TREE NEWS update: Broadcom CEO Hock Tan said six frontier AI customers drive most global AI compute demand. Google received Ironwood TPU v7 and TPU v8i, with multi-year deals worth hundreds of billions annually; OpenAI’s Jalapeno accelerator outperforms Grace Blackwell Ultra at half cost, with 1.3GW planned for FY2027. Entropic and Meta outlined multi-gigawatt deployments through 2028.
Broadcom CEO details custom AI chip deals with Google, OpenAI, Meta
AI take
This signals a structural shift in AI compute procurement: hyperscalers and frontier labs are locking in custom silicon on multi-year, multi-gigawatt timelines, making Broadcom a central arbiter of AI infrastructure capacity. The cost-performance advantage cited for OpenAI's chip challenges Nvidia's dominance, but the real story is the scale of committed power and capital, which could strain supply chains and energy grids. Watch whether these deals consolidate AI demand among a few buyers and how rivals respond on pricing and delivery schedules.
Generated by AI for reference only.
Share on WeChat
Open WeChat → Scan → then tap "…" to send to a chat or Moments.
Tap "…" in the top-right corner to send to a chat or share to Moments.
Related News
2h ago
Chinese Brokers Hold $7.6B in 239 Stocks as Top-10 Shareholders
2h ago
Westlake University Launches World Model Research Center
3h ago
DeepSeek open-sources Personal Agent architecture
3h ago
What Is AI Reward Hacking and Why It Matters
3h ago
2026 World Internet of Things Expo opens in Wuxi
4h ago
Changxin Technology Completes Engineering Sample Validation for 4F²-Based Sixth-Gen DRAM Platform