NVIDIA Q4 Fiscal 2026 Earnings Summary
Agentic AI Inflection Drives Record Revenue as Blackwell Ramps and Vera Rubin Looms
Here’s Our Take
NVIDIA just reaffirmed that this is not a one-year AI spike — it is a multi-year infrastructure cycle.
Blackwell is ramping faster than expected. Networking is becoming a structural growth driver. Hyperscaler CapEx continues to rise. And Vera Rubin is positioned to extend performance leadership in H2 2026.
The absence of China revenue in both Q4 and guidance removes a key debate point and highlights how broad global demand has become.
The only real question is not demand — it’s sustainability and valuation.
NVIDIA is executing at a scale rarely seen in semiconductor history. Free cash flow generation is massive, backlog extends into 2027, and platform control spans compute, networking, and software.
This is not simply a chip company. It is the operating system of AI infrastructure.
The stock may trade sideways at times as expectations reset, but fundamentally, the earnings power trajectory continues upward.
Execution remains elite.
Quarter Overview
NVIDIA delivered another historic quarter, reinforcing its position as the foundational infrastructure provider of the AI industrial revolution.
Revenue surged to $68.1 billion (+73% y/y), accelerating sequentially and exceeding expectations. Data Center revenue reached $62.3 billion (+75% y/y), fueled by accelerating Blackwell adoption, explosive networking growth, and rising inference demand tied to agentic AI systems.
Importantly, management guided Q1 revenue to $78 billion (+7% above consensus) — and did so without assuming any Data Center compute revenue from China, removing a key geopolitical overhang.
The message was unmistakable: demand is broad, deep, and extending into calendar 2027. NVIDIA is no longer just riding AI hype — it is capitalizing on an infrastructure supercycle.
Key Financial Highlights
Revenue: $68.1B (+73% y/y)
Data Center Revenue: $62.3B (+75% y/y)
Networking Revenue: $11B (+3.5x y/y; $31B annual run rate)
Gaming Revenue: $3.7B (+47% y/y; below consensus due to supply constraints)
Non-GAAP Gross Margin: 75.2% (above consensus)
Non-GAAP Operating Margin: 67.7%
Free Cash Flow (Q4): $34.9B
Free Cash Flow (FY26): $97B
Capital Returned (FY26): $41B
Q1 Guidance
Revenue: $78B ±2% (well above consensus)
Gross Margin: ~75%
No China Data Center compute assumed
Management also noted that inventory and supply commitments now extend into 2027, reflecting unusually strong visibility.
Data Center: Blackwell Ramp Accelerates
Blackwell is scaling rapidly and now accounts for roughly two-thirds of Data Center revenue. Nearly 9 gigawatts of infrastructure are deployed across hyperscalers, model builders, enterprises, and sovereign customers.
Demand is no longer limited to chatbot training. It now spans:
Inference workloads
Agentic AI systems
Search and recommendation upgrades
Content generation
Sovereign AI infrastructure
Even Hopper and Ampere products remain sold out in the cloud — a rare dynamic during a product transition.
The shift toward inference is especially important. Blackwell delivers up to 50x performance per watt and materially lower cost per token versus Hopper, reinforcing NVIDIA’s leadership in performance-per-dollar economics.
Networking: The Quiet Superpower
Networking revenue hit $11B in Q4, up more than 3.5x year-over-year. NVLink scale-up switches and Spectrum-X Ethernet were major drivers as hyperscalers build giga-scale AI factories.
For the full year, networking exceeded $31B, more than 10x revenue since the Mellanox acquisition.
The key insight: AI infrastructure is not just GPUs. NVIDIA’s extreme co-design across chips, networking, and software creates switching costs and ecosystem lock-in.
Agentic AI: The Next Compute Multiplier
CEO Jensen Huang emphasized that the “ChatGPT moment of agentic AI has arrived.”
Partnerships and momentum include:
OpenAI GPT-5.3-Codex trained and inferencing on Grace Blackwell
Meta deploying millions of Blackwell and Rubin GPUs
$10B investment in Anthropic
Partnership momentum with xAI and others
Agentic systems (Claude Code, Codex, etc.) are driving token profitability and compute urgency. Compute now directly translates to revenue growth for customers — reinforcing hyperscaler CapEx expansion.
Analysts estimate the top five hyperscalers’ 2026 CapEx now approaches $700B, up roughly $120B year-to-date.
Vera Rubin: The Next Platform Leap
Rubin remains on track for H2 2026 production.
The Rubin platform includes:
Vera CPU
Rubin GPU
Next-gen NVLink fabric
ConnectX-9 SuperNIC
Spectrum-6 Ethernet
Rubin is expected to:
Train MoE models with ¼ the GPUs
Reduce inference token costs by up to 10x
Deliver improved resiliency and serviceability
Management expects broad adoption across cloud model builders.
The cadence is clear: Hopper → Blackwell → Rubin → next-gen leaps in performance per watt.
Beyond Data Center: Physical AI Expands
Physical AI contributed over $6B in FY26 revenue, including robotics, autonomous vehicles, and industrial AI.
Automotive revenue rose modestly, while partnerships with Waymo, Mercedes-Benz, and robotics leaders highlight long-term expansion into real-world AI.
This is early-stage optionality, but increasingly material.
Gross Margins and OpEx
Gross margins remain in the mid-70s — exceptional at this scale.
Operating expenses are rising as NVIDIA invests aggressively (annual R&D approaching $20B). Stock-based compensation will now be included in non-GAAP results starting Q1, modestly tempering EPS optics.
Inventory and purchase commitments increased meaningfully — reflecting strong forward demand visibility rather than inventory risk.
P.S. Know someone who’d appreciate smarter stock insights and clearer investing strategies? Forward this email or share this link: subscribe.triplegains.com
Triple Gains - Stock Analysis - Thematic Insights - Portfolio Strategy
DISCLAIMER: The content provided in this newsletter does not constitute investment advice, financial advice, trading advice, or any other form of personal recommendation. Nothing in this newsletter should be interpreted as a suggestion to buy, sell, or hold any investment or security. All content is for general informational purposes only and should not be relied upon for making investment decisions. Readers should conduct their own research and consult qualified financial advisors before making any investment decisions. To read our full disclaimer, click here.