NVIDIA surpasses $40 billion in AI equity investments in 2026.

      OpenAI received $30 billion, while the remaining funds are allocated among CoreWeave, IREN, Corning, Nebius, and about two dozen private funding rounds. This trend leans more towards vertical integration than traditional venture capital investing, prompting questions about potential circular deals.

      According to CNBC, NVIDIA has invested over $40 billion in AI equity within the first four months of 2026, based on public records and corporate updates. The largest component of this total is the $30 billion invested in OpenAI in late February. The rest, exceeding $10 billion, is distributed across seven multi-billion-dollar arrangements with publicly traded companies and around two dozen private startup investments.

      Among the publicly disclosed investments is up to $3.2 billion in Corning, which produces optical fiber and ceramics for AI data centers, and as much as $2.1 billion in IREN, a data center operator shifting from Bitcoin mining to GPU computing. Both investments were made through warrants or structured commitments rather than straightforward equity, with cash outflows scheduled at Nvidia’s discretion. During this period, the chipmaker also increased its stakes in CoreWeave and Nebius.

      The $2 billion investment in CoreWeave made last January is now valued at approximately $4.4 billion, accounting for around 28% of Nvidia’s listed equity portfolio. The smaller $2 billion Nebius investment from March includes a specific commitment for five gigawatts of deployment, aligned similarly to the new $2.1 billion warrant for IREN.

      The pattern shows that capital is directed towards companies purchasing Nvidia GPUs in large quantities and renting them out to hyperscalers and frontier model developers, a framework now referred to as a neocloud. NVIDIA has articulated this strategy clearly; CFO Colette Kress stated in the most recent earnings call that the company invests where it deems necessary to ensure compute capacity is developed around its hardware.

      Last fiscal year, Nvidia invested $17.5 billion in private companies and infrastructure funds, mainly targeting early-stage startups, according to their 10-K report. The pace of investment in 2026 already surpasses that of the previous full year. These investments are relatively small compared to Nvidia’s approximately $200 billion in cash and equivalents, meaning they do not strain the balance sheet; what is significant is the signal they send about Nvidia's perspective in the AI value chain.

      Nvidia's role is becoming increasingly integrated both upstream and downstream of the chip itself. The investment in OpenAI is not merely an isolated bet; it is coupled with multi-year compute commitments and alignment with silicon roadmaps. The investments in CoreWeave and Nebius include capacity reservations and joint architecture agreements, while the investment in Corning bolsters the optical interconnect supply chain crucial for next-generation data center fabrics.

      Examining the entire situation, Nvidia is gaining influence over pricing, deployment, and connectivity regarding its silicon. Some analysts describe this as vertical integration, while others label it circular financing. The criticism regarding potential circular deals has intensified over the last two quarters. NVIDIA invests in a company, which subsequently enters into a long-term GPU purchase agreement with NVIDIA; some of the revenue generated from these GPU sales could be interpreted as a return on the equity NVIDIA just invested.

      Oracle’s $300 billion deal with OpenAI and the concentration concerns arising from it serve as the most frequently referenced example of this broader issue; the concentration of revenue counterparties has led analysts to adopt a more cautious stance on Oracle despite rising headline figures. The pattern observed with Nvidia’s smaller portfolio companies is similar, albeit with a greater number of counterparties.

      There are some reasons this comparison may be somewhat unfair. Nvidia's investments typically involve minority stakes in companies that have diverse customer bases; for instance, Meta’s $21 billion investment in CoreWeave shows that CoreWeave serves clients beyond Nvidia. Companies like Mistral AI, Wayve, Lambda Labs, Genesis Therapeutics, Recraft, and JetBrains are all either customers or investments with their own independent commercial rationale.

      The critique holds more weight in cases where Nvidia is both a significant equity investor and a contractually obligated customer of the same company, with CoreWeave’s $6.3 billion capacity purchase agreement being the most frequently cited example.

      The more pressing concern is what occurs with the portfolio when the demand for AI compute normalizes. Many of Nvidia’s investments are relatively minor in comparison to the parent company’s overall revenue and cash reserves, suggesting that a write-down would not adversely affect the core business.

      The greater risk lies in reputational damage. Each new deal resembling a prior one enhances the perception that Nvidia is essentially financing its own demand curve. Both Wall Street and the SEC are beginning to scrutinize whether the disclosure practices regarding these arrangements are keeping up with their scale.

      Currently, the strategy is achieving the desired results for Nvidia. AI infrastructure capacity is being established where Nvidia silicon operates, model providers are obtaining compute resources they could not create independently, and the chipmaker