Nvidia Posts First Trillion-Dollar Quarter. Jensen Huang Says Demand Still Exceeds Supply.

H200 chips selling faster than they can be made. Every major AI lab is competing for allocation.

The Numbers

Nvidia's Q4 FY2026 results shattered records:

MetricQ4 FY2026Q4 FY2025Change Revenue$112B$62B+81% Net Income$67B$29B+131% Data Center Revenue$98B$47B+108% Gross Margin76.2%74.1%+2.1pp First trillion-dollar quarter in tech history. Annualized revenue run rate exceeds $400B.

---

Where the Money Comes From

Revenue Breakdown

SegmentRevenue% of Total Data Center (AI)$98B87.5% Gaming$7B6.3% Professional Visualization$4B3.6% Automotive$2B1.8% Other$1B0.9%

Data Center Customers

Customer TypeEstimated % Hyperscalers (Microsoft, Google, Meta, Amazon)55% Cloud providers20% AI labs (OpenAI, Anthropic, xAI)15% Enterprise10%

---

The Jensen Huang Quotes

On Demand

'Every major technology company in the world is racing to build AI infrastructure. Demand for our chips exceeds supply and will for the foreseeable future. We're in the iPhone moment of AI.'

On Competition

'AMD, Intel, custom silicon—they're all welcome to try. CUDA has 4 million developers. The ecosystem is the moat.'

On the Future

'We're not in the chip business. We're in the accelerated computing business. The total addressable market is all computing. That's a $3 trillion annual opportunity.'

---

The Chip Situation

Current Products

ChipPerformancePriceAvailability H2004 petaflops$30,00012-month waitlist H1003.9 petaflops$25,0006-month waitlist A1002.5 petaflops$15,000Available B100 (announced)5+ petaflops$40,000+H2 2026

Production Capacity

- Current: ~4 million H-series chips/year - 2026 target: 8 million chips/year - Constraint: TSMC advanced packaging capacity - Solution: Building with TSMC, Samsung, Intel

---

Who's Buying

The Hyperscaler Race

CompanyEstimated AI Capex 2026Primary Use Microsoft$80BAzure AI, OpenAI Meta$50BLlama training, Instagram/FB Google$45BGemini, Cloud AI Amazon$40BAWS AI, Alexa ByteDance$20BTikTok, China AI

The AI Lab Buildout

LabEstimated GPU Count OpenAI100,000+ H100s Anthropic50,000+ H100s xAI100,000+ H100s Google DeepMind200,000+ TPUs + GPUs

---

The Competition Landscape

Alternatives to Nvidia

CompetitorProductStatus AMDMI300X10-15% market share IntelGaudi 3<5% market share GoogleTPU v5pInternal only AmazonTrainium 2Internal only GroqLPUInference only

Why Nvidia Still Wins

1. CUDA ecosystem: 15 years of software optimization 2. Network effects: Everyone trains on CUDA, so everyone deploys on CUDA 3. Full stack: Chips + networking + software 4. Execution: They ship on time, at scale 5. Roadmap: B100, B200 already announced

---

Stock Performance

MetricValue Market Cap$4.2 trillion P/E Ratio65x 1-Year Return+180% 5-Year Return+1,800% YTD 2026+45% Nvidia is now the world's most valuable company, surpassing Apple and Microsoft.

---

The Bull Case

- AI infrastructure buildout continues for 5+ years - Total addressable market expands as AI reaches more industries - Software ecosystem is unassailable - Inference demand grows as AI deployments scale - New markets: robotics, autonomous vehicles, scientific computing

The Bear Case

- Customers building custom silicon to reduce dependence - AMD/Intel eventually close the gap - AI spending peak may be sooner than expected - Regulatory risk (China restrictions, antitrust) - Valuation assumes flawless execution

---

The Bigger Picture

This Is Historically Unusual

CompanyPeak Revenue/QuarterContext Nvidia$112BAI infrastructure Apple$124BHoliday iPhone peak Amazon$170BHoliday retail peak Walmart$165BHoliday retail peak

Nvidia—a chip company—is approaching Apple-level revenue. That's unprecedented.

The AI Arms Race in Numbers

Total AI infrastructure spend 2026: ~$500B Nvidia's share: ~60% That's $300B going to one company.

---

Bottom Line

Nvidia's dominance is real and likely to continue for the foreseeable future. The question isn't whether they're winning—they are. The question is:

1. How long does the AI buildout continue? 2. Can anyone build a competitive alternative? 3. Is this the new normal or a bubble?

For now, Jensen Huang is sitting on the most valuable company in history, built on selling the infrastructure for the AI revolution.

---

Related Reading

- NVIDIA's Blackwell Chips Face 12-Month Backlog as AI Demand Surges - Jensen Huang Is Now Worth $120 Billion. He Wears the Same Leather Jacket. - NVIDIA Is Now Worth More Than Every Country's GDP Except the US and China - NVIDIA's Blackwell Chips Are Delayed Again—Here's Why It Matters - NVIDIA's Blackwell B300 Ships: 10x Faster AI Training Is Here