● Covered by 4 sources · 4 reportsMedium impact
Google Develops 'Frozen v2' AI Chip to Enhance Gemini Model Efficiency by 2028
Key points
- Google develops 'Frozen v2' for Gemini models.
- Chip enhances efficiency, aims for 2028 release.
- Projected efficiency: 6-10 times current AI chips.
- Chip embeds Gemini architecture in silicon.
- Responds to AI compute shortage, reduces Nvidia reliance.
Overview
Google is working on a new AI chip called "Frozen v2", aimed at enhancing the efficiency of its Gemini models. The chip will incorporate elements of Gemini's architecture directly into its silicon, potentially increasing efficiency by 6 to 10 times compared to current tensor processing units (TPUs).
Details of Development
Set for release around 2028, 'Frozen v2' is designed to address the AI compute shortage by making Gemini models more power-efficient. This would enable Google Cloud to handle more workloads without needing additional external resources.
The information was first reported by The Information, and though Google has not officially confirmed the specifics, it emphasized its ongoing efforts to optimize AI performance.
Industry Context
The development of 'Frozen v2' signifies a trend in the tech industry towards in-house hardware production. This approach aims to optimize hardware for specific AI models, thereby reducing reliance on third-party suppliers like Nvidia, who dominate the current AI chip market.
Such tailored chips could lead to reduced latency and increased efficiency, influencing Google's AI and cloud operations substantially.
Implications
If successful, 'Frozen v2' could allow Google to offer more efficient and cost-effective cloud services. With increased efficiency, AI applications may expand, potentially unlocking new technological possibilities and market opportunities for the company.
The focus on in-house AI chips reflects broader industry efforts to tackle global AI computing capacity issues.
✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →
The daily brief
One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.
One email a day. Unsubscribe in one click, any time.
Today's brief
Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.
▶ Play today's brief
New every morning, and the back catalogue is archived by date.
How outlets covered it
Google is developing a new chip, dubbed 'Frozen v2', which integrates its Gemini architecture into silicon. This chip aims to significantly enhance token processing efficiency, responding to AI compute shortages and expected to launch by 2028.
Google is developing a chip designed specifically for its Gemini AI model, which aims to improve inference efficiency significantly. This specialized silicon could potentially provide six to ten times more tokens per watt compared to current AI chips, addressing the AI compute demand in cloud services.
Google is developing a new AI chip called 'Frozen v2' to enhance the efficiency of its Gemini models, potentially achieving six to ten times the efficiency of existing chips. This effort reflects the industry's trend towards in-house chip production amid global AI computing shortages and aims to reduce dependence on Nvidia's hardware.
Alphabet's shares rose 3% following a report on its new AI chip, 'Frozen v2', which aims to enhance efficiency for Gemini models. The chip's design could significantly reduce computational demands, addressing internal compute shortages while potentially impacting Google's cloud operations.