What are the targets invested in by NVIDIA's "Three-Track Architecture" investment system?

Original Title: "What Targets Has Nvidia's 'Three-Track Architecture' Investment System Invested In?"
Original Author: Ada, Deep Tide TechFlow

Recently, Nvidia's venture capital arm NVentures made a new investment in France-based quantum computing company Alice & Bob, focusing on fault-tolerant quantum computing.

It is a common misconception to think that all of Nvidia's external investments are consolidated under NVentures. In fact, this venture arm, established in 2021, made only 30 investments throughout the year, with the total scale far below a single strategic investment by the corporate development team. The latter alone invested $2 billion in Synopsys at the end of 2025, which is several times the total of NVentures' nearly three-year cumulative investments.

To understand how Nvidia weaves an AI ecosystem with capital, one must start with its "Three-Track Architecture" of its investment system. The corporate development team handles strategic, large-scale investments and mergers worth billions to over a hundred billion dollars; NVentures is responsible for early-stage, industry-wide financial investments; NVIDIA Inception is an accelerator that does not invest but provides resources and connections.

Together, these three form the largest and fastest-paced capital deployment machine in Silicon Valley history, also becoming the core target of "cyclical financing" skepticism among short-sellers.

The True Face of NVentures: 2-Person Team, 79 Companies, 20 Unicorns

Although NVentures carries Nvidia's brand, its internal scale is surprisingly small.

According to private equity data firm Tracxn, as of May 2026, the entire team consists of only 2 people, having invested in 79 companies, nurturing 20 unicorns, including AI video generation platform Synthesia, clinical AI company Abridge, and quantum computing firm PsiQuantum. Over the past 12 months, the team completed 43 new investments, with 20 deals in the first five months of 2026 alone, showing a clear acceleration.

Leading NVentures is Mohamed "Sid" Siddeek, vice president and head of NVentures. Siddeek's resume reflects Nvidia's positioning of this department.

He worked at Morgan Stanley in the late 1990s, participating in Nvidia's IPO roadshow with Jensen Huang; then served nearly 10 years as head of TMT and telecom investments at Abu Dhabi's sovereign fund Mubadala; later led enterprise software and healthcare investments at SoftBank Vision Fund; returning to Nvidia in 2021 to establish NVentures.

Siddeek describes the scope of investments as: "The only two filters are, first, anything Nvidia can reach, and second, which fields are investable." In an interview with Global Corporate Venturing, he revealed that this means covering almost all AI-transformable industries such as healthcare, manufacturing, robotics, autonomous driving, and quantum, from underlying tools to application layers within NVentures' investment scope.

Three-Track Architecture: Corp Dev Handles Strategy, NVentures Focuses on Early-Stage, Inception Builds Ecosystem

Nvidia's external investment system consists of three distinct parts with clear roles.

The first layer is the Corporate Development team, led by Vishal Bhagwati, responsible for all strategic large-scale investments, joint ventures, and mergers.

The scale of these investments is entirely different from NVentures. Notable deals from late 2025 to mid-2026 include leading a $30 billion investment in OpenAI in February 2026 (as part of a roughly $110 billion funding round), with a future commitment to add up to $100 billion; a $10 billion commitment to Anthropic in November 2025; injecting $2 billion into Synopsys at the end of 2025; an additional $2 billion investment in CoreWeave in early 2026, along with a $6.3 billion cloud capacity procurement agreement; investing $2 billion in Nebius in March 2026; and a maximum $2 billion equity commitment to xAI.

According to CNBC, in just the first four months of 2026, the corporate development team led AI equity investments exceeding $40 billion. Nvidia's total investment in private companies and infrastructure funds in fiscal year 2025 was $17.5 billion.

The second layer is NVentures, led by Sid Siddeek, positioned as a traditional VC seeking financial returns.

Single investments range from a few million to tens of millions of dollars, mainly at Seed to Series B stages. Siddeek explicitly told Global Venturing that NVentures "mainly focuses on early-stage investments, while the corporate development team handles larger, more strategic investments." Behaviorally, NVentures primarily co-invests, with only about one-eighth of investments led by them, often participating as Nvidia's backer in rounds led by top VCs like Accel, a16z, Sequoia.

The third layer is NVIDIA Inception, essentially a startup accelerator program that does not directly invest but provides Nvidia hardware credit lines, technical support, marketing, and VC connection channels.

In 2025, Nvidia upgraded and launched the "VC Alliance," partnering with firms like Accel, Elaia, Partech, Sofinnova, distributing Nvidia DGX Cloud Lepton compute vouchers to its portfolio companies, extending Inception's reach into Europe.

There is a clear "funnel" relationship among the three.

Inception discovers early projects and introduces them into Nvidia's ecosystem. Those with investment potential may receive early checks of a few hundred thousand to tens of millions of dollars and, once they grow enough to influence Nvidia's strategic layout (becoming key customers, suppliers, or potential acquisition targets), they are "upgraded" to the corporate development team, entering cooperation frameworks worth billions or even trillions of dollars.

Recent NVentures Deals: Quantum, Reasoning Routing, AI Security

In May 2026, NVentures was quite active. In just the past month, four disclosed deals.

On May 22, French quantum computing company Alice & Bob announced NVentures' participation in its €100 million Series B extension round. Alice & Bob's core technology is based on fault-tolerant quantum architecture using "cat qubits," with deep collaboration with Nvidia's CUDA-Q, cuQuantum, Dynamiqs, and NVQLink hybrid classical-quantum computing stacks;

On May 26, AI model routing platform OpenRouter completed a $113 million Series B, with NVentures co-investing alongside Google CapitalG, Snowflake, etc. OpenRouter provides developers a unified API to access dozens of global model providers;

On May 28, AI inference infrastructure startup Tensormesh completed a $20 million seed extension round, with NVentures co-investing with CoreWeave, AMD;

On May 6, AI cybersecurity company Xbow completed a $35 million Series C extension, with NVentures participating.

Looking at the targets, NVentures' recent focus clearly tilts toward three directions: quantum computing (Alice & Bob, Quantinuum, PsiQuantum), AI biotech (Relation Therapeutics, Genesis Therapeutics), and AI agents and reasoning layers (OpenRouter, Tensormesh, etc.).

This aligns with Siddeek's statement that "any field Nvidia can reach" is within scope, matching Nvidia's ongoing investments in next-generation software stacks like CUDA-Q, CUDA-X, Triton.

Regionally, NVentures' European footprint is accelerating significantly. In 2025, it completed 14 European investments, double the 7 in 2024.

Three-Layer Investment Portfolio Panorama

If the portfolios of the three layers are mapped together, Nvidia's "capital radiation" across the AI ecosystem can be summarized into five main quadrants.

The foundational model layer includes OpenAI, Anthropic, xAI, Mistral, Cohere, Thinking Machines Lab, Reflection AI, Black Forest Labs. This layer is mainly funded by the corporate development team, with NVentures participating as a smaller co-investor.

The cloud and infrastructure layer includes CoreWeave, Nebius, Lambda, Crusoe, Nscale, Firmus Technologies. Also led by the corporate development team, with single investments often in the tens of billions of dollars, accompanied by long-term compute capacity procurement contracts.

The application and development tools layer features Cursor, Perplexity, Synthesia, Runway, Lovable, Together AI, Weka. NVentures' involvement here is relatively high, with smaller amounts.

Robotics and autonomous driving include Figure AI (latest valuation $39 billion), Wayve ($8.6 billion valuation). Both the corporate development team and NVentures jointly participate.

Quantum computing and biotech include PsiQuantum, Quantinuum, Alice & Bob, Relation Therapeutics. Mainly early investments led by NVentures, as a hedge against Nvidia's "post-GPU era" compute paradigm shift.

According to VC research firm F4 Fund, from 2025 to early 2026, at least 10 companies in Nvidia's (corporate + NVentures) investment rounds had valuations exceeding $1 billion, including OpenAI, Anthropic, xAI, Mistral, Figure AI, Cursor, Perplexity, Scale AI, Wayve.

Controversy: Burry's Shorting and the "Cyclical Financing" Question

However, Nvidia's vast external investment landscape is increasingly attracting skepticism. The most notable criticism comes from Michael Burry, the hedge fund manager famous for the movie "The Big Short."

According to Scion Asset Management's Q3 2025 13F filing, Burry built a short position against Nvidia and Palantir before September 30, 2025, including about 1 million shares of Nvidia put options, with a nominal exposure of approximately $187 million at the then-current stock price; and 50k put options on Palantir (each covering 100 shares), with an actual premium paid of about $9.2 million.

Burry posted on his X account "Cassandra Unchained" a photo with "The Big Short" poster, captioned "Sometimes, we see bubbles," and retweeted a Bloomberg chart on Nvidia's cyclical financing, directly criticizing Nvidia's capital deployment model.

His specific allegations are technical. He estimates on Substack that from 2026 to 2028, cloud providers like Microsoft, Google, Oracle, Meta, by extending Nvidia GPU depreciation periods, have collectively understated depreciation by about $176 billion, artificially inflating profits during that period. This accounting adjustment resonates with Nvidia's equity investments in clients, enabling buyers to report higher "book profits" to absorb larger capital expenditures, while also providing funds for purchasing Nvidia hardware.

At the institutional level, similar doubts are accumulating. The EU competition regulators explicitly included "cyclical expenditure risks" in Nvidia's investment system review scope in March 2026.

Seaport Research estimates that for every $1 Nvidia invests in equity, there is approximately $3.50 in downstream chip procurement revenue. Bloomberg's March 2026 "AI Cyclical Trading" feature mapped the flow of funds among Nvidia, CoreWeave, OpenAI, Oracle, and Anthropic into a dense network diagram.

Nvidia holds about 7% of CoreWeave, which uses Nvidia GPUs as collateral for financing, then uses cash to purchase more GPUs from Nvidia, which in turn signs a $6.3 billion cloud capacity procurement agreement, committed to absorbing CoreWeave's excess capacity until 2032; Nvidia commits up to $100 billion to OpenAI, which promises to buy Nvidia hardware and build a $300 billion data center via Oracle; Nvidia invests $10 billion in Anthropic, which commits to deploying Claude on Microsoft Azure, which then procures Nvidia Grace Blackwell and Vera Rubin systems.

Proponents' rebuttals also exist.

Asset management firm Janus Henderson characterizes this model as a "benign cycle," believing that in an era of extreme compute scarcity, binding supply and demand through "equity + long-term procurement contracts" is a reasonable business arrangement.

Morningstar's analysis points out that Nvidia's commitment to purchase excess capacity from CoreWeave effectively makes Nvidia bear the inventory risk, which could actually restrain short-term hardware sales impulses.

In this controversy, NVentures' position is quite delicate. Its early, small-scale, co-investment, industry-diverse investment style sharply contrasts with the corporate development team's "cyclical trading" model. The companies NVentures invests in, like Alice & Bob, Tensormesh, OpenRouter, are too small to form a "both Nvidia client and investor" cycle; their investments are more aligned with traditional corporate venture capital financial logic.

However, from the perspective of Nvidia's overall investment system, whether NVentures in some way acts as a "VC compliance cloak" disclosed externally—making it easier for outsiders to interpret Nvidia's investment activities as normal VC behavior rather than systemic seller financing—is an implicit question that Burry and EU regulators have not explicitly stated but are implicitly asking.

Nvidia's consistent official stance is that all investments are based on independent commercial judgment, not linked to hardware sales. But increasingly, market observers quote a question: in an era of compute shortages, is it just a coincidence that "equity and procurement contracts are entangled"? Trust itself becomes a core issue.

Original Link

Click to learn about BlockBeats' recruitment positions

Join the official BlockBeats community:

Telegram Subscription Group: https://t.me/theblockbeats

Telegram Discussion Group: https://t.me/BlockBeats_App

Twitter Official Account: https://twitter.com/BlockBeatsAsia