Wire · founder news, decoded · opportunities
Google plans new chip to run Gemini models more efficiently, the Information reports
◆ Published
20 July 2026
◆ Topic
opportunities
◆ Geography
◆ Source
◆ Verified
Fusion42 · 20 July 2026 · Fusion42 review
Google is developing a custom server chip called 'Frozen v2' that embeds Gemini model elements directly into hardware, targeting 2028 deployment with 6-10x better efficiency than current TPUs to address internal AI capacity constraints.
◆ The Wire takeaway
If you're selling inference optimisation, cooling, power delivery, or tensor software to AI labs, Google just signalled it's building these capabilities in-house by 2028—your buyer is becoming your competitor. Start mapping what Google can't bake into silicon and sell into their supply chain before they do.
◆ Related on Wire
- Samsung Electronics will join hands with Google Cloud to introduce the generative artificial intelli..13 July 2026
- Intel and Google deepen AI ties for chip design19 July 2026
- Gemini Enterprise Agent Platform Leads Enterprise AI Governance as OpenAI Starts Billing ...20 July 2026
- Samsung's 2nm Node Lands Tesla's AI5 Chip at Its Texas Taylor Fab, Silencing Yield Doubters13 July 2026
- Meta to put AI chip into production in September as it looks to double computing capacity ...10 July 2026
- Rapidus Pursues 2-nanometer Semiconductor Production in Japan8 July 2026
◆ Topics
custom-silicon · inference-optimization · gemini-hardwiring · capacity-crunch · 2028-deployment