By James Aspinwall | February 26, 2026
NVIDIA just posted $68.1 billion in quarterly revenue. That’s not a typo. A company that made $39.3 billion in the same quarter last year grew 73% year-over-year to a number that would have been an impressive annual figure for most tech companies just a few years ago. Full-year revenue hit $215.9 billion, up 65% from fiscal 2025.
Let’s break down what matters, what’s concerning, and whether NVDA at ~$201/share and a $4.7 trillion market cap still has room to run.
The Numbers That Matter
Revenue Acceleration Is Real
| Metric | Q4 FY26 | Q3 FY26 | Q4 FY25 | Q/Q | Y/Y |
|---|---|---|---|---|---|
| Revenue | $68.1B | $57.0B | $39.3B | +20% | +73% |
| Data Center | $62.3B | $51.0B* | $35.6B* | +22% | +75% |
| Operating Income | $44.3B | $36.0B | $24.0B | +23% | +84% |
| Net Income | $43.0B | $31.9B | $22.1B | +35% | +94% |
| EPS (Diluted) | $1.76 | $1.30 | $0.89 | +35% | +98% |
The Q/Q acceleration from 20% revenue growth to 35% net income growth tells you about operating leverage. When your gross margins are 75% and your product is software-defined silicon, every incremental dollar of revenue drops disproportionately to the bottom line.
Margins Are Expanding, Not Compressing
This is the number that skeptics should pay attention to. Gross margin improved to 75.0% GAAP (75.2% non-GAAP) — up 1.6 points sequentially and 2.0 points year-over-year. The full-year margin of 71.1% was dragged down by the $4.5 billion H20 charge in Q1 (China export restrictions), but the exit rate tells the real story: NVIDIA is getting more profitable as it scales, not less.
Q1 FY27 guidance projects 74.9-75.0% gross margins, confirming this isn’t a one-quarter anomaly.
Cash Flow Is Extraordinary
- Operating cash flow: $36.2B in Q4, $102.7B for the full year
- Free cash flow: $34.9B in Q4, $96.6B for the full year
- Cash + marketable securities: $62.6B on the balance sheet
NVIDIA generated $96.6 billion in free cash flow this year. They returned $41.1 billion to shareholders through buybacks and dividends, and still grew their cash pile by ~$19 billion. The balance sheet is a fortress: $157.3 billion in shareholders’ equity against only $8.5 billion in total debt.
Segment Deep Dive
Data Center: 91% of Revenue and Accelerating
Data Center revenue of $62.3 billion represents 91.4% of total revenue. This concentration would normally be a risk factor, but when your addressable market is “every company building AI infrastructure,” concentration is just focus.
Key signals:
- Meta partnership — multiyear, multigenerational deal spanning millions of Blackwell and Rubin GPUs
- Inference economics — Blackwell Ultra delivers 50x better performance and 35x lower cost for agentic AI vs. Hopper
- Cloud lock-in — AWS, Google Cloud, Azure, and OCI will be first to deploy Vera Rubin instances
The inference market is the next growth vector. Training dominated the narrative in 2023-2024, but Jensen Huang’s emphasis on “agentic AI” and inference cost reduction signals where the puck is heading. Every enterprise deploying AI agents needs inference compute — and the economics favor NVIDIA’s platform.
Gaming & AI PC: Cyclical but Growing
Gaming revenue of $3.7B was down 13% Q/Q (seasonal normalization after holiday) but up 47% Y/Y. Full-year gaming hit a record $16.0B (+41%). The RTX platform is evolving from a gaming product to an AI PC platform — DLSS 4.5 and local LLM inference capabilities make the GPU relevant beyond gaming.
Professional Visualization: The Sleeper
$1.3B in Q4, up 159% Y/Y. Small in absolute terms but the growth rate is remarkable. DGX Spark and RTX PRO workstations are creating a new category: the personal AI supercomputer for professionals. This segment could triple again.
Automotive & Robotics: The Long Game
$604M in Q4 — small today, but the DRIVE Hyperion ecosystem expansion and partnerships (Mercedes-Benz, the Cosmos/Isaac robotics stack) are planting seeds. Physical AI is a decade-long opportunity. NVIDIA is positioning to be the platform, not just a chip vendor.
Competitive Moat Assessment
NVIDIA holds ~86% of AI data center chip revenue
The competitive landscape:
| Player | Position | Threat Level |
|---|---|---|
| AMD | ~15-20% AI accelerator target for 2026, MI350 competitive on price | Medium |
| Intel | Gaudi line struggling (~$500M), losing CPU share | Low |
| Custom ASICs (Google TPU, Amazon Trainium) | Growing but hyperscaler-specific | Medium-Low |
| Groq | Inference-focused, now NVIDIA licensee | Low |
The CUDA ecosystem is NVIDIA’s real moat. It’s not just hardware — it’s the software stack, the developer tools, the libraries (CUDA-X, Omniverse, NeMo, BioNeMo), and the network effects of millions of developers trained on NVIDIA’s platform. AMD can compete on silicon; competing on ecosystem is a different game entirely.
The Groq licensing deal is telling: rather than compete, Groq chose to license NVIDIA technology. When your competitors become your customers, your moat is deep.
Valuation: The Hard Question
At ~$201/share and a $4.7T market cap:
| Metric | Value |
|---|---|
| Market Cap | ~$4.7T |
| Trailing P/E (GAAP) | ~41x ($4.90 EPS) |
| Forward P/E (est.) | ~28-30x |
| Price/FCF | ~49x |
| EV/Revenue (TTM) | ~22x |
Bull case: Q1 FY27 guidance of $78B implies ~$340B+ annualized revenue run rate. If NVIDIA hits $350-400B in FY27 revenue with 75% gross margins, you’re looking at ~$7-8 EPS, putting the forward P/E at 25-29x. For a company growing 50%+, that’s reasonable. Analyst consensus target of $257-269 implies 30%+ upside.
Bear case: The 41x trailing P/E prices in perfection. Any deceleration — China restrictions tightening, hyperscaler capex pullback, custom ASIC adoption — could compress multiples. Operating expenses grew 45% Y/Y in Q4, and the SBC inclusion in non-GAAP starting FY27 ($1.9B/quarter) will change optics. Inventory nearly doubled to $21.4B — either they’re building for massive demand, or there’s risk of oversupply.
Base case: Revenue grows 50-60% in FY27 to $330-345B. Margins hold at 74-75%. EPS of $6.50-7.50. Fair value range: $195-260. The stock is fairly valued today with upside if execution continues.
Key Risks
-
China revenue exclusion — NVIDIA isn’t counting any Data Center compute revenue from China in Q1 guidance. This is either conservative or signals permanent market loss.
-
Customer concentration — The hyperscalers (Microsoft, Meta, Google, Amazon) likely represent 40-50%+ of revenue. A capex cycle pause would be felt immediately.
-
Inventory build — $21.4B in inventory (up from $10.1B a year ago) needs to convert to revenue. If demand disappoints, write-downs follow.
-
Valuation compression — At $4.7T, NVIDIA needs to keep delivering 50%+ growth to justify its multiple. The law of large numbers is undefeated.
-
The SBC accounting change — Including stock-based compensation in non-GAAP numbers starting FY27 is the right thing to do (and long overdue across the industry), but it will make Y/Y non-GAAP comparisons look worse and increases reported opex by ~$1.9B/quarter.
The Verdict
NVIDIA is executing at a level that’s historically unprecedented for a company of this scale. $96.6 billion in annual free cash flow. 75% gross margins. 73% revenue growth. These are numbers that belong in a textbook chapter on monopoly economics.
For long-term investors: NVIDIA remains the best pure-play on AI infrastructure. The Blackwell-to-Rubin product roadmap provides multi-year visibility. The software ecosystem (CUDA, NeMo, Omniverse) creates switching costs that hardware alone cannot. If you believe AI infrastructure spending is secular (not cyclical), NVIDIA is the toll booth on the highway.
For value-conscious investors: The stock is fairly priced, not cheap. Wait for a pullback to the $160-175 range for a margin of safety, or dollar-cost average. The Q1 guidance beat is already partially priced in.
For traders: The post-earnings move will depend on guidance tone, not the backward-looking beat. Watch the $78B Q1 revenue guide — if management signals acceleration beyond that, the stock breaks $220. If the China exclusion narrative dominates, expect consolidation.
Rating: Overweight. The fundamentals justify the valuation at current levels, with meaningful upside if the agentic AI thesis plays out at the pace Jensen Huang is projecting. The primary risk isn’t competition — it’s whether the world’s appetite for AI compute continues to grow exponentially, or whether we hit a plateau. So far, every skeptic who bet on that plateau has been wrong.
Disclaimer: This analysis is for informational purposes only and does not constitute investment advice. Always do your own research and consult a financial advisor before making investment decisions.