Alphabet Plans New "Frozen" AI Chip Integrating Gemini Model for 6-10x Efficiency

Alphabet is developing a new server chip, dubbed "frozen," that directly integrates the blueprint of its Gemini AI model. Developers estimate this chip will be 6 to 10 times more efficient than Google's current line of AI chips, enabling more efficient AI model delivery to users.

View Tweet→

Sources (10)

Google Develops 'Frozen v2' AI Chip to Etch Gemini Architecture into Silicon

PRIMARYTom's Hardware · RSS · 11h ago

Read article→

Google Develops "Frozen v2" AI Chip for 6-10x More Efficient Gemini Model Inference

Tessara Analysts Feed · RSS · 1d ago

Read article→

Related Signals

2h agoLaunchAMD Unveils Helios to Challenge Nvidia in AI Accelerator Market→3h agoSignalIntel EMIB-T Packaging Addresses TSMC CoWoS Capacity Shortfall→5h agoSignalTSMC Former Deputy Manager Indicted for Leaking Military-Exclusive Process to China→6h agoModel ReleaseGoogle Begins Training Gemini 4, Its 'Most Ambitious Pre-Training Run Yet'→6h agoModel ReleaseGoogle DeepMind Launches Gemini 3.6 Flash with Improved Token Efficiency and New Agentic Models→

Credibility

Lead sourceTom's Hardware

Corroboration3 distinct outlets

VerificationCorroborated

Entities

Alphabetsubject TSMCmentioned Groqmentioned Intelmentioned NVIDIAmentioned

Unlock the full signal with Pro

Tessara synthesizes 80+ data sources into actionable intelligence for AI infrastructure investors.

Unlock with Pro→Sign in

← SignalsSignal

Watch7.5LaunchComputeCorroboratedactiveZH Press

Updated 1d ago · 10 sources

Alphabet Plans New "Frozen" AI Chip Integrating Gemini Model for 6-10x Efficiency

Alphabet is developing a new server chip, dubbed "frozen," that directly integrates the blueprint of its Gemini AI model. Developers estimate this chip will be 6 to 10 times more efficient than Google's current line of AI chips, enabling more efficient AI model delivery to users.

View Tweet→