Nvidia’s $40 Billion Bet: How Dual-Track Investing Is Rewiring AI Infrastructure—and Triggering Global Antitrust Probes
The chip giant is deploying equity capital and commercial contracts simultaneously, locking customers into GPU ecosystems while regulators in five jurisdictions question whether demand is organic or balance-sheet-manufactured.
Nvidia committed over $40 billion in equity investments across AI infrastructure companies in 2026 alone—a dual-track strategy that pairs capital deployment with commercial GPU contracts, creating structural dependencies that now face antitrust scrutiny across five jurisdictions.
The scale is unprecedented in semiconductor history. During fiscal 2025, Nvidia invested CNBC reports, the company deployed $17.5 billion into private companies. This year’s $40 billion-plus pace represents a tripling of that velocity, executed through a model where equity options and commercial agreements lock portfolio companies into multi-gigawatt GPU deployments. The OpenAI partnership alone—announced in May 2026 and structured as up to $100 billion across 10 separate tranches—ties each $10 billion investment to the deployment of one gigawatt of Nvidia systems, with the first gigawatt targeted for the second half of 2026, according to Nvidia’s newsroom.
Nvidia’s non-marketable equity securities on its balance sheet surged to $22.25 billion as of January 31, 2026, up from $3.39 billion one year prior. Realized gains on these holdings hit $8.92 billion versus $1.03 billion the prior year, driven largely by a stake in Intel now valued above $25 billion. The company participated in 67 venture deals in 2025 compared to 54 in 2024.
The Circular Financing Model
The investment-plus-contract structure repeats across Nvidia’s recent deals. On May 7, the company signed a $3.4 billion five-year AI cloud contract with IREN alongside a right to invest up to $2.1 billion in equity at a $70-per-share exercise price, per Nvidia’s announcement. The deal commits IREN to deploying up to 5 gigawatts of AI Infrastructure. Days earlier, Nvidia announced a long-term optical networking partnership with Corning valued at up to $3.2 billion, coupled with an equity investment option and agreements to build three new fiber facilities in North Carolina and Texas creating over 3,000 jobs, according to CNBC.
Earlier this year, Nvidia deployed $2 billion into CoreWeave, the neocloud provider that has since appreciated 65% year-to-date. It invested another $2 billion each in photonics suppliers Marvell Technology, Lumentum, and Coherent in March, and $2 billion in Nebius Group—a cloud infrastructure spinout now up 80% year-to-date as of early May, per CNBC.
“It smells like you are pre-funding the purchase of your own GPUs and products.”
— Jordan Klein, Chip Analyst, Mizuho
The concern among analysts and short sellers is whether this capital cycle manufactures demand. Jim Chanos, founder of Chanos & Co., framed it bluntly in recent commentary: “They’re putting money into money-losing companies in order for those companies to order their chips.” Ben Bajarin of Creative Strategies warned that “if the cycle turns, the market starts questioning how much of the demand was organic versus supported by Nvidia’s own balance sheet,” according to CNBC.
Regulatory Crosshairs
The dual-track model has triggered formal investigations across five jurisdictions. The U.S. Department of Justice issued civil investigative demands in late 2024 as part of a broader Antitrust probe into Nvidia’s market dominance. France’s Autorité de la Concurrence concluded in June 2024 that Nvidia likely abused its dominant position through discriminatory pricing, production restrictions, and investments in AI-focused cloud providers like CoreWeave that could foreclose rivals, per TechPolicy.Press. The UK’s Competition and Markets Authority, the European Commission, and China’s State Administration for Market Regulation have all opened parallel probes.
China’s SAMR issued a preliminary finding of anti-monopoly law violations related to Nvidia’s acquisition of Mellanox, with potential fines of up to 10% of China revenue under consideration as of March 2026, according to AI CERTs News. Nvidia controls approximately 90% of the AI data center GPU market, making any preferential supply arrangements particularly sensitive under competition law.
Rebecca Haw Allensworth, an antitrust professor at Vanderbilt Law School, noted the incentive distortion inherent in the OpenAI partnership: “They’re financially interested in each other’s success. That creates an incentive for Nvidia to not sell chips to, or not sell chips on the same terms to, other competitors of OpenAI,” according to analysis by Mogin Law LLP.
Strategic Positioning Ahead of the Infrastructure Cycle
Nvidia’s investment blitz precedes what Goldman Sachs estimates as a $7.6 trillion cumulative AI capital expenditure cycle from 2026 through 2031, covering compute, data centers, and power infrastructure. AI-optimized facilities cost $15 million to $20 million per megawatt—50% to 100% more than traditional cloud infrastructure—creating enormous capital requirements that few players can self-finance, per Goldman Sachs analysis from March 2026.
CEO Jensen Huang acknowledged on the Dwarkesh Podcast in April 2026 that Nvidia missed early investment opportunities in OpenAI and Anthropic. The company now holds stakes valued at approximately $30 billion and $10 billion in those firms respectively, according to CNBC. Huang has publicly stated the company’s philosophy as agnostic: “There are so many great, amazing foundation model companies, and we try to invest in all of them. We don’t pick winners. We need to support everyone.”
Yet the portfolio approach creates systemic interdependencies. Bloomberg mapped the circular investment ecosystem connecting Nvidia, OpenAI, Microsoft, CoreWeave, and Anthropic—a web where capital flows in multiple directions while all parties remain dependent on Nvidia’s GPU supply. The architecture raises questions about concentration risk: if one node fails or regulatory action fragments the network, the knock-on effects could cascade.
Geopolitical Implications
Beyond antitrust concerns, Nvidia’s infrastructure consolidation carries national security dimensions. The company is effectively architecting the U.S. AI compute stack at a time when U.S.-China AI competition has elevated semiconductor access to a strategic priority. Export controls restrict Nvidia’s most advanced chips from Chinese markets, but the company’s equity stakes in Western infrastructure providers lock in long-term alignment with U.S. and allied deployment.
The Corning partnership illustrates this dynamic. Nvidia cannot acquire Corning outright without triggering severe antitrust review, but financing three domestic factories, taking an equity option, and securing preferred access to output achieves vertical integration by proxy while supporting domestic manufacturing—a structure that aligns with industrial policy objectives, per analysis by CNBC.
- Nvidia has committed over $40 billion in equity investments in 2026, triple its fiscal 2025 pace, using a dual-track model that pairs capital with commercial GPU contracts.
- The OpenAI partnership alone represents up to $100 billion tied to 10 gigawatts of system deployment, with the first gigawatt targeted for late 2026.
- Regulators in the U.S., France, UK, EU, and China have opened antitrust probes focused on whether Nvidia’s investments create preferential supply access or foreclose rivals.
- Analysts warn that if the AI infrastructure cycle slows, the market may question how much demand was organic versus balance-sheet-supported.
- The strategy positions Nvidia as the de facto architect of U.S. AI infrastructure, raising both competitive and geopolitical stakes.
What to Watch
The first gigawatt deployment milestone under the OpenAI partnership, expected in the second half of 2026, will test whether the 10-tranche structure proceeds as planned or faces delays that could reset market expectations for infrastructure build-out velocity. Regulatory timelines matter equally: the DOJ’s civil investigative demands could escalate to formal enforcement action if investigators find evidence of exclusionary conduct, while China’s preliminary Mellanox finding may result in fines that set a precedent for other jurisdictions.
Portfolio company performance will signal whether Nvidia’s conviction translates to sustainable business models. CoreWeave and Nebius have posted strong year-to-date gains, but both remain pre-profitability and dependent on continued access to capital and GPUs. IREN’s stock jumped 7% on the Nvidia deal announcement, yet the company trades at 63x forward earnings with execution risk around deploying 150,000 GPUs, according to 24/7 Wall St. Analyst consensus remains split, with 10 buy ratings against 2 strong sells.
Finally, any signs that Nvidia withholds chips from competitors of portfolio companies—or offers preferential pricing—will accelerate antitrust enforcement. The French regulator’s June 2024 finding already identified “likely abuse” tied to investments in firms like CoreWeave. If similar patterns emerge elsewhere, the dual-track strategy that has turbocharged Nvidia’s ecosystem reach may face structural limits imposed by competition law.