Alphabet (GOOGL) stock rose 1.51% on Monday, closing higher after The Information reported Google is developing a new server chip internally called “Frozen v2.”
The chip is designed to run Gemini AI models more efficiently by permanently embedding parts of the model’s architecture directly into the silicon. That approach reduces the number of calculations and the amount of data movement needed to answer queries.
$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 movement and the number of decisions… pic.twitter.com/O6dhl6twdm
— Wall St Engine (@wallstengine) July 20, 2026
Google engineers project the chip could serve between six and ten times more tokens per unit of power compared to the company’s current TPUs — tensor processing units — according to the report.
Google told CNBC its teams are “constantly researching and experimenting with new innovations” and that it co-designs hardware and software “from the ground up” to optimize for real-world workloads.
The project is aimed at addressing a serious internal compute shortage. That crunch has created tensions inside Google and reportedly led Google Cloud to turn away outside business customers.
Just last month, Google agreed to pay SpaceX nearly $1 billion a month to help bridge the gap and meet enterprise compute commitments.
Frozen v2 would sit alongside Google’s existing TPU lineup rather than replace it — a more specialized chip for a specific job.
The trade-off is flexibility. The chip would only work with future Gemini models if Google keeps the same underlying architecture. Google currently views it partly as a trial run and does not plan to produce it at the same scale as its TPUs.
Deployment is targeted for 2028, though engineers are still finalizing the design.
The chip news comes as Google faces real headwinds in AI. The next Gemini Pro release has been delayed after it fell short of internal goals, according to Bloomberg.
Google has also lost several senior researchers to rivals in recent months.
Chinese AI models are gaining ground with U.S. businesses, now accounting for 45% of U.S. company token use. New releases from Moonshot AI and Alibaba this past weekend are narrowing the capability gap further.
Google DeepMind CEO Demis Hassabis was on Capitol Hill this week pitching lawmakers on a FINRA-style federal watchdog for AI, funded largely by the industry, that would test advanced models for national-security risks before release.
Alphabet stock was up a further 0.97% in after-hours trading Monday.
The post Alphabet (GOOGL) Stock Pops as Google Quietly Builds a Chip to Supercharge Gemini appeared first on CoinCentral.
