NVIDIA

NVIDIA designs the accelerators (GPUs) that train and run AI models, and sells them increasingly as complete systems: GPUs, CPUs, high-speed networking (NVLink, InfiniBand, Spectrum-X Ethernet) and the software to run them. Taiwan Semiconductor manufactures the chips; NVIDIA designs, integrates and sells.

The customers are the hyperscale clouds (Microsoft, Amazon, Google, Meta, Oracle), AI labs, and increasingly governments and enterprises. Gaming GPUs, once the core, are now a small share.

Revenue in FY2026 (year ending January 2026) was $215.9B, up 65.5% after more than doubling in each of the two prior years.

The silicon is excellent, but the durable advantage is CUDA: twenty years of libraries, compilers and developer habit. Almost every AI framework and model is optimised for NVIDIA first. A rival chip has to match not only performance but the software that lets engineers use it without rewriting their code.

The moat is being tested from the inside: its largest customers design their own chips (Google's TPU, Amazon's Trainium, Microsoft's Maia, Meta's MTIA) precisely to reduce dependence, and AMD offers a credible GPU with an open software stack. Inference, which is larger and more price-sensitive than training, is where alternatives have the best chance.

These are software-like margins on hardware: gross margin 71% and operating margin 60% in FY2026. Free cash flow was $96.7B and stock pay a modest $6.4B, so nearly all of it belongs to owners. The capital is light because manufacturing is outsourced.

The catch is concentration and cyclicality. A handful of customers account for a large share of revenue, and their spending depends on their own returns from AI. Semiconductor history is a series of shortages followed by gluts.

If AI spending grows into a durable new layer of computing, NVIDIA is the toll road and still worth more than today's earnings suggest. Our base case assumes growth slows sharply from here and margins come down from the peak, which still leaves a very valuable company.

We would own it below our stated price; at prices that assume the peak lasts, we would wait.

• The capex cycle turning: hyperscalers cutting AI spending after a period of overbuilding.
• Custom chips and AMD taking inference, the larger long-run market.
• Export controls limiting sales to China and other markets.
• Concentration in a few customers and one manufacturer (TSMC, Taiwan).

• Gross margin falling below 60% as competition prices in.
• A major customer moving most of its inference to its own silicon.
• Hyperscaler capex guidance falling year on year.