AI Matters
Why it matters
Google designs chip that bakes Gemini architecture directly into silicon
The Decoder Google's "Frozen v2" chip reportedly bakes Gemini's architecture directly into silicon for efficiency gains
Google is reportedly developing a server chip called "Frozen v2" that integrates Gemini's architecture into hardware to drastically reduce AI processing costs.
Why it matters
Hardwiring AI models into chips can make them up to 10 times more efficient, lowering the cost and energy required to run Google's everyday AI services.
Latest brief · Jul 22, 2026