AI chip startups are now attacking Nvidia by eliminating DRAM, interconnects and compute-memory splits
deedydas · x · 2026-07-26
AI chip startups are splitting Nvidia’s stack into four new substrate bets
The post updates a prior list of next-gen AI chip startups and adds more coverage of interconnect and alternative substrates. The accompanying chart frames the sector as an attempt to attack data movement — not just compute — in order to break Nvidia’s dominance.
Key approaches highlighted:
- Eliminate DRAM: Groq
- Eliminate the interconnect: Cerebras
- Eliminate the compute/memory split: d-Matrix
- Eliminate server compute-centrism: Majestic
- Eliminate generality: Etched, Taalas, MatX
- Eliminate the $400M lithography machine: substrate-focused players
The image also maps a larger landscape of interconnect, optical compute, thermodynamic, and superconducting startups. It notes that roughly $18B of value now sits in interconnect, far more than the optical-compute, thermodynamic, or superconducting segments individually.
Related event: AI Chip Startups Target Nvidia's Dominance with $58B Valuation(4 posts)→
More from Infra
- 12 KV Cache Reduction Techniques Every AI Engineer Should Understand, Explained — blaizedsouza · 2026-09-11
- The shadow GPU capacity market is formalizing, with Meta selling excess compute to outside buyers — DavidLinthicum · 2026-09-11
- Engram's random reads don't suit SSDs; CPU-memory over NVLink could serve all 72 GPUs — bookwormengr · 2026-09-11
- 80% of the DIY LLM inference hype posters have already quit — it's brutally hard systems work — abhijithneil · 2026-09-11
- Hugging Face's Ultra Scale Playbook: a free book on training LLMs on GPU clusters — mdancho84 · 2026-09-11
- Is inference latency becoming the biggest bottleneck for production AI agents? — Euphoric_Sea632 · 2026-09-11