Google Is Building a New AI Chip That Could Cut Power Costs Tenfold

Key points

• Google is developing a new server chip, internally codenamed "Frozen v2," with a planned release in 2028.
• The chip could be 6 to 10 times more efficient than Google's existing AI chips, measured by how many text responses it can produce per unit of electricity.
• Google's stock rose roughly 3% on Monday morning after news of the chip became public.
• OpenAI announced its own first custom chip, called Jalapeño, in June 2025; Anthropic is reportedly in talks with Samsung over a chipmaking partnership.

Google is quietly designing a new kind of computer chip built specifically to run its Gemini AI models, the technology behind Google's chatbots and AI search features, more cheaply and quickly. The chip, codenamed "Frozen v2" inside the company, is targeted for release in 2028, first reported by The Information, citing anonymous sources familiar with the project.

The headline figure is striking. Frozen v2 could produce 6 to 10 times as many AI-generated responses for every unit of electricity compared with Google's current chips. In plain terms: the same electric bill could power far more AI work.

Google did not confirm the report when TechCrunch AI asked directly. It also did not deny it. "Our teams are constantly researching and experimenting with new innovations to deliver maximum performance and efficiency for our users and customers," a spokesperson said.

Why does any of this matter to ordinary people? Running AI is expensive, and those costs flow downstream. Companies that can cut their chip bills have more room to keep AI tools free or affordable, rather than passing costs on to subscribers or advertisers.

The wider industry is pushing hard in the same direction. Nvidia, the California chipmaker, currently supplies most of the specialised processors that AI companies need. That dependence worries the big labs: if Nvidia raises prices or supply runs short, everyone from Google to startups feels the squeeze. Building their own chips is the most direct way to reduce that risk.

OpenAI took that step in June 2025, announcing Jalapeño, its first in-house inference chip, a processor designed specifically to generate AI responses rather than to train new models. Anthropic is reportedly in early discussions with Samsung about a similar arrangement.

For Google, the financial pressure is acute. Earlier this year, the company said it plans to spend between $180 billion and $190 billion on AI infrastructure through 2025. That is a number large enough to unsettle investors, who want evidence the spending will produce returns. News of Frozen v2 appears to have offered some reassurance: Google's stock climbed around 3% on Monday morning following The Information's report, just before the company's quarterly earnings call.

What does this mean for people who use Google products?

Faster, cheaper chip technology generally means AI features reach more people at lower cost. If Frozen v2 performs as described, Google could run Gemini more efficiently, which could translate into quicker responses in Search, Google Docs, or Android, without a price increase for users. The 2028 target date is still years away, and not every internal project reaches production, as Google's own statement acknowledged.