Marvell Designed Amazon's Trainium and Microsoft's Maia. Almost Nobody Notices.
Two of the largest cloud providers in the world run custom AI silicon that Marvell helped design. The company's own name rarely appears next to either chip's.
Tagged AI accelerators · show all articles
Two of the largest cloud providers in the world run custom AI silicon that Marvell helped design. The company's own name rarely appears next to either chip's.
A next-generation product usually promises more performance than the last. Huawei's 2026 Ascend roadmap breaks that pattern on paper — and the reason why is a rational, deliberate trade-off, not a stumble.
Every other accelerator in this cohort is designed to run whatever model architecture a customer throws at it. Etched's chip runs transformers and nothing else — an enormous bet that the architecture behind ChatGPT isn't going anywhere.
Every chipmaker in this cohort dices a wafer into individual dies. Cerebras builds one processor out of an entire, uncut wafer — an idea the rest of the industry tried and abandoned decades ago.
Nvidia versus AMD gets the headlines. The rivalry that decides which hyperscalers can build competitive custom silicon at all runs between two quieter companies most readers couldn't name.
Google's TPU, and at least five other hyperscalers' custom accelerators, share a co-designer almost nobody outside the industry can name.
OpenAI's custom-accelerator agreement with Broadcom is larger than any other single deal in this cohort by power commitment. This briefing works through the numbers, the timeline, and what building at this scale actually requires.
Almost every company in this cohort's accelerator and startup tracks is American, Korean, or Chinese. Europe's near-total absence from that list is itself a story — and Axelera AI is one of the few companies trying to close the gap.
Most accelerator deals get announced as a unit count or a dollar figure. AMD and OpenAI described theirs in gigawatts of power consumption — a framing that says as much about the industry's real constraint as it does about the chips themselves.
AMD's absorption of inference-chip startup Untether didn't make headlines the way a multi-gigawatt hyperscaler deal does. It may say more about where AMD thinks the accelerator market is actually headed.