Five months after announcing a landmark $100 billion investment plan with OpenAI, Nvidia CEO Jensen Huang now says the eye-popping figure was “never a commitment,” revealing the fragile nature of AI industry partnerships and sending shockwaves through the technology sector. What began as a September 2025 announcement of what would have been the largest AI infrastructure project in history – requiring power equivalent to 10 nuclear reactors – has quietly fizzled out, exposing deeper tensions between the two AI giants and raising questions about the sustainability of current AI investment strategies.
The Deal That Wasn’t
In September 2025, Nvidia and OpenAI announced a letter of intent for Nvidia to invest up to $100 billion in OpenAI’s AI infrastructure, with both companies expecting to finalize details “in the coming weeks.” The project would have matched Nvidia’s total GPU shipments for the entire year, according to Huang’s initial comments to CNBC. “This is a giant project,” Huang said at the time, describing a plan that would deploy 10 gigawatts of Nvidia systems for OpenAI.
But five months later, no deal has closed. Huang has been walking back the number, telling reporters in Taiwan that the $100 billion was “never a commitment.” He clarified that OpenAI had invited Nvidia to invest “up to” that amount and that Nvidia would “invest one step at a time.” While Huang confirmed Nvidia will “definitely participate” in OpenAI’s latest funding round and called it “probably the largest investment we’ve ever made,” he explicitly denied it would reach $100 billion when pressed by reporters.
Underlying Tensions and Performance Concerns
The unraveling of the deal coincides with reports of deeper issues between the companies. According to Reuters, OpenAI has been quietly seeking alternatives to Nvidia chips since last year, with sources indicating dissatisfaction with the speed of some Nvidia chips for inference tasks – the process by which trained AI models generate responses to user queries. The performance limitations reportedly became apparent in OpenAI’s Codex, an AI code-generation tool, with OpenAI staff attributing some of its constraints to Nvidia’s GPU-based hardware.
A Wall Street Journal report added fuel to the fire, suggesting Nvidia insiders had expressed doubts about the transaction and that Huang had privately criticized what he described as a lack of discipline in OpenAI’s business approach. The Journal also reported Huang had expressed concern about the competition OpenAI faces from Google and Anthropic, though Huang later called those claims “nonsense” in comments to TechCrunch.
The Circular Investment Dilemma
Financial analysts have raised eyebrows at the nature of such investments in the AI sector. Bryn Talkington, managing partner at Requisite Capital Management, noted the circular nature to CNBC: “Nvidia invests $100 billion in OpenAI, which then OpenAI turns back and gives it back to Nvidia. I feel like this is going to be very virtuous for Jensen.”
Tech critic Ed Zitron has been particularly critical of Nvidia’s investment strategy, which touches dozens of tech companies, including major players and startups – all of whom are also Nvidia customers. “NVIDIA seeds companies and gives them the guaranteed contracts necessary to raise debt to buy GPUs from NVIDIA,” Zitron wrote on Bluesky last September. “Even though these companies are horribly unprofitable and will eventually die from a lack of any real demand.”
OpenAI’s Diversification Strategy
While publicly maintaining that “we love working with NVIDIA and they make the best AI chips in the world,” as CEO Sam Altman posted on X, OpenAI has been actively diversifying its hardware partnerships. The company announced a $10 billion deal with Cerebras in January, adding 750 megawatts of computing capacity for faster inference through 2028. Sachin Katti, who joined OpenAI from Intel in November to lead compute infrastructure, said the partnership adds “a dedicated low-latency inference solution” to OpenAI’s platform.
Beyond Cerebras, OpenAI struck an agreement with AMD in October for six gigawatts of GPUs and announced plans with Broadcom to develop a custom AI chip to reduce dependence on Nvidia. The company had also reportedly discussed working with startups Cerebras and Groq, but Nvidia’s $20 billion licensing deal with Groq in December – which included hiring Groq’s founder and CEO Jonathan Ross – effectively ended OpenAI’s talks with that company.
Broader Market Implications
The Nvidia-OpenAI saga reflects a larger trend in the AI industry where strategic partnerships are becoming increasingly complex and sometimes contradictory. Snowflake recently entered into a $200 million multi-year AI deal with OpenAI, giving its 12,600 customers access to OpenAI models across major cloud providers. This came just months after Snowflake announced a similar $200 million deal with Anthropic, OpenAI’s competitor. ServiceNow has followed a similar pattern, announcing multi-year deals with both OpenAI and Anthropic in January.
Baris Gultekin, Vice President of AI at Snowflake, explained this approach: “Our partnership with OpenAI is a multi-year commercial commitment focused on reliability, performance, and real customer usage. At the same time, we remain intentionally model-agnostic. Enterprises need choice, and we do not believe in locking customers into a single provider.”
Financial Fallout and Investor Concerns
The uncertainty surrounding the Nvidia-OpenAI deal has had tangible financial consequences. Nvidia shares fell about 1.1 percent following the reports, with Sarah Kunst, managing director at Cleo Capital, telling CNBC that the back-and-forth was unusual. “One of the things I did notice about Jensen Huang is that there wasn’t a strong ‘It will be $100 billion.’ It was, ‘It will be big. It will be our biggest investment ever.’ And so I do think there are some question marks there.”
The situation also highlights the massive financial commitments companies are making in the AI race. Oracle recently raised $25 billion in a bond offering despite investor concerns over its rising debt from AI infrastructure spending, including a $300 billion deal with OpenAI. Oracle’s net debt increased from $78 billion to $105 billion year-over-year, and its stock fell nearly 50% from its September peak.
What This Means for the AI Industry
The collapse of the Nvidia-OpenAI mega-deal serves as a reality check for an industry that has been operating on seemingly limitless optimism and investment. It reveals several critical insights: First, even the most promising partnerships between industry leaders can falter under the weight of performance concerns and strategic differences. Second, the circular nature of AI investments – where chipmakers invest in companies that then buy their chips – raises questions about sustainable business models. Third, diversification has become a survival strategy, with major AI companies actively working to reduce dependence on any single hardware provider.
As the AI industry matures, we’re likely to see more such recalibrations. The days of unconditional partnerships and blank-check investments may be giving way to more measured, performance-based collaborations. For businesses relying on AI technologies, this means paying closer attention to the stability of their providers’ partnerships and the diversity of their technology stacks. For investors, it suggests a need for more scrutiny of AI companies’ hardware dependencies and the sustainability of their growth strategies.
The Nvidia-OpenAI story isn’t just about a failed deal – it’s about an industry learning to balance ambition with pragmatism, innovation with sustainability, and partnership with independence. As Huang himself put it when discussing the investment: “We will invest one step at a time.” Perhaps that’s the approach the entire AI industry needs to adopt as it moves from explosive growth to sustainable development.