Nvidia Earnings: Revenue Boom Fuels AI Hype

Key Morningstar Metrics for Nvidia

What We Thought of Nvidia’s Earnings

Wide-moat Nvidia NVDA reported another quarter of outstanding revenue and earnings while providing investors with a forecast for the April quarter that was well ahead of our expectations. Leading cloud computing companies plan to boost their capital expenditures to satisfy demand for artificial intelligence training and inference, and it appears that virtually all this spending will fall into Nvidia’s pockets. More importantly, we anticipate healthy growth for Nvidia’s data center revenue beyond 2024, as Nvidia’s Cuda software should contribute to strong customer stickiness for existing AI models and workloads.

We raise our fair value estimate for Nvidia to US$730 from $480 as we boost our near-term and long-term revenue and profitability assumptions, thanks to the tremendous rise in AI computing demand. We reiterate our Very High Uncertainty Rating, as the competitive landscape in AI seems to be changing almost weekly.

Nvidia’s total revenue in the January quarter was $22.1 billion, up 22% sequentially, up 265% year over year, and ahead of guidance of $20 billion. Data center remains the only segment that should matter to investors. DC revenue of $18.4 billion was up 27% sequentially as more graphics processor packaging supply came online and was up 409% year over year and exceeded the $15.0 billion Nvidia earned in the DC segment all of last year.

We think fiscal 2025 will be a banner year for Nvidia as demand continues to exceed supply. In the April quarter, Nvidia forecasts total revenue of $24 billion, which implies over $21 billion of DC revenue, which would again be 5 times greater than the year-ago quarter. This revenue should again be virtually all profit for Nvidia, as the company forecasts a 76% GAAP gross margin, and we model a 63% GAAP operating margin. We anticipate revenue will rise by a couple of billion each quarter throughout fiscal 2025 for Nvidia as more chip supply comes online.