$NVDA commits $2B to Nebius ($NBIS), sparking debate over whether the neocloud deal is a smart infrastructure bet or a circular financing arrangement.

NVDA Drops $2B on Nebius (NBIS): Neocloud Bet or Circular Financing Trap?

by
Nebius neocloud investment - NVDA, $20억 베팅: 엔비디아가 네오클라우드 Nebius에 투자한 진짜 이유 AI Illustration
Generated by OpenAI DALL-E 3
💡 Key Takeaways

  • Nvidia deployed $2 billion into Nebius via pre-funded warrants — not a direct equity purchase — a structure that grants NVDA major shareholder status only upon exercise, giving it optionality while Nebius gets immediate capital to buy more NVDA chips.
  • Nebius is targeting a staggering 5GW of NVDA-powered compute capacity by end of 2030, beginning with a 1.2GW AI factory near Independence, Missouri, slated for power delivery in H2 2026.
  • With an identical $2 billion bet placed in CoreWeave just two months prior, Nvidia is systematically financing its own largest customers — a circular capital strategy that drew “mafia-like” criticism from investor Michael Burry in March 2026.

When Nvidia quietly structured a $2 billion private placement into Nebius on March 11, 2026, it wasn’t just another venture line item — it was the clearest signal yet that the company is no longer content being a chip supplier. NBIS shares surged roughly 16% that day, briefly vaulting the company’s market capitalization to approximately $24 billion, and the AI infrastructure world collectively took notice. The question isn’t whether the money is big; the question is what Nvidia is actually buying — and what it expects in return.

The architecture of this deal tells a more sophisticated story than a simple equity investment. Nvidia received pre-funded warrants for 21,065,936 Class A shares at an exercise price of just $0.0001 per share. In practical terms, Nebius collects the $2 billion upfront to fund its buildout, while Nvidia holds the right to convert those warrants into a major ownership stake at virtually zero cost. It is a financing structure engineered to keep both parties strategically intertwined for years to come.

Nvidia’s Neocloud Bet: Why the Nebius Deal Is More Than a Balance Sheet Entry

To understand why this structure matters, consider what Nebius actually does with the capital. The company’s capital expenditures had already exploded to $2.1 billion in Q4 2025, a staggering leap from just $416 million the prior quarter. That velocity of spending has one primary destination: procuring Nvidia’s own GPU clusters at industrial scale. The warrant structure therefore creates an elegant — and controversial — feedback loop.

Nvidia funds a customer. That customer deploys the capital to purchase Nvidia hardware. Nvidia’s revenue rises, its order backlog deepens, and it simultaneously builds a long-term equity stake in one of the fastest-growing AI infrastructure operators on the planet. Jensen Huang framed the thesis publicly as scaling “agentic AI” infrastructure to meet what he called “surging global demand for intelligence” — language that signals Nvidia’s ambitions now extend well beyond the chip fabrication layer into the broader AI compute stack.

5 Gigawatts by 2030: Building the AI Factory of the Future

selective focus photography of graph
Photo by m. on Unsplash

The scale of Nebius’s infrastructure ambition is almost difficult to contextualize without a physical reference point. The company’s publicly stated target is more than 5 gigawatts of compute capacity by the end of 2030 — a figure Bloomberg and Yahoo Finance translate to the equivalent of powering roughly 3.8 to 4 million American homes. This is not cloud infrastructure in the legacy sense; this is purpose-built AI compute fabric designed from the ground up for the agentic workloads of the next decade.

The first concrete milestone in that roadmap is already approved and underway. Nebius secured permits to develop a 1.2GW AI factory on a 400-acre site near Independence, Missouri, with power delivery targeted for the second half of 2026. That single facility alone represents a capital commitment that would have been unthinkable for any cloud operator just three years ago. The sheer land footprint and power density of the project underscores how fundamentally different next-generation AI data centers are from the hyperscaler warehouses of the 2010s.

The Missouri project also reflects a deliberate geographic diversification strategy. By anchoring heavy AI compute capacity in the Midwest — closer to established power grid infrastructure and away from coastal permitting bottlenecks — Nebius is effectively solving the site-selection problem that has stalled many competing buildouts. Power availability, not GPU supply, is increasingly the binding constraint in the AI infrastructure arms race, and Nebius appears to have gamed that equation early.

The Neocloud Wars: How Nebius and CoreWeave Are Competing for Nvidia’s Capital

Nebius AI data center 2026 - a long hallway with glass doors leading to another room
Photo by Paul Hanaoka on Unsplash

The Nebius deal did not exist in isolation. Just two months earlier, in January 2026, Nvidia committed an identical $2 billion to CoreWeave (CRWV) under the same strategic framework and the same 5GW-by-2030 compute target. The symmetry is not coincidental — it reveals a deliberate portfolio approach where Nvidia is seeding multiple neocloud operators simultaneously, hedging execution risk while ensuring that the largest buyers of its GPU clusters are also its strategic equity partners.

Both Nebius and CoreWeave occupy a critical structural position in the AI value chain: they are the primary intermediaries — the so-called neocloud layer — sitting between Nvidia’s silicon and the hyperscalers that ultimately consume the compute. Microsoft’s $17 billion deal with CoreWeave and Meta’s $3 billion agreement illustrate just how aggressively the hyperscaler tier is outsourcing raw GPU capacity rather than building it entirely in-house. Nebius and CoreWeave are the picks-and-shovels play for companies that need AI compute at scale but can’t or won’t absorb the capital intensity themselves.

The same week as the Nebius announcement, Nvidia also disclosed investments in Lumentum, Coherent, and Thinking Machines Lab — each at $2 billion — signaling that this is a systematic capital deployment campaign, not a one-off strategic bet. Nvidia is effectively financing the entire vertical stack of the agentic AI era, from optical interconnects to neocloud operators to frontier AI labs.

The Circular Financing Controversy: When Michael Burry Calls It “Mafia-Like”

Not everyone views Nvidia’s investment spree through a bullish lens. In March 2026, investor Michael Burry publicly accused Nvidia of operating with “mafia-like” behavior in the AI chip market, raising pointed questions about antitrust exposure. The accusation cuts to the structural heart of the warrant strategy: if Nvidia is simultaneously the dominant GPU supplier and a major equity investor in its own largest customers, the line between competitive market dynamics and vertical market control begins to blur uncomfortably.

The core concern is the circularity embedded in the model. Nvidia injects capital into a neocloud operator. That operator is contractually and economically incentivized to deploy the capital on Nvidia hardware — the same hardware that carries the highest gross margins in the semiconductor industry. Nvidia’s revenue inflates, its order visibility extends, and its equity stakes appreciate in lockstep with the AI infrastructure buildout it is simultaneously financing. Critics argue this is less a competitive market and more a self-reinforcing ecosystem that structurally disadvantages any GPU alternative.

Jensen Huang’s public framing deliberately sidesteps the circularity critique by anchoring the narrative in demand — specifically, the notion that the global appetite for agentic AI workloads is simply too large for any organic financing mechanism to address at the required speed. That argument has merit on its face; the scale and velocity of AI infrastructure spending in 2026 is genuinely unprecedented. But the antitrust optics of a company financing its own customers to buy its own products at scale are unlikely to escape regulatory attention indefinitely, particularly as the neocloud tier accumulates more market power in the AI compute stack.

The Nebius deal ultimately crystallizes a pivotal tension in the AI infrastructure economy: the companies building the physical foundation of the agentic AI era are doing so with capital provided by the single supplier they cannot afford to replace — and that supplier now holds warrants on their equity. Whether that arrangement represents inspired vertical integration or a structurally captive market is the question that will define the next phase of scrutiny around Nvidia’s dominance.

References

Stay Ahead of the Curve

Get the latest insights on AI, tech, and markets — delivered in both Korean and English.

Explore More Posts

Disclaimer: This article is for informational purposes only and is not investment advice. It does not recommend buying or selling any specific security or digital asset. All investment decisions and their outcomes are the sole responsibility of the reader.