NVIDIA broke records once again. The company reported revenue of $68.1 billion for the fourth quarter of fiscal year 2026 (ended January 25, 2026), beating Wall Street expectations by $3 billion. This represents 73% growth year over year and 20% growth quarter over quarter.
Key numbers
| Metric | Q4 FY2026 | Q4 FY2025 | Change |
|---|---|---|---|
| Total Revenue | $68.1B | $39.3B | +73% |
| Data Center | $62.3B | $35.6B | +75% |
| Gaming | $2.5B | $3.3B | -24% |
| Net Income (GAAP) | ~$43B | ~$22B | +94% |
| Full Year Revenue (FY2026) | $215.9B | $130.5B | +65% |
| Q1 FY2027 Guidance | $78B (+/-2%) | — | — |
Data centers: the real engine
The data center segment generated $62.3 billion, representing 91.5% of total company revenue. That is 75% growth year over year. This segment includes GPUs used for AI model training and inference by companies like Microsoft, Google, Amazon, and Meta.
After testing servers with Blackwell GPUs for an inference project, I can confirm the performance per watt is impressive. The demand for these GPUs is not speculative — there is real workload behind it.
Gaming is now just 11.45% of NVIDIA
Perhaps the most striking figure for gamers: the gaming segment generated only $2.5 billion, a 24% decline year over year. Gaming represented more than 50% of NVIDIA revenue just four years ago. Today it is just 11.45%.
This explains why consumer GPUs like the RTX 5090 and 5080 have increasingly high prices — NVIDIA no longer depends on the consumer market for growth.
Blackwell dominates, Rubin is coming
The Blackwell architecture was the main driver of the quarter. Jensen Huang, NVIDIA CEO, confirmed that the next architecture Rubin is already in development and will begin shipping to customers in the second half of 2026.
NVIDIA's upgrade cycle is accelerating: from 2-year cycles (Ampere to Hopper) to roughly 1-year cycles (Hopper to Blackwell to Rubin). This keeps pressure on competitors AMD and Intel.
$78B guidance for next quarter
NVIDIA guided for revenue of $78 billion in Q1 FY2027, which would imply 14.5% sequential growth. This guidance crushed concerns about an AI demand slowdown and sent a clear message: the AI era is just getting started for NVIDIA.
What this means for investors
NVIDIA stock (NVDA) had a mixed reaction after the results — a common pattern when expectations are already sky-high. However:
- Fundamentals are rock solid: $43B net income in a single quarter is extraordinary
- Guidance beats consensus: $78B vs the ~$75B analysts expected
- Backlog keeps growing: demand exceeds supply for Blackwell GPUs
Troubleshooting for developers
I cannot get Blackwell GPUs for my project
Demand exceeds supply. Alternatives include using cloud providers (AWS, Azure, GCP) that have priority access, or considering previous-gen GPUs like H100 that remain capable for many workloads.
# Check NVIDIA GPU availability on your cloud provider
# AWS
aws ec2 describe-instance-types --filters "Name=instance-type,Values=p5*" --query "InstanceTypes[].InstanceType"
# Azure
az vm list-sizes --location eastus --query "[?contains(name, 'ND')]"GPU cloud costs are too high
Consider spot or preemptible instances that cost 60-70% less. For inference, use quantized models (INT8/INT4) that require less VRAM. Tools like vLLM optimize memory usage.
Is AMD MI300X a viable alternative?
Yes, for certain workloads. The software ecosystem (ROCm) has improved significantly, but CUDA still has an edge in compatibility and tooling. In my experience, AMD is competitive for inference; for large-scale training, NVIDIA still dominates.
Looking ahead
With $215.9 billion in annual revenue and guidance pointing to over $300B in the next fiscal year, NVIDIA has become the most important company of the AI era. As long as Big Tech keeps investing $650B annually in AI infrastructure, NVIDIA will remain the primary beneficiary.
