$30 billion went to OpenAI; the rest is spread across CoreWeave, IREN, Corning, Nebius and about two dozen private rounds. The pattern is closer to vertical integration than to venture capital, and raises the inevitable circular deal questions.
NVIDIA has invested more than $40 billion in artificial intelligence capital in the first four months of 2026, CNBC reports, citing public filings and corporate disclosures.
That total is $30 billion that the largest line chip maker put into OpenAI at the end of February. The remaining $10 billion is spread across seven multibillion-dollar deals in publicly traded companies, plus nearly two dozen private startup rounds.
On the public side, disclosed checks include up to $3.2 billion in optical fiber and ceramics maker Corning, which provides AI-data center fabrics, and Up to $2.1 billion in IRENA data center operator transitioning from Bitcoin mining to GPU computing.
Both took the form of guarantees or structured obligations rather than direct equity, with cash outflows determined at Nvidia’s discretion. The chip maker also added CoreWeave and Nebius positions during this period.
The CoreWeave stake, which was valued at $2 billion last January, is now valued at about $4.4 billion and represents about 28% of Nvidia’s listed equity portfolio.
2 dollarsIn March, bn Nebius invested is smaller in dollar terms but clearly has a five-gigawatt deployment commitment; A new $2.1 billion warrant to IREN it rests on a similar logic.
Among these, the pattern is consistent: capital flows to companies that buy Nvidia GPUs at scale and lease them back to hyperscalers and frontier model builders, what the industry now calls neocloud.
NVIDIA’s own strategic framework is simple. CFO Colette Kress said on the most recent earnings call that the company is investing where it sees a need to ensure computing capacity is built around its hardware.
Last fiscal year, the company invested $17.5 billion in private companies and infrastructure funds, primarily early-stage startups. The pace of 2026 already exceeds the previous year.
The investments themselves are largely small relative to Nvidia’s roughly $200 billion in cash and equivalents, meaning they don’t strain the balance sheet; what matters is what they signal about how the chipmaker sees its place in the AI value chain.
That location is increasingly upstream and downstream of the chip itself. The OpenAI investment is not a standalone bet; it is coupled with multi-year computing commitments and silicon roadmap alignment.
The CoreWeave and Nebius positions come with capacity reserves and joint architecture agreements. Corning’s investment supports the optical interconnect supply chain on which Nvidia depends for next-generation data center materials.
From start to finish, Nvidia influences how its silicon is paid for, placed and bonded. Some analysts call this vertical integration; others call it crowdfunding.
Criticism of circular bargaining has gained attention over the past two quarters. NVIDIA holds a position in the company; that company then signs a long-term GPU purchase commitment with NVIDIA; Part of the GPU revenue flowing back to NVIDIA can be characterized as a return on the same capital it just invested.
Related to Oracle’s $300 billion OpenAI deal and the concentration it creates is the most frequently cited example of a broader problem; The concentration of revenue counterparties was one of the reasons analysts were more cautious about Oracle, even as headline numbers rose.
With Nvidia’s smaller portfolio companies, the model is the same, just with more upside.
There are reasons why the comparison is partly unfair. Nvidia’s investments are typically minority positions in companies with many other customers; Meta’s $21 billion addition to CoreWeave It demonstrates that CoreWeave’s customer base is wider than Nvidia’s.
Mistral AI, Wayve, Lambda Labs, Genesis Therapeutics, Recraft and JetBrains are all clients or investments with independent commercial logic.
The criticism is more acute in deals where Nvidia is both a meaningful equity investor and a contractually committed customer of the same company; CoreWeave’s $6.3 billion capacity purchase agreement with Nvidia is the most cited example of this.
The bigger question is what happens to the portfolio when AI computing demand normalizes. Most of Nvidia’s bets are financially small relative to the parent’s revenue and cash position, so a write-off event would not hurt the core business.
TA more significant risk is reputation. Each new deal, which is structurally similar to the previous one, adds to the perception that Nvidia is financing its own demand curve.
Both Wall Street and the SEC are beginning to question whether the disclosure regime around these regulations is up to their scale.
So far, the strategy is getting the results Nvidia wants. Artificial intelligence infrastructure is built where Nvidia’s silicon works, model providers provide compute they might not otherwise be able to build independently, and the chipmaker’s data center revenue grows accordingly.
The pace of 2026 capital commitments suggests that Nvidia intends to continue writing the same type of check as long as the supply-demand mismatch persists.
CNBC’s tally of multibillion-dollar deals in seven public markets and 24 or more private rounds is, by its own terms, a record. It also positions Nvidia as the largest single source of AI infrastructure funding in the market, alongside major hyperscalers.
The role fits Jensen Huang’s story of being the platform of the AI era. Whether this is acceptable to auditors and regulators is the next question.
