Google Develops 'Frozen V2' Chip to Run Gemini Models 6 to 10 Times More Efficiently Than TPUs
Google is developing a new server chip, informally called Frozen V2, that integrates the blueprint of its Gemini artificial intelligence model directly into silicon. The chip aims to run Gemini models 6 to 10 times more efficiently than the company’s latest tensor processing units, according to a report from The Information.
Google is targeting deployment as early as 2028 to ease a severe AI compute shortage that has already forced its cloud division to turn away external customers. The design represents a new branch of custom silicon rather than a TPU replacement, trading model flexibility for gains in inference speed and power efficiency.
From the sources (20 posts)
@unusual_whalesGoogle, $GOOGL, to launch new Frozen chip in 2028, per Blooomberg.
@cointelegraph🔥 NEW: Google is developing a new “Frozen V2” chip that could run Gemini 6–10x more efficiently than its existing TPUs, per The Information.
@mtsliveSITUATION BREWING: Google is developing a new AI chip, code-named Frozen, projected to be 6-10x more efficient than its TPUs, per The Information. Frozen would have decisions for Gemini models built into the chip. It is a new branch of cus
@kobeissiletterBREAKING: Alphabet, $GOOGL, is planning on launching a new “frozen” chip to run its AI models more efficiently, per The Information. Details include: 1. This new server chip would directly integrate the blueprint of its Gemini AI model 2
@zephyr_z9So, this chip makes a lot of sense for real-time voice models and stuff that requires extremely low latency
@kimmonismusGoogle may be preparing to freeze parts of Gemini’s architecture directly into silicon. Informally called "Frozen v2," the chip reportedly targets 6–10× more tokens per watt than Google’s newest TPUs. Deployment is planned for as early as
@zephyr_z9RT @jukan05: So Google is essentially developing Taalas-like chips that bake the model weights directly into the silicon?
@polymarketmoneyBREAKING: Google developing a new AI chip called “Frozen v2” to run Gemini models up to 10x more efficiently than its latest TPUs.
@jukan05
@jukan05So Google is essentially developing Taalas-like chips that bake the model weights directly into the silicon?
@stocksavvyshay$GOOGL is developing a new AI chip called “Frozen v2” that could run Gemini models nearly 10x more efficiently than its latest TPUs. The chip is targeted for 2028 and would hardwire parts of Gemini to improve speed and efficiency while eas
@techmemeSources: Google is developing a specialized server chip, informally dubbed "Frozen v2", that integrates Gemini AI model blueprints in the silicon, set for 2028 (The Information) (Visit Techmeme dot com for the link and full context!)
@stockmktnewzGOOGLE PLANS NEW ‘FROZEN’ CHIP TO RUN ITS AI MODELS MUCH MORE EFFICIENTLY Google is working on a new server chip that would directly integrate the blueprint of its Gemini AI model, enabling the company to serve its AI models to users much
@negligible_cap*GOOGLE PLANS NEW CHIP TO BOOST AI MODEL EFFICIENCY: INFORMATION "Google is working on a new server chip that would directly integrate the blueprint of its Gemini AI model, enabling the company to serve its AI models to users much more eff
@wallstengine$GOOGL is developing a new AI chip that could run Gemini models 6 to 10 times more efficiently than its latest TPUs. The chip, internally called “Frozen v2,” would bake parts of Gemini’s architecture directly into silicon, reducing data mo
@financialjuiceGoogle Plans New ‘Frozen’ Chip to Run Its AI Models Much More Efficiently - The Information
@financialjuiceGoogle's new chip to run AI models more efficiently - The Information
@deitaone*GOOGLE PLANS NEW CHIP: INFORMATION
@firstsquawkGOOGLE PLANS TO DEPLOY THE FROZEN V2 CHIPS AS SOON AS 2028- THE INFORMATION
@financialjuiceGoogle plans new chip - Information.