Alphabet Inc. shares rose Monday following reports that the company is developing a new server-grade AI chip called “Frozen v2” [1].
This development is significant because it represents a strategic move to reduce dependence on external hardware providers and lower the massive operational costs associated with generative AI. By optimizing hardware specifically for its own software, Alphabet aims to gain a competitive edge in processing speed and energy efficiency.
The new chip is designed to embed parts of the Gemini model architecture directly into the hardware [2]. This integration is intended to allow Gemini models to run more efficiently than they do on general-purpose hardware [3].
Market reaction was immediate. Alphabet shares on the NASDAQ exchange rose 1.15% on the day the report surfaced [4]. The news originated from a report by The Information and was subsequently picked up by several financial news outlets [1].
Industry analysts said the push for the Frozen v2 chip is driven by a need to curb AI infrastructure spending [5]. By designing its own silicon, Alphabet can better control the cost of scaling its AI services across global data centers [6].
Furthermore, the project is seen as a way to reduce the company's reliance on Nvidia GPUs [5]. While Google has previously developed its own AI hardware, the Frozen v2 project focuses specifically on the architectural needs of the Gemini series of models to maximize throughput, and reduce latency [3].
“Alphabet is reported to be developing a new server‑grade AI chip called “Frozen v2””
Alphabet's move toward vertical integration—controlling both the AI model and the silicon it runs on—mirrors strategies used by other tech giants to escape the 'Nvidia tax.' If Frozen v2 successfully reduces infrastructure costs, it could allow Google to offer more competitive pricing for Gemini API access while improving the profit margins of its cloud business.
