NVDA

Nvidia delivered a blockbuster quarter that exceeded even the elevated expectations set by the pre-earnings brief, reporting $57 billion in revenue (up 62% year-over-year) and record data center revenue of $51 billion (up 66% YoY). The company's gross margin of 73.4% exceeded guidance, driven by favorable data center mix and improved cycle times. Most significantly, management provided forward guidance of $65 billion for Q4 (implying 14% sequential growth), and disclosed visibility to $500 billion in cumulative Blackwell and Rubin revenue through the end of calendar 2026, a figure that management explicitly stated will grow as new orders continue to flow in. The Blackwell architecture transition is executing flawlessly, with GB300 now contributing roughly two-thirds of Blackwell revenue and already shipping to major cloud service providers. Networking revenue surged 162% year-over-year to $8.2 billion, reflecting the onset of NVLink scale-up and broad adoption of Spectrum X Ethernet across hyperscaler AI factories. Management's commentary revealed three critical dynamics underpinning the infrastructure build: the transition from CPU to GPU-accelerated computing (addressing hundreds of billions in existing cloud spend), the replacement of classical machine learning with generative AI in hyperscaler workloads (driving revenue gains for cloud providers), and the emerging wave of agentic AI systems (a net-new category driving incremental consumption). Jensen Huang emphasized that Nvidia is the only architecture capable of addressing all three transitions simultaneously, and that the company's competitive moat rests not just on performance but on ecosystem breadth, the ability to run every frontier AI model, support every cloud, and provide diverse offtake for new infrastructure providers. The company also announced strategic partnerships with Anthropic (committing up to 1 gigawatt of compute capacity) and a deepened relationship with OpenAI, alongside a partnership with Saudi Arabia's Public Investment Fund and Humane to deploy up to 150,000 GB300 accelerators. Management's tone was notably confident about supply chain resilience, with Huang stressing that Nvidia's long-standing relationships with TSMC, memory vendors, and ODMs have enabled the company to secure supply well in advance and maintain credibility through balance sheet strength. On margins, Colette Kress guided to holding gross margins in the mid-seventies for fiscal 2027 despite rising input costs, citing cost improvements, cycle time optimization, and favorable mix as offsetting factors. The only material headwind disclosed was China, where geopolitical constraints continue to prevent shipments of competitive data center compute products, though management reiterated its commitment to advocacy for U.S. competitiveness globally. Investors should focus on three forward-looking metrics: the trajectory of the