Nvidia Commits $18 Billion in Equity Investments, Betting Big on AI Ecosystem Expansion
Nashnova编辑部
Nvidia has committed $18 billion in equity investments for the rest of this fiscal year, targeting AI model builders and infrastructure players; its private-company holdings hit $47.9 billion by end-July — more than double last year — as the chipmaker pivots from selling shovels to owning the mine.
Where is the $18 billion going?
The money targets three buckets: AI model developers, infrastructure financing firms, and other private companies — spanning the full stack from compute to applications.
By end-July, Nvidia held $47.9 billion in private-company equity, up from $22.3 billion at last fiscal year-end — more than doubled in a year.
This means → Nvidia is no longer just a GPU vendor. It is using real capital to embed itself in its customers' ownership structures.
Which deals have landed recently?
Infrastructure: $1.5 billion invested in SB Energy, plus a stake in Cloverleaf Infrastructure — both tackling the land and power bottlenecks choking data-center buildout.
Financing: Nvidia co-organized a Wall Street consortium to raise over $500 billion in external funding for AI infrastructure.
Models and chips: a multi-billion-dollar deal with coding startup Poolside; talks underway with AI search startup Perplexity and Korean inference-chip startup Rebellions; $2.94 billion in cash already paid for AI hardware startup Groq.
Why not just acquire these companies outright?
Analyst Luke Lango notes that licensing, talent partnerships, and minority stakes let Nvidia sidestep the antitrust scrutiny and integration risk of full acquisitions.
In plain terms = buying a company whole invites regulators to the door. A small stake locks in the relationship without tripping the wire.
What is the "circular financing" risk?
The core concern: Nvidia is chip supplier, financier, and equity holder to many of the same companies — three roles stacked on one counterparty.
Lango's warning: "If demand slows, it will hit Nvidia's revenue and its equity portfolio at the same time."
This means → a cooling in the AI investment cycle would compress both the income side and the asset side simultaneously — a high concentration of risk.
How does management respond?
CFO Colette Kress said on the earnings call: "We believe these investments, measured by demand strength, the business they create for us, the ecosystem built on the Nvidia platform, and the equity returns on capital deployed, will deliver strong performance with limited risk."
In plain terms = management's logic is that every dollar invested comes back as chip orders, plus equity upside — "invest once, collect twice."
What is the key variable to watch?
The backdrop: Amazon, Google, and Microsoft — Nvidia's biggest customers — are developing their own chips. Nvidia is therefore racing to extend beyond chips into infrastructure, software, and model layers.
This reflects a deeper anxiety — when your largest buyers start making their own hardware, selling chips alone is not enough. You need an ecosystem to keep them locked in.
Whether this strategy can sustain ecosystem stickiness amid the custom-chip wave is the core variable going forward.
市场有风险，内容仅供研究参考，不构成投资建议。