BREAKING
OpenAI Unveils Jalapeño AI Chip
Who Built What
1
OpenAI design
↓
2
Broadcom silicon
↓
3
Celestica integration
0
mo
design to tape-out
0
GW
deployment scale
Custom ASIC vs General GPU
Jalapeño ASIC
inference-only
●
Built for LLM inference
●
Less data movement
●
Higher perf per watt
General GPU
NVIDIA
●
General purpose
●
Industry standard
●
Higher dependence
0
B
financing at risk
0
GW
first phase
First Chips Due Late 2026–2027
AI NEWS BLITZ
OpenAI has revealed Jalapeño, its first in-house AI chip, built with Broadcom.