We rank 36 companies based in the United States in Chips & accelerators, part of our coverage of AI infrastructure. The first 5 by rank are shown here.
| Rank | Company | What it does | Place on our list |
|---|---|---|---|
| 1 | NVIDIA | NVIDIA designs and sells GPUs, networking and integrated AI computing systems to the largest cloud companies, enterprises and system builders | No. 4 of the AI Infra 100 |
| 2 | Broadcom | Broadcom designs custom AI accelerator chips and Ethernet networking chips for the largest cloud companies and AI labs | No. 9 of the AI Infra 100 |
| 3 | AMD | AMD sells Instinct AI accelerator chips and Helios rack-scale systems to the largest cloud companies, AI labs and enterprises | No. 22 of the AI Infra 100 |
| 4 | Marvell | Marvell designs custom AI chips, optical connection chips and memory-expansion chips for AI data centers | No. 46 of the AI Infra 100 |
| 5 | Cerebras | Cerebras designs and sells AI computing systems built on wafer-scale chips to enterprises, cloud providers, research institutions and governments for training and running AI models | No. 87 of the AI Infra 100 |
31 more companies in the United States are ranked in this field. Members see all of them, with scores and the facts behind each place.
Ranked by our Overall score, November 2026 (updated 17 Nov) edition. How the scores are built
By our ranking: NVIDIA, Broadcom, AMD, Marvell, Cerebras.
36, out of 100 companies ranked in Chips & accelerators worldwide.
Every ranked company, its scores, its signals and its peers. $99 a month, 7-day trial. What Frontier Tech Intelligence is →