Nvidia chief executive Jensen Huang has cast doubt on the possibility of investing $100 billion (£75bn) in OpenAI, the San Francisco-based artificial intelligence startup that has become a household name thanks to the explosive popularity of ChatGPT. Speaking at a Morgan Stanley conference, Huang stated that such a massive financial commitment is “probably not in the cards,” even as Nvidia recently participated in a smaller $30 billion funding round for the company.
“I think the opportunity to invest $100 billion in OpenAI is probably not in the cards,” Huang said, adding that because of OpenAI’s anticipated initial public offering (IPO) later this year, “this might be the last time we’ll have the opportunity to invest in a consequential company like this.” The comments underscore evolving dynamics between Nvidia, the world’s leading designer of graphics processing units (GPUs) crucial for training large language models, and OpenAI, the organization that ignited the current generative AI frenzy.
Huang’s remarks come amid months of speculation about the two companies’ relationship. In September 2024, Nvidia had initially announced plans to invest up to $100 billion into OpenAI over several years, with tranches tied to OpenAI’s successive deployments of Nvidia’s chips in its data centers. That deal, however, was never finalized and reportedly stalled in January 2025. The earlier announcement had spurred further similar investment moves across the tech industry and temporarily boosted stock prices, but the underlying economics of the AI boom have shifted dramatically since then.
Optimistic projections from last year have given way to the stark realities of building massive data centers required to power generative AI. Such facilities consume enormous amounts of electricity, water, and other natural resources, and their construction has faced increasing backlash from local communities worried about rising utility costs and environmental impacts. Huang acknowledged these headwinds, noting that the changing resources could affect future investment strategies.
In the same conference, Huang commented on Nvidia’s recent $10 billion investment in Anthropic, another leading AI startup known for its Claude chatbot and focus on safety. He described that investment as “probably the last” opportunity to invest in Anthropic before it, too, goes public. Anthropic, founded by former OpenAI researchers, has been seen as a major competitor in the AI space, and its expected IPO would likely follow a similar trajectory to OpenAI’s.
The relationship between Nvidia and OpenAI is symbiotic but often tense. Nvidia supplies the vast majority of high-performance chips used to train and run OpenAI’s models, while OpenAI’s massive demand for those chips has helped drive Nvidia’s revenue to unprecedented levels. However, OpenAI has also been exploring the development of its own custom AI accelerators, a move that could reduce its dependence on Nvidia. Industry analysts suggest that OpenAI’s efforts to build proprietary hardware, combined with Nvidia’s desire to maintain leverage over its most important customer, may explain why the $100 billion investment was never finalized.
Historically, Nvidia’s business model has revolved around selling hardware components rather than making large equity investments in end-users. The company’s foray into significant startup stakes—such as the $30 billion round in OpenAI and the $10 billion in Anthropic—represents a strategic shift. Huang has positioned these investments as aligning Nvidia’s financial interests with the success of the AI ecosystem, but the scale of the proposed $100 billion commitment was unusual even for a highly profitable company like Nvidia, whose annual revenue in fiscal 2025 surpassed $90 billion.
The AI industry has been undergoing a rapid transformation since the launch of ChatGPT in late 2022. OpenAI has since released GPT-4, GPT-4o, and various other models, cementing its leadership. But the race has attracted fierce competition from Anthropic, Google’s DeepMind, Meta, and a host of startups. The high capital expenditure required for training state-of-the-art models has led to consolidation and strategic alliances. Nvidia’s decision to invest heavily in AI startups signals its bet on the long-term dominance of these players, but Huang’s latest comments reveal caution about over-commitment.
Another key fact from the conference is Huang’s emphasis on the “last chance” nature of investing in pre-IPO companies. He described both OpenAI and Anthropic as “consequential” companies that would shape the future of computing. The implied timeline suggests that Nvidia may use its balance sheet to acquire equity stakes before these companies go public, profiting from post-IPO valuations while also ensuring continued chip deals.
Moreover, the economics of AI have changed. In 2023 and 2024, many investors poured capital into AI startups without clear paths to profitability. Today, the market is demanding evidence of revenue growth and sustainable business models. OpenAI, for instance, reportedly spends billions of dollars annually to run inference and training workloads on Nvidia chips. The company’s revenue has grown rapidly, but its costs have also soared. Huang’s tempered outlook reflects a broader industry recalibration.
Environmental concerns add another layer. Data centers for AI require huge amounts of electricity. A single training run for a large model can consume as much energy as hundreds of homes over a year. Water usage for cooling has drawn criticism in drought-prone regions. Nvidia itself has been working on more energy-efficient chip designs, but the scale of demand continues to rise. Huang acknowledged these challenges, noting that future investments would need to account for sustainability.
The $100 billion investment that never happened may have been an attempt to secure a long-term exclusive supply arrangement. By tying investment to chip deployment, Nvidia could have locked in OpenAI as a guaranteed customer for years. However, OpenAI’s exploration of alternative chip suppliers—such as AMD or custom ASICs—may have made such an agreement less attractive to Nvidia. Huang’s statement that the deal is “probably” not happening leaves the door slightly open, but the tone suggests that the opportunity has passed.
In summary, Jensen Huang’s comments at the Morgan Stanley conference provide a clear snapshot of the current state of the Nvidia-OpenAI relationship. The key facts are as follows: Huang stated the $100 billion investment is unlikely; the reason given is OpenAI’s expected IPO; Nvidia only participated in a $30 billion round; similar reasoning was given for Anthropic’s $10 billion investment; the earlier $100 billion plan was never finalized and stalled in January; changing economics and resource constraints are reshaping AI investments; and Nvidia’s strategy now favors pre-IPO equity stakes over long-term commitments. Huang closed by emphasizing that these two investments might be the last chances to back such consequential companies before they go public.
Source: Silicon UK News