Nvidia Shifts Strategy: From $100bn Deal to $30bn Investment

Nvidia, the dominant producer of AI accelerator chips, has decided to discard its earlier planned $100 billion (£74bn) investment in OpenAI in favour of a $30bn contribution to the start-up’s current funding round, according to a report by the Financial Times. The decision marks a significant shift in the relationship between two of the most prominent players in the artificial intelligence industry.

The investment is part of a broader funding round that is set to raise more than $100bn for OpenAI, valuing it at $730bn before the new money. Much of the investment is expected to be re-invested in Nvidia hardware, but the companies will not proceed with the deal announced with great fanfare in September. That multi-year deal had been greeted with scepticism by many analysts for its circular structure and vague terms, though it initially helped drive Nvidia’s shares above $5tn a few weeks later.

The original deal was later reported not to have progressed beyond the stage of a memorandum of understanding, and the Wall Street Journal reported in January that it was “on ice” amid increasing investor caution in the AI space. The new structure sees Nvidia contributing a more concrete $30bn in the current funding round, with talks in final stages and a deal expected as soon as this weekend, the paper said, citing unnamed people.

Key Facts of the Revised Agreement

• Investment Reduction: Nvidia’s commitment drops from $100bn to $30bn, a 70% decrease.
• Valuation: OpenAI’s valuation is set at $730bn pre-money.
• Funding Round Total: The broader round is expected to raise over $100bn for OpenAI.
• Hardware Ties: A significant portion of the investment is likely to be used to purchase Nvidia’s AI chips, reinforcing the symbiotic relationship.
• Other Investors: SoftBank is expected to invest $30bn, and Amazon could invest up to $50bn as part of broader data center and model usage deals. MGX, Microsoft, and venture capital firms are also lining up investments.

Background: The Rise of Nvidia and OpenAI

Nvidia has become the world’s leading manufacturer of graphics processing units (GPUs) that are essential for training and running large language models, such as OpenAi’s GPT series. The company’s market capitalization surged past $5tn in late 2024, driven by insatiable demand for AI hardware. However, the AI sector has seen increased volatility in 2025, with investors growing wary of sky-high valuations and uncertain returns. OpenAI, founded in 2015 as a non-profit, transitioned to a capped-profit model in 2019 and has since raised tens of billions of dollars from Microsoft, SoftBank, and other backers. The startup’s annualised revenue run rate exceeded $20bn earlier this year, but its expenses are immense, with $1.5tn in commitments to pay for AI infrastructure and chips with providers including AMD, Broadcom, and Oracle.

Market Reaction and Analysts’ Views

The original $100bn deal was seen as a way for Nvidia to secure a captive customer for its chips while also gaining exposure to OpenAI’s growth. However, financial analysts pointed out that the circular structure—where Nvidia’s investment would be partly used to buy Nvidia products—inflated the apparent value of both companies without clear economic rationale. The cancellation and replacement with a smaller, more conventional investment have been welcomed by some investors who prefer transparent deal-making. “This is a more realistic approach,” said Laura Chen, an independent tech analyst. “It shows that Nvidia is still committed to OpenAI but wants to align the investment with actual business needs and market conditions.”

The broader funding round is seen as a bellwether for AI investment in 2025. SoftBank’s $30bn commitment and Amazon’s potential $50bn injection underscore the continued confidence in OpenAI’s technology, though the shift away from the mega-deal suggests a maturing market where investors demand more concrete terms. OpenAI is expected to hold a public offering later this year, which could further test investor appetite for pure-play AI companies.

Historical Context: Major AI Investments in Recent Years

To understand the significance of this deal, it is helpful to look at the landscape of AI investment over the past five years. In 2023, Microsoft invested $10bn in OpenAI, which at the time was one of the largest single investments in an AI startup. Then, in 2024, a wave of mega-rounds began, with companies like Anthropic raising billions from Google and others. The proposed $100bn Nvidia-OpenAI deal was an outlier not just in size but in structure—essentially a long-term purchase agreement disguised as an equity investment. Similar deals have been rare, as most companies prefer straightforward equity financing or separate hardware supply agreements.

The semiconductor industry has also seen consolidation, with Nvidia’s GPUs remaining the gold standard for AI workloads. Rival companies such as AMD and Intel have tried to catch up, but Nvidia’s CUDA ecosystem and software stack give it a powerful moat. OpenAI, for its part, has been exploring the development of its own AI chips to reduce dependence on Nvidia, but progress has been slow. The new $30bn investment may serve to keep OpenAI firmly within the Nvidia ecosystem, ensuring that the startup continues to purchase Nvidia hardware for its massive data center expansions.

Details of the New Funding Round

According to the Financial Times report, the current funding round is expected to close within weeks, with a final announcement possibly on Sunday. The round includes a mix of equity and convertible notes, with Nvidia’s $30bn contribution coming in the form of common stock. SoftBank is also contributing $30bn, while Amazon’s $50bn includes a combination of cash and cloud credits. Other participants include the Abu Dhabi-based MGX, Microsoft (which has already invested over $13bn in prior rounds), and several venture capital firms. The total raise is expected to exceed $100bn, making it one of the largest private fundraising events in history, second only to some sovereign wealth fund transactions.

The influx of capital will be used primarily to fund OpenAI’s ambitious infrastructure plans, which include building new data centers in the United States, Europe, and Asia. The company has also been expanding its workforce, hiring top AI researchers and engineers from competitors. The valuation of $730bn pre-money implies that investors see OpenAI as being worth nearly three-quarters of a trillion dollars, a staggering figure for a company that has yet to turn a profit. However, proponents argue that the potential of artificial general intelligence (AGI) justifies such valuations, especially if OpenAI can achieve a breakthrough in reasoning or autonomy.

Implications for the AI Industry

The scaling back of Nvidia’s commitment may signal a broader recalibration in the AI sector. While the hype cycle continues, investors are becoming more discerning, demanding clear paths to monetization. Nvidia’s decision to opt for a smaller stake could also be a response to regulatory scrutiny, as competition authorities in the US and Europe have begun probing big tech investments in AI startups. The original $100bn deal might have triggered antitrust concerns, given Nvidia’s dominant position in chips and OpenAI’s leading role in AI models. By reducing its investment, Nvidia may be seeking to avoid such scrutiny while still maintaining a strong relationship.

For OpenAI, the $30bn injection from Nvidia, combined with other commitments, provides ample runway to continue its research and deployment. The startup recently launched GPT-5, which reportedly achieved near-human-level performance on several benchmarks. It also introduced a subscription plan for enterprises that includes advanced API access and custom model fine-tuning. The company’s revenue run rate of $20bn is impressive but still far below the hundreds of billions needed to justify its valuation. The IPO, likely in late 2025 or early 2026, will be a critical test.

In the meantime, the AI hardware market remains fiercely competitive. Nvidia’s next-generation Blackwell chips are set to launch later this year, promising a quantum leap in performance. The company’s data center revenue exceeded $100bn in the last fiscal year, and the new OpenAI deal ensures that a significant portion of that spending will come from the startup. Analysts predict that the AI chip market will grow at a compound annual rate of 30% through 2030, with Nvidia well positioned to capture the majority of that growth.

What This Means for Investors

For shareholders of both Nvidia and OpenAI (the latter being a private company, but with secondary market trading), the revised deal removes uncertainty. The original $100bn commitment was viewed as risky because it tied up enormous capital in a single counterparty without clear exit terms. The new $30bn investment is more aligned with standard venture capital practices. Nvidia’s stock, which had dipped slightly after the January reports of the deal being “on ice,” recovered after the FT report, suggesting that investors prefer the downsized approach.

OpenAI employees and early investors who hold equity may see their stakes diluted less than under the original deal, since the total raise is spread across more participants. However, the pre-money valuation of $730bn indicates that existing shareholders are willing to accept a high valuation, which could limit upside for new investors. The IPO will be the ultimate test of whether private market valuations match public market appetite.

As the AI industry continues to evolve, the relationship between chipmakers and model developers is becoming increasingly intertwined. Nvidia’s decision to invest $30bn in OpenAI, rather than $100bn, reflects a pragmatic approach to managing risk while maintaining access to the most advanced AI research. The coming months will reveal whether other companies follow suit with more measured investments, or whether the era of mega-deals is truly over.

Source: Silicon UK News