Nvidia's 9.3% Stake in Nebius, Explained

Nvidia quietly became the largest shareholder in Nebius Group NV, disclosing a 9.3% stake in the Amsterdam-based AI cloud company via SEC filing on July 21, 2026. Nebius, which most Americans have never heard of, now carries a $46 billion market cap and has seen its stock climb 250% over the past twelve months. Shares jumped another 7% on the disclosure.

On its own, the investment looks like a chip giant backing a fast-growing customer. Zoom out, and it fits a much bigger pattern: Nvidia has now invested meaningfully in OpenAI ($30 billion), Anthropic ($30 billion), and the compute infrastructure layer underneath them. It's not just selling GPUs anymore — it's buying pieces of every floor in the AI building.

What Nebius Actually Does

Nebius is what the industry calls a "neocloud" — a company that builds and operates the massive data centers (power, cooling, GPU orchestration, the works) needed to train and run AI models at scale. Think of it as the physical engine room behind products like ChatGPT or Meta's AI tools. It's one of Europe's few indigenous AI infrastructure players in a market dominated by AWS, Google Cloud, and Microsoft Azure.

The company has emerged as a serious contender in 2026. Meta committed $27 billion to Nebius AI infrastructure in March 2026 — a radical bet by Meta to own its own compute layer rather than renting from hyperscalers. Nebius also raised $775 million in its first senior secured debt facility, backed by deployed GPU infrastructure and contracted cash flows. Freedom Capital Markets called the debt raise a "positive catalyst" for the company. When you can borrow against actual AI models running on your servers, you've crossed a threshold from startup into real infrastructure business.

Nvidia's Multi-Layered Relationship

Here's where it gets interesting — and complicated. Nvidia is simultaneously:

• Supplier: Nebius buys Nvidia's GPUs to fill its data centers.
• Investor: Nvidia now owns 9.3% of the company, on top of a previously announced $2 billion investment.
• Partner: The two companies collaborate on AI factory design and deployment.

This is like a semiconductor fab supplier in the 1980s taking a stake in a PC maker — hedging by making money on components while ensuring a major customer stays solvent and competitive. Except Nvidia is doing this simultaneously across the entire AI value chain: chips, models, and infrastructure.

The Circular Economy of AI

Nvidia's investment pattern reveals a deliberate strategy. It has now put capital into all three of the "AI labs" reshaping computing — OpenAI, Anthropic — and into the physical compute layer those labs depend on. Every dollar these companies spend on infrastructure flows back to Nvidia as GPU revenue. Every stake Nvidia takes gives it a seat at the table and upside if those companies succeed.

The tension worth watching: Nebius's entire value proposition depends on Nvidia's chips. Nvidia could theoretically build competing inference and compute infrastructure itself. Is Nvidia genuinely betting on Nebius's independence, or is it locking in a customer while maintaining optionality to eventually compete with — or acquire — it?

Why the European Angle Matters

Nebius being Dutch is more than a footnote. The AI infrastructure race has been almost entirely a US story so far, with AWS, Azure, and Google Cloud controlling the bulk of global compute capacity. Nebius gaining traction — and winning a $27 billion Meta contract — signals a broader shift that few observers are discussing: major tech companies are willing to go outside the US cloud oligopoly when the capacity and economics are right. That has geopolitical implications as AI compute becomes a strategic asset.

Key Takeaways

• Nvidia's 9.3% stake makes it Nebius's largest shareholder and validates the neocloud model.
• Combined with $30 billion into OpenAI and $30 billion into Anthropic, Nvidia is investing across every layer of the AI stack.
• Nebius's $27 billion Meta deal and $775 million debt raise show the company is scaling fast — but it remains deeply dependent on Nvidia's chips.
• This is one of the clearest examples yet of AI's circular economy: Nvidia sells the GPUs, invests in the buyers, and profits either way.

The question worth watching: does Nvidia stay a kingmaker in AI infrastructure, or does it eventually decide to become the king?

Educational content only — not financial advice.