Nvidia’s equity investments stood at $99 billion as of July 26, up sharply from about $7 billion a year earlier and $2.2 billion two years ago, according to the company’s latest earnings report. The chip giant has become one of the world’s largest corporate backers of technology companies, with nearly $50 billion of the portfolio concentrated in AI laboratories and infrastructure providers.
The growth trajectory tells the story. Nvidia’s equity holdings were worth $2.2 billion in mid-2024. By mid-2025, that figure had risen to approximately $7 billion. In the twelve months since, the portfolio exploded to $99 billion, driven by a combination of market appreciation and aggressive new commitments across the AI sector. The company has assumed investment obligations of more than $40 billion since the beginning of 2026 alone.
CNBC first reported the figures on September 4, drawing on Nvidia’s latest quarterly filing with the Securities and Exchange Commission. The numbers place Nvidia behind only Alphabet and Amazon among corporate equity investors, both of which hold more than $100 billion accumulated over more than a decade of corporate venturing. Nvidia reached near parity in roughly two years, a pace that has drawn comparisons to the early days of Alphabet’s venture arm.
The AI lab concentration
Nearly half of the portfolio sits in frontier AI companies. The largest single commitment is a $30 billion stake in OpenAI, part of the company’s $110 billion funding round announced in February. Nvidia has also invested $2 billion each in CoreWeave and Nebius, two cloud providers that rent computing power built on Nvidia’s own accelerators.
These investments serve a dual purpose. They help fund the companies that buy Nvidia’s most expensive products, and they give Nvidia equity exposure to the businesses that depend on its chips. It is a feedback loop that critics call a circular economy: Nvidia sells chips to companies, then invests in those same companies to ensure they keep buying more chips.
The company has also committed approximately $6.5 billion since March to photonics and optical networking firms Lumentum, Coherent, and Marvell, embedding itself throughout the AI supply chain beyond just the processor itself. A $12.9 billion acquisition of Hugging Face, the open-source AI model repository, expanded Nvidia’s reach into the software layer of the AI ecosystem. Hugging Face hosts more than one million machine learning models and is used by researchers and developers worldwide to distribute and fine-tune AI systems.
SpaceX, Intel, and the non-AI bets
Not all of the portfolio is AI-focused. Nvidia’s stake in SpaceX was valued at $21 billion in June, making it one of the largest positions in the book. A $5 billion investment in Intel last year has turned into a stake worth approximately $30 billion, though that value fluctuates with Intel’s volatile stock price. A $3.5 billion commitment to MediaTek, the Taiwanese chip designer, targets the mobile and edge computing market where AI models are increasingly running on devices rather than in the cloud.
Nokia received $1 billion from Nvidia as part of a push into telecommunications infrastructure, where AI-powered network optimization is becoming a selling point for carrier upgrades. The breadth of these investments shows Nvidia positioning itself as something closer to a holding company than a traditional semiconductor firm, one that profits whether its own chips or someone else’s infrastructure wins.
The risk side
The portfolio carries real risks that Nvidia has begun to disclose in more detail. The company has prepared to take up to $125 billion in risk related to potential declines in the residual value of accelerators provided to customers under financing arrangements. These arrangements, where Nvidia essentially helps customers finance their chip purchases, create exposure to counterparty risk if those customers face financial difficulty.
Nvidia also committed up to $105 billion in credit guarantees for a data center project in Ohio built to meet OpenAI’s computing needs. The guarantee means Nvidia would be on the hook if the project’s financing collapses, a significant exposure for a company whose core business is selling chips, not underwriting real estate and infrastructure projects.
Nvidia’s last quarter revenue increased 106 percent on an annual basis to $96.2 billion. Of that total, $48.7 billion came from the hyperscale segment, which includes the largest cloud providers like Microsoft, Amazon, and Google. The concentration of revenue in a small number of very large customers creates dependency, which the investment portfolio is partly designed to address by building relationships across a broader base of AI companies and startups.
A new kind of tech giant
The investment strategy reflects a broader shift in how Nvidia sees itself. CEO Jensen Huang has spoken about building an AI ecosystem, not just selling chips. The investments in cloud providers, AI labs, networking companies, and software platforms all serve that vision, creating a web of dependencies that benefits Nvidia regardless of which specific AI company succeeds or fails.
The portfolio also gives Nvidia influence over the direction of AI development. By investing in OpenAI, CoreWeave, and Hugging Face, the company gains insight into how its chips are being used and what capabilities researchers want next. That information feeds back into Nvidia’s product development cycle, potentially giving it an edge over competitors like AMD and Intel.
By comparison, Alphabet and Amazon each hold more than $100 billion in equity investments accumulated over more than a decade of corporate venturing. Nvidia has nearly caught up in just two years, a pace of portfolio construction that has few precedents in corporate finance. The question is whether the portfolio’s value will hold if AI spending growth slows, or if the circular nature of the investments creates hidden fragility that only becomes apparent in a downturn.
discussion