Alphabet’s Google announced two additions to its artificial intelligence chip lineup on Monday, targeting customers with sensitive data or particularly demanding computing workloads. The processors broaden Google’s offering as it competes with Nvidia in the fast-growing AI chip market.
The products, called TPU 8tx and TPU 8i Max, adapt existing tensor processing units to different customer requirements rather than introducing an entirely new processor design.
For applications that must work through substantial volumes of information simultaneously, the 8i Max brings two chips together in one package. That configuration increases both memory capacity and computing resources. Google identified coding assistance, mathematical reasoning and drug discovery as tasks suited to the processor.
The 8tx addresses a different requirement: safeguarding information while computations are underway. Its intended customers include public-sector agencies and businesses operating under industry-specific regulation.
Google combines its own chip technology with AMD processors in the 8tx, using hardware-based encryption to secure data during processing. The company said the design can run as much as twice as fast as its predecessor, with almost no performance penalty from the protection.
Cost is another element of Google’s competitive position. Its processors reportedly carry prices about 30% below Nvidia’s while delivering comparable performance and consuming less electricity for some AI workloads.
Google’s cloud business has become a channel for offering the chips to enterprise AI customers. Historically, the processors have largely supported Google’s own services, including Gemini, Search and YouTube.
Alphabet said TPU sales contributed meaningfully to its September-quarter results. Revenue at Google Cloud reached $30.71 billion, more than twice the year-earlier level.
Customers are expected to gain access to both new variants in 2027.
