Nvidia faces growing chip competition but retains a wide lead

Nvidia remains the dominant AI chip supplier despite growing competition from GPU makers, custom silicon designers and its own cloud customers. It reported $89.0 billion in data center revenue for the quarter ended July 26, 2026. Morgan Stanley estimated in March 2026 that Nvidia held roughly 85% of AI processor revenue, underscoring the gap challengers still face.

AMD is the closest direct GPU rival, with Instinct accelerators and Helios rack-scale systems. Its latest data center revenue reached $8.5 billion, up 97% year over year, although that business also includes server CPUs. Broadcom competes through custom accelerators and networking, reporting $16.7 billion in AI semiconductor revenue, up 221% year over year. Marvell also designs custom chips for cloud operators.

Google, Amazon, Microsoft and Meta are deploying proprietary accelerators alongside Nvidia hardware to lower costs and reduce supplier dependence. Their chips increasingly target both training and inference. Cerebras offers another alternative through wafer-scale computing, while Nvidia has licensed Groq technology and hired key executives as it strengthens its inference position.

Huawei is the strongest challenger in China, where export restrictions constrain Nvidia. IDC put Huawei’s share at 28% against Nvidia’s 34% in the first half of 2026. Elsewhere, Nvidia’s CUDA ecosystem makes switching costly, while limited manufacturing, memory and packaging capacity complicates rivals’ efforts to expand shipments.