Topic
AI Accelerator Startups
The merchant challengers to NVIDIA: Groq LPU, Cerebras wafer-scale, Etched transformer-ASIC, and the rest.
- Verified Facts
- 124
- Companies Tracked
- 12
- Business Signals (90d)
- 10
- Avg Fit Score
- —
For chip engineers
Hiring focus (this topic)
- other544
Scarcest skills
Full board →- DFT1 cos.
- STA1 cos.
- CUDA1 cos.
- emulation1 cos.
- analog mixed-signal2 cos.
Companies by sub-segment
Custom ASIC
Inference Startup
Power Semiconductors
Modules
Related topics
Other topics on the same stack layer (L4).
L4
AI ASICs
Hyperscaler silicon: TPU, Trainium, MTIA, Maia — and the merchant gap they create.
L4
EDA & Chip-Design IP
The design-stack chokepoint behind every AI chip: Cadence, Synopsys, Siemens EDA, and Arm IP.
L4
Semiconductor Equipment (WFE)
ASML, AMAT, Lam, KLA, TEL: the wafer-fab equipment behind every AI chip.
L4
Power Semiconductors (SiC & GaN)
The SiC/GaN devices behind 800V datacenter distribution and high-density GPU power.
L4
Foundry Capacity
Who gets the wafers — TSMC, Samsung, Intel Foundry — now decides who ships AI chips.
Explore deeper
Engine-synthesised analysis for this topic, plus the ranked supplier leaderboard for its stack layer.
Tools
Subscribe to weekly updates on AI Accelerator Startups
One email Friday morning. Engineers only.
Subscribe to aiinframap Weekly →