Nvidia just became the first company in history to cross a $5 trillion market capitalization. The Information reports the chipmaker passed the milestone this week, extending a run that has turned Jensen Huang’s company into the most valuable business on the planet by a wide margin. No other public company has ever touched this number.
What stands out here is the speed. Nvidia crossed $1 trillion in May 2023. It hit $2 trillion in February 2024. Then $3 trillion that June, $4 trillion in July 2025, and now $5 trillion. That’s five times in roughly two and a half years. Apple and Microsoft, the only other companies sitting in the $3 to $4 trillion range, took decades to build the businesses that got them there.
What’s driving the climb
The answer is simple: Nvidia sells the picks and shovels for the AI gold rush, and demand isn’t slowing down. Every major AI lab, hyperscaler, sovereign AI project, and enterprise data center buildout runs on Nvidia GPUs. The H100, the H200, the Blackwell B200, and now the upcoming Rubin generation are the default substrate for training and serving frontier models.
A few numbers that explain the valuation:
- Microsoft, Meta, Google, and Amazon are collectively spending well over $300 billion in 2025 on AI infrastructure, with the lion’s share flowing to Nvidia silicon.
- OpenAI, Anthropic, xAI, and other labs are signing multi-billion dollar compute deals that lock in Nvidia capacity for years.
- Sovereign AI projects across the Middle East, Europe, and Asia are adding tens of billions more in committed orders.
- Nvidia’s data center revenue has been growing at triple-digit percentages year over year for multiple consecutive quarters.
The moat is software too. CUDA, the company’s developer platform, is where every serious AI engineer works. Switching to AMD or custom silicon means rewriting tooling, retraining teams, and accepting performance gaps. That lock-in is why Nvidia commands gross margins north of 70 percent on its top GPUs.
Why this matters for the AI industry
A $5 trillion Nvidia tells you something important about where AI economics are concentrated right now. The labs and hyperscalers building models capture headlines. The company supplying their compute captures the cash. For practitioners and operators, this has direct consequences.
- GPU pricing isn’t coming down meaningfully anytime soon. If you’re building on AI infrastructure, budget accordingly.
- The secondary market for compute (rental tiers, regional clouds, GPU brokerages) keeps growing because buying directly is hard and slow.
- Every alternative chip story (Google TPUs, AWS Trainium, Cerebras, Groq, custom ASICs from Meta and OpenAI) now carries more weight, because the industry has a clear interest in breaking Nvidia’s pricing power.
The political angle is also sharpening. A single American company controlling the substrate of global AI is a national security topic in Washington and a sovereignty topic in Brussels, Beijing, Riyadh, and Delhi. Export controls on Nvidia chips to China have already shaped the geopolitical map. Expect more of that, not less.
What comes next
The Blackwell ramp is the immediate story. Nvidia is shipping its newest architecture into a market that wants every unit it can get. Rubin, the next generation after Blackwell, is on track for 2026 and is expected to widen the performance lead again.
The risks are real but mostly long-dated. A serious slowdown in AI capex, a credible competitor at scale, a regulatory crackdown, or a macro shock could all hit the stock. None of those look imminent based on current order books.
For now, one company sells the engine that powers the AI era, and the market just priced that reality at $5 trillion. The chipmaker has effectively become a proxy for AI itself. When you watch Nvidia, you’re watching the industry.
More details at The Information.