The common misconception that all of Nvidia's external investments flow through its venture arm NVentures obscures a far more complex reality. In truth, the two-person NVentures team manages about 30 deals per year, yet its total spending is dwarfed by a single transaction from the Corporate Development team. At the end of 2025, that team poured $2 billion of equity into Synopsys alone—already several times NVentures' cumulative investment over the past three years. Nvidia's investment apparatus operates on three distinct tracks: Corporate Development handles strategic, multi-billion to hundred-billion-dollar investments and M&A; NVentures focuses on early-stage, broad-based financial investments; and NVIDIA Inception functions as an accelerator providing resources without direct capital. Together, they form the largest and fastest capital deployment machine in Silicon Valley history, and the very core target of short-sellers' "circular financing" concerns.
NVentures: A Two-Person Team and 20 Unicorns
Despite the Nvidia brand, NVentures operates with an unexpectedly lean structure. According to private-market data provider Tracxn, as of May 2026, the team had just two members but had already invested in 79 companies, breeding 20 unicorns including AI video platform Synthesia, clinical AI firm Abridge, and quantum computing company PsiQuantum. Over the past 12 months, the team closed 43 new deals, with 20 executed in the first five months of 2026 alone—a clear acceleration. The team is led by Vice President Mohamed “Sid” Siddeek, whose career trajectory mirrors Nvidia's positioning for this unit. He worked at Morgan Stanley in the late 1990s, accompanying Jensen Huang on the Nvidia IPO roadshow, then spent nearly a decade leading TMT and telecom investments at Abu Dhabi's sovereign fund Mubadala, and later headed enterprise software and healthcare investments at SoftBank Vision Fund before returning to Nvidia in 2021 to build NVentures. Siddeek describes the investment criteria simply: “The true filtering is two layers—first, any domain Nvidia can touch; second, which of those domains are investable.” This spans horizontally across almost every industry AI can transform—healthcare, manufacturing, robotics, autonomous driving, quantum computing—and vertically from foundational tools to application layers.
Corporate Development: Strategic Deals at Hundred-Billion Scale
In stark contrast stands the Corporate Development team led by Vishal Bhagwati, responsible for all strategic, large-scale investments, joint ventures, and M&A. The pace and magnitude of its recent moves have been breathtaking: in February 2026, it spearheaded a $30 billion investment in OpenAI (part of an approximately $110 billion funding round) with a commitment to potentially add up to $100 billion more; in November 2025, a $10 billion commitment to Anthropic; at the end of 2025, $2 billion into Synopsys; in early 2026, an additional $2 billion into CoreWeave alongside a $6.3 billion cloud capacity purchase agreement; in March 2026, another $2 billion into Nebius; and a commitment of up to $2 billion in equity to xAI. According to CNBC, in just the first four months of 2026, the Corporate Development team directed over $40 billion in AI equity investments, far exceeding Nvidia's total of $17.5 billion deployed into private companies and infrastructure funds in fiscal 2025.
NVentures: Follower Style and Recent Focus
NVentures, by contrast, operates as a traditional VC seeking financial returns, with check sizes ranging from a few million to tens of millions of dollars, primarily in Seed to Series B stages, and predominantly as a co-investor—leading only about one-eighth of its deals—joining rounds alongside top VCs like Accel, a16z, and Sequoia to provide Nvidia's endorsement. May 2026 alone saw four disclosed deals: on May 22, participation in the €100 million Series B extension of French quantum computing firm Alice & Bob; on May 26, joining the $113 million Series B of AI model routing platform OpenRouter; on May 28, the $20 million seed extension of AI inference infrastructure startup Tensormesh; and on May 6, the $35 million Series C extension of AI cybersecurity company Xbow. These bets converge on three areas: quantum computing (Alice & Bob, Quantinuum, PsiQuantum), AI biomedicine (Relation Therapeutics, Genesis Therapeutics), and AI agents/inference layers (OpenRouter, Tensormesh), aligning precisely with Nvidia's next-generation software stack development on CUDA-Q, CUDA-X, and Triton. Geographically, Europe has become an expanding focus, with 14 deals in 2025 doubling the 7 made in 2024.
Five Quadrants of Capital Influence
Mapping out all three tracks reveals five principal quadrants of Nvidia's capital influence across the AI ecosystem. The foundational model layer—OpenAI, Anthropic, xAI, Mistral, among others—is predominantly funded by Corporate Development. The cloud and infrastructure layer—CoreWeave, Nebius, Lambda, Crusoe, Nscale—is likewise led by Corporate Development, with each deal often worth billions and tied to long-term computing capacity purchase agreements. The application and developer tools layer—Cursor, Perplexity, Synthesia, Runway, Lovable, Together AI, Weka—sees higher NVentures participation but with smaller checks. The robotics and autonomous driving layer—Figure AI (latest valuation $39 billion) and Wayve ($8.6 billion)—features joint participation from both Corporate Development and NVentures. The quantum computing and biomedicine layer—PsiQuantum, Quantinuum, Alice & Bob, Relation Therapeutics—is primarily funded by NVentures' early-stage investments, serving as Nvidia's hedge against a “post-GPU” computing paradigm. According to VC research firm F4 Fund, between 2025 and early 2026, at least 10 companies in Nvidia-backed rounds crossed the $1 billion valuation threshold, including OpenAI, Anthropic, xAI, Mistral, Figure AI, Cursor, Perplexity, Scale AI, and Wayve.
Short-Seller's Spotlight: Burry and the Circular Financing Debate
The sprawling investment footprint is drawing increasing scrutiny. The sharpest critique comes from Michael Burry, the hedge fund manager immortalized in The Big Short. In Q3 2025, his Scion Asset Management disclosed new short positions on Nvidia and Palantir, holding put options on approximately 1 million Nvidia shares with a notional exposure of around $187 million, and 50,000 put contracts on Palantir costing about $9.2 million in premium. Burry posted a Big Short movie still on X's Cassandra Unchained account with the caption “Sometimes, we can see the bubble,” then reposted a Bloomberg chart on Nvidia's circular financing, directly targeting the capital deployment model. In his Substack, he estimated that between 2026 and 2028, major cloud providers including Microsoft, Google, Oracle, and Meta would collectively understate depreciation by about $176 billion by extending the accounting depreciation lives of Nvidia GPUs, thereby inflating their profits. This accounting maneuver resonates with Nvidia's equity investments in its customers: the former gives buyers higher “paper profits” to absorb larger capex, while the latter directly supplies funds for hardware purchases. At the institutional level, the EU competition regulator explicitly included Nvidia's “circular spending risks” in its review in March 2026. Seaport Research estimated that for every $1 of equity Nvidia invests, it yields approximately $3.50 in downstream chip revenue. Bloomberg's “AI Circular Transactions” feature mapped a dense web of money flows among Nvidia, CoreWeave, OpenAI, Oracle, and Anthropic: Nvidia holds about a 7% stake in CoreWeave, which uses GPUs as collateral to finance more GPU purchases from Nvidia, and Nvidia signed a $6.3 billion cloud capacity purchase agreement to absorb CoreWeave's excess capacity through 2032. Similar chains exist with OpenAI, Anthropic, and xAI.
Virtuous Cycle or Vendor Financing?
Not everyone views the model critically. Asset manager Janus Henderson characterizes it as a “virtuous cycle,” arguing that binding supply and demand through “equity plus long-term procurement contracts” is a sensible business arrangement in an era of extreme compute scarcity. Morningstar's analysis points out that Nvidia's commitment to buying CoreWeave's excess capacity actually puts Nvidia on the hook for inventory risk, creating a counterbalance to short-term incentives to sell hardware. Yet a growing refrain among market observers is that in a time of compute shortage, whether one believes “the entanglement of equity and procurement contracts is coincidence” is itself a question of trust.
The Ambiguous Position of NVentures
Within this controversy, NVentures occupies a delicate position. Its early-stage, small-check, follower-oriented, sector-diversified style stands in sharp contrast to the Corporate Development team's “circular trading” patterns. Companies like Alice & Bob, Tensormesh, and OpenRouter are far too small to form the “both customer and investment target” loop; their investments more closely resemble traditional corporate VC financial logic. But from the perspective of Nvidia's overall investment system, the question—unspoken but implicit in Burry's and EU regulators' scrutiny—is whether NVentures unwittingly serves as a “VC compliance cloak” in external disclosures, making it easier for outsiders to perceive Nvidia's investment activities as normal venture capital rather than systematic vendor financing. Nvidia's official stance remains that all investments are based on independent business judgment and are not tied to hardware sales.