Nvidia's AI Equity Bets Cross $40 Billion in 2026, Anchored by a $30B OpenAI Stake

Nvidia has crossed $40 billion in announced AI equity commitments for 2026, according to a CNBC tally published May 9, cementing the chipmaker's transformation from arms dealer to financier of the entire AI build-out. The single biggest line item is its previously disclosed $30 billion stake in OpenAI, which alone accounts for three-quarters of the year's investment activity. The rest is spread across roughly two dozen public-company stakes and private startup rounds, layered on top of 67 venture deals Nvidia did in 2025.

Two of the freshest deals — both unveiled in the past week — extend the pattern into infrastructure and components. Nvidia is committing up to $2.1 billion to Australian-listed data-center operator IREN, in exchange for a five-year, $3.4 billion managed-GPU-cloud arrangement that gives Nvidia access to capacity for its own internal research workloads. As part of the same partnership, IREN will deploy up to 5 gigawatts of Nvidia's DSX-branded reference designs across global facilities, effectively turning the operator into a preferred Nvidia-spec landlord for AI compute.

The Corning investment, sized at up to $3.2 billion, is a bet on the physical layer of next-generation racks. Corning will build three new U.S. plants dedicated to optical components for Nvidia, signaling that the company plans to migrate its rack-scale systems from copper to fiber as power and bandwidth requirements outrun what copper can handle. Combined, the two deals position Nvidia inside both the megawatts and the millimeter-scale interconnect that future deployments will depend on.

CEO Jensen Huang framed the strategy on Nvidia's February earnings call as "focused very squarely, strategically on expanding and deepening our ecosystem reach." Wedbush analyst Matthew Bryson told CNBC the deals fit "squarely into the circular investment theme" — Nvidia is, after all, taking equity in many of its own customers — but argued that, if the bets pay off, the chipmaker will have built "a competitive moat" that's hard for AMD or hyperscaler in-house silicon teams to dislodge.

The circular-deal critique is gaining traction with regulators and short-sellers who note that capital is increasingly cycling between the same handful of names — OpenAI, CoreWeave, IREN, Lambda, and now Corning — with Nvidia at the center of every flow. Even so, the strategy is reshaping how AI infrastructure gets financed: rather than waiting for capital markets to fund customers' GPU buildouts, Nvidia is underwriting the demand directly and getting equity upside in return.