Nvidia has committed more than $40 billion to equity investments in artificial intelligence companies in the first five months of 2026, according to data compiled by CNBC and FactSet. The figure is anchored by a single $30 billion stake in OpenAI, but seven additional deals in publicly traded companies have followed in rapid succession. The most recent two arrived this week: an agreement giving Nvidia the right to invest up to $3.2 billion in glassmaker Corning, and a separate pact with data center operator IREN worth up to $2.1 billion. Both announcements sent the respective shares higher on the day of disclosure. "AI factories are becoming foundational infrastructure for the global economy," CEO Jensen Huang said in a statement accompanying the IREN deal.
What the Deals Actually Buy
Each investment is structured around a commercial relationship, not a passive financial bet. The Corning agreement comes with a commitment from the 175-year-old glass manufacturer to build three new U.S. facilities dedicated to optical networking technologies—Nvidia is transitioning its rack-scale systems from copper interconnects to fiber-optic cables, and Corning is now a key supplier in that shift. The IREN deal pairs the equity right with a five-year, $3.4 billion agreement for the data center operator to provide Nvidia with managed GPU cloud services at its Childress, Texas facilities, and a commitment to deploy up to five gigawatts of Nvidia's DSX infrastructure designs globally. IREN had previously signed a $9.7 billion deal with Microsoft in November 2025 for GB300 GPU infrastructure, making it one of the more actively courted operators in the AI infrastructure stack. Earlier in 2026, Nvidia placed $2 billion each in Marvell Technology, Lumentum, and Coherent—all companies working on silicon photonics—and $2 billion in CoreWeave and Nebius Group for data center and AI cloud deployment.
The OpenAI Revision and the Circular Investment Critique
The original Nvidia-OpenAI arrangement was conceived at a far larger scale. In September 2025, the companies announced Nvidia would invest up to $100 billion in OpenAI over time as the AI lab deployed ten gigawatts of Nvidia systems. That plan was scaled back after OpenAI pivoted toward relying on infrastructure partners rather than building its own data centers. Huang said in March that the $100 billion figure was probably "not in the cards," and that the $30 billion deal might be the last investment before an OpenAI IPO. Critics have noted the structural tension in the broader portfolio: Nvidia is investing in companies that are also customers, creating what Wedbush Securities analyst Matthew Bryson described in a note as "squarely into the circular investment theme"—capital that flows to firms which then turn around and buy Nvidia's hardware. Bryson acknowledged the strategy could build a competitive moat if successful, but the logic of a chip company funding its own demand is drawing scrutiny from institutional investors.
A $5.2 Trillion Company Doubling Down
Nvidia's stock has risen more than eleven-fold over four years, pushing its market capitalization to approximately $5.2 trillion as of May 8. Goldman Sachs, ahead of Nvidia's May 20 earnings, raised its revenue and earnings estimates by roughly 12 percent, with calendar-year 2027 forecasts running 34 percent above Street consensus. Altimeter Capital CEO Brad Gerstner has suggested the company could become the world's first $10 trillion business. An earlier bet that illustrates the return potential: Nvidia's $5 billion investment in Intel is now valued at more than $25 billion. The company participated in roughly 67 venture rounds in 2025 and has already joined approximately 24 private startup rounds in 2026 alone—a pace that is less characteristic of a chipmaker than of a late-stage venture fund that happens to manufacture the hardware its portfolio companies need.
By now, the pattern looks deliberate. Nvidia's investment strategy has moved from opportunistic participation to something closer to ecosystem management: back the data centers that need the GPUs, back the component suppliers that build the interconnects, back the cloud operators that run the inference workloads. Whether that constitutes a durable competitive moat or a house of cards built on circular capital will depend entirely on whether AI infrastructure spending continues to expand at its current rate—a question the May 20 earnings call may begin to answer.
