Nvidia (NVDA 5.27%) reported fourth-quarter fiscal 2025 earnings on February 26, 2025, delivering another record quarter amid surging artificial intelligence (AI) demand. The company posted Q4 revenue of $39.3 billion, up 78% year-over-year, and provided strong guidance for continued growth.
Blackwell Launch Demonstrates Unprecedented Scale
The new Blackwell architecture saw remarkable initial uptake, with $11 billion in revenue during its first quarter of availability. This represents the fastest product ramp in Nvidia's history.
We successfully and incredibly ramped up Grace Blackwell, delivering some $11 billion of revenues last quarter. We're going to have to continue to scale as demand is quite high, and customers are anxious and impatient to get their Blackwell systems.
-- Jensen Huang, CEO
The complex manufacturing process spans 350 plants producing 1.5 million components per rack, showcasing Nvidia's ability to execute at scale. Early deployments include major cloud providers like Microsoft Azure, Google Cloud Platform, AWS and Oracle Cloud Infrastructure.
AI Workloads Evolving Beyond Initial Use Cases
Management highlighted three distinct scaling laws driving compute demand: pre-training, post-training, and inference time scaling. Post-training tasks can now require more compute than initial training.
The amount of computation you use for post-training is actually higher than pre-training. ... The amount of tokens generated [and] the amount of inference compute needed is already 100x more than the one-shot examples and the one-shot capabilities of large language models in the beginning. And that's just the beginning.
-- Jensen Huang, CEO
This evolution suggests expanding infrastructure requirements as AI applications mature and become more sophisticated.
NASDAQ: NVDA
Key Data Points
Data Center Growth Remains Strong Despite Geographic Shifts
Data center revenue reached $35.6 billion in the fourth quarter, up 93% year-over-year. While China revenue remains around half of pre-export-control levels, strong U.S. demand helped drive 16% sequential growth.
Data center revenue of $35.6 billion was a record, up 16% sequentially and 93% year-on-year, as the Blackwell ramp commenced and Hopper 200 continued sequential growth.
-- Colette Kress, CFO
The company expects China to maintain its current percentage of revenue going forward, suggesting other regions can sustain growth trajectories.
Product Roadmap Shows Consistent Innovation Pace
Management confirmed the upcoming launch of Blackwell Ultra in the second half of 2025, maintaining its annual product cadence. They also referenced development of the next-generation "Vera Rubin" architecture.
The next train is on an annual rhythm and Blackwell Ultra with new networking, new memories and of course, new processors, and all of that is coming online. We have been working with all of our partners and customers, laying this out.
-- Jensen Huang, CEO
The system architecture compatibility between Blackwell and Blackwell Ultra should enable smoother customer transitions compared to previous generations.
Margin Profile Evolving with Product Mix
Gross margins will remain in the low 70s as Blackwell ramps, with improvement opportunities once initial production stabilizes.
During our Blackwell ramp, our gross margins will be in the low 70s. At this point, we are focusing on expediting our manufacturing ... to make sure that we can provide to customers as soon as possible.
-- Colette Kress, CFO
Management cited multiple opportunities to improve costs and margins once initial production goals are met.
Looking Ahead
Nvidia's management expressed confidence in sustained demand, highlighting the expanding role of AI across industries. The company appears well-positioned with its full stack approach spanning hardware, networking, and software. As CEO Jensen Huang emphasized:
No technology has ever had the opportunity to address a larger part of the world's GDP than AI. No software tool ever has. ... When you take a step back and look at it from that perspective, we're really just in the beginning.
