NextFin News - In a revealing segment on CNBC’s 'Power Lunch' on Monday, February 2, 2026, Dan Niles, the founder of Niles Investment Management, detailed a significant cooling in the relationship between Nvidia and OpenAI. According to Niles, the semiconductor giant has fundamentally altered its perspective on the AI startup that originally sparked the generative AI boom. The discussion highlighted a growing risk that the long-standing deal between the two entities could fall apart as their strategic interests begin to diverge in a maturing market.
The shift in sentiment comes at a critical juncture for the technology sector. While OpenAI remains a dominant force in large language models (LLMs), the cost of maintaining its lead has become a point of contention. Niles noted that Nvidia, led by CEO Jensen Huang, is increasingly viewing OpenAI not just as a primary customer, but as a potential competitor and a source of systemic risk due to its massive capital expenditure requirements. This change in stance is occurring against the backdrop of a broader market re-evaluation of AI valuations, where investors are demanding more tangible returns on the billions spent on H200 and Blackwell-series chips.
The friction is partly driven by OpenAI’s reported efforts to secure independent chip manufacturing capabilities—a move that directly threatens Nvidia’s market dominance. According to Niles, Nvidia is responding by prioritizing 'Sovereign AI' initiatives and enterprise partnerships that offer more stable, long-term growth. This pivot is supported by the current political climate; U.S. President Trump has recently emphasized the importance of American-led infrastructure, leading to a surge in domestic data center projects that do not rely solely on OpenAI’s ecosystem. By diversifying its portfolio, Nvidia is insulating itself from the potential 'AI bubble' that some analysts fear is concentrated within the high-burn-rate startup sector.
From a financial perspective, the data suggests a widening gap in operational efficiency. While Nvidia’s margins remain robust due to its software moat and CUDA platform, OpenAI’s path to profitability remains obscured by the sheer energy and hardware costs of training next-generation models. Niles pointed out that if OpenAI cannot demonstrate a sustainable business model by the end of 2026, Nvidia may further reduce its allocation of top-tier silicon to the firm in favor of cloud service providers (CSPs) like Microsoft and Amazon, who possess deeper pockets and more diversified revenue streams.
Looking ahead, the 'Nvidia-OpenAI' divorce, if it proceeds, could signal a new era of fragmented AI development. We are likely to see Nvidia transition from a hardware supplier to a full-stack AI infrastructure provider, competing directly with the software layers that OpenAI currently occupies. Meanwhile, OpenAI may be forced to accelerate its hardware independence, potentially through a massive joint venture with Middle Eastern sovereign wealth funds or domestic industrial partners. As U.S. President Trump’s administration continues to push for 'Compute First' policies, the battle for control over the AI stack will likely move from the laboratory to the foundry, with Nvidia positioning itself to remain the indispensable architect of the era, regardless of which model provider leads the pack.
Explore more exclusive insights at nextfin.ai.
