfold the original investment; additionally, the restructuring agreement gives Microsoft cleaner exclusive access to IP and models before 2032. But on the product level, Microsoft has not fully realized this advantage. Parmy Olson directly points out the key issue: If OpenAI can continuously produce one of the world’s strongest AI models, why do Microsoft’s flagship Copilot products—built on OpenAI’s technology—still lag behind competition? She wrote that user feedback on Copilot centers on it being “confusing, limited, and difficult to use,” highlighting the gap between model quality and product delivery. Competitive pressure is erupting on the application side. Parmy Olson wrote that last month, Anthropic launched Claude Cowork, an app reportedly built with its own AI coding tool in ten days; after being authorized, it can operate PCs, organize files, generate PowerPoint and Excel from documents, and reply to LinkedIn messages. By comparison, she pointed out, even though Microsoft owns Windows, Office, and LinkedIn, Copilot cannot achieve these capabilities. This gap makes the market more likely to attribute the issue to Microsoft’s internal research and productization mechanism, rather than just the models themselves. Parmy Olson cited David Rainville, head of Sycomore Sustainable Tech, saying that some industry observers are watching Microsoft AI project leadership closely. He commented that if Microsoft cannot launch a product comparable to Claude Cowork in the next six months, “someone will have to step down,” and said Microsoft has a clear disconnect between “model quality” and “execution ability.” Risk Disclaimer The market carries risk; investments should be made cautiously. This article does not constitute personal investment advice, nor does it consider an individual user’s specific investment goals, financial situation, or needs. Users should consider whether any opinions, perspectives, or conclusions in this article fit their particular circumstances. Investing based on the article is at your own risk.