Running a large AI model at scale has become less a scientific feat than a brutal math problem about electricity bills and datacenter rent. Every token Gemini generates costs Google something real—silicon cycles, memory bandwidth, and watts drawn from the wall. As query volume grows, fractions of a cent add up to sums that can swallow margins whole. That quiet urgency is behind Frozen v2, an internal server chip project now taking shape inside Google. Rather than refining its general-purpose Tensor Processing Units for another generation, the company is attempting something far more radical: casting the skeleton of the Gemini model directly into the silicon itself.

From Flexible Accelerators to Model-Specific Silicon

Google’s TPUs have been the workhorses of its infrastructure for nearly a decade. They train models, power search ranking algorithms, and even rent out by the hour to cloud customers including Meta and others looking for an alternative to Nvidia’s GPUs. That versatility is exactly what makes a TPU a TPU. It speaks a general vocabulary of matrix multiplication and memory movement, usable by almost any neural network you can describe in software.

Frozen v2 deliberately trades that flexibility away. The chip is being designed as a domain-specific accelerator whose circuits physically mirror portions of Gemini’s own architecture. Where a TPU fetches instructions and interprets them as software operations, Frozen v2 would burn the model’s structural blueprint—the arrangement of its layers and data paths—directly into the chip’s layout. Google expects this tight marriage of model and metal to make the chip six to ten times more efficient than current TPUs at serving AI responses. Fewer compute steps per query means less time waiting for a token to appear, and far less energy spent generating it.

This is not merely a faster version of the same idea. It is a different category of chip, one that trades generality for devotion to a single model family.

Why the First “Frozen” Melted

This approach has roots in an earlier concept attributed to Jeff Dean, Chief Scientist at Google DeepMind. The original “Frozen” proposal suggested pushing specialization even further by hardcoding not just the architecture, but the actual model weights—the billions of tuned parameters that constitute Gemini’s learned behavior—directly into the chip itself.

The logic was sound. If you know exactly which numbers the model will use, why bother fetching them from external memory? You could etch them into the transistors and eliminate entire categories of delay.

The problem was permanence. AI models do not stand still. Google updates Gemini continuously, retraining on new data, adjusting parameters, releasing improved versions. A chip with weights frozen in silicon would become a paperweight the moment a new model revision shipped. That lack of flexibility killed the original concept.

Architecture Without the Anchor

Frozen v2 solves the obsolescence trap by hardcoding the architecture while leaving the weights free to change. Think of it as pouring a custom racetrack instead of welding the car to the road. The shape of the circuit stays fixed, optimized for Gemini’s specific patterns of computation, but the contents flowing through those circuits can be refreshed by loading new weights from memory.

This distinction matters in practice. When engineers train a new Gemini checkpoint, they can deploy it to Frozen v2 hardware without fabricating a new chip. The exact degree of hardcoding is still an open question inside Google; teams must decide precisely which structural elements deserve silicon immortality and which should remain configurable. But the principle is settled. By freezing the shape and fluidly swapping the parameters, Google keeps the efficiency upside without sacrificing the ability to iterate.

The Economics of Keeping It In-House

There is another reason you will not see Frozen v2 listed on Google Cloud’s pricing sheet. Because the chip is so intimately shaped around Gemini’s internals, it would make little sense to outside developers running PyTorch or custom Transformer variants. Google has no plans to sell it as a general-purpose product. It will remain an internal tool, aimed squarely at the crushing demand for inference capacity inside Google’s own datacenters.

Chaguo hilo linaakisi ukweli mgumu wa kiuchumi. Katika soko la sasa la AI inayozalisha, uwezo wa mifumo unakaribia kufanana kwa haraka. Pengo kati ya washindani mara nyingi linategemea nani anaweza kumudu kuendesha mfumo mkubwa zaidi kwa gharama ndogo zaidi kwa kila token. Utoaji wa matokeo (inference) si jambo la ziada baada ya mafunzo; kwa bidhaa inayotumiwa sana kama Gemini, ndiyo gharama kuu. Ikiwa Frozen v2 itapunguza gharama hiyo kwa mara sita au zaidi, Google itapata nafasi ya ziada ambayo washindani hawawezi kuipata kwa urahisi. Inaweza kuweka akiba hiyo kama faida au kuipitisha kama bei nafuu kwa watumiaji wa API na mifumo iliyounganishwa, hivyo kuongeza shinikizo kwa OpenAI, Anthropic, na wengine.

Hii Inamaanisha Nini kwa Sekta

Hatua ya Google pia inaashiria mwelekeo wa mkakati mpana wa vifaa (hardware). Kwa miaka mingi, mkakati wa kawaida ulikuwa ni kutengeneza kifaa cha kuharakisha (accelerator) chenye uwezo mkubwa wa kubadilika na kuacha programu (software) ishughulikie utaalamu maalum. GPU za Nvidia zinatawala kwa sababu zinaweza kuendesha kila kitu kuanzia mienendo ya molekuli hadi michezo ya video na mifumo mikubwa ya lugha (large language models). TPUs za Google zilionza kwa ari hiyo hiyo ya matumizi mapana.

Frozen v2 inavunja utamaduni huo. Ni kukiri kwamba wakati familia moja ya mifumo inapoendesha idadi kubwa ya maswali, silikoni maalum iliyoundwa kwa ajili ya mfumo huo inaweza kujilipa mara nyingi zaidi. Washindani wengine wakubwa (hyperscalers) wamefuata mantiki kama hiyo—kwa mfano, chip za Trainium na Inferentia za Amazon—lakini mbinu ya Google inaenda mbali zaidi kwa kusanifu vifaa (hardware) kulingana na usanifu maalum wa mfumo badala ya aina ya jumla ya mitandao.

Hatari, bila shaka, ni kukosa uwezo wa kubadilika (rigidity). Ikiwa usanifu wa Gemini utabadilika kuelekea upande ambao saketi zilizowekwa (hardcoded circuits) haziwezi kukidhi, Google inaweza kujikuta na silikoni ghali ambayo haiwezi kuendesha mawazo yake mapya zaidi. Hiyo ndiyo sababu hasa kwa nini suluhu ya usanifu pekee (architecture-only compromise) ni muhimu. Inatoa njia ya kati: utaalamu wa kutosha ili kupata ongezeko kubwa la ufanisi, na uwezo wa kutosha wa kubadilika ili kuzuia kampuni kujifunga katika hali ngumu.

Hitimisho la Kweli

Frozen v2 inapaswa kueleweka si kama tangazo la chip, bali kama dau la kimkakati kuhusu sura ya baadaye ya ushindani wa AI. Google inabeteka kwamba washindi hawatajenga tu mifumo bora zaidi, bali watamiliki mfumo mzima (entire stack)—kuanzia ramani ya mfumo hadi elektroni zinazopita kwenye transista. Ikiwa mradi utafanikiwa, matokeo hayatakuja kuonekana kwenye alama za vipimo (benchmark scores). Yataonekana kwenye safu ya gharama ya ripoti ya mapato ya robo mwaka, ambapo senti chache zilizookolewa kwa kila milioni ya token zinaweza kuchora upya mipaka ya kile kinachowezekana kibiashara katika AI inayozalisha.