Google is stepping up its efforts in chip manufacturing as part of a broader strategy to lessen its reliance on third-party chips for its cloud computing operations. On Monday, the company introduced two new variations of its tensor processing units (TPUs)—chips developed in-house for AI workloads—expanding its chip lineup into additional use cases, including confidential inference and greater memory capacity for large workloads.

The new units, TPU 8tx and TPU 8i Max, belong to the company's eighth generation of chips that were introduced in April this year, and they build upon the capabilities those chips offered, namely increased inference capacity and support for demanding training requirements.

TPU 8tx, designed in collaboration with AMD, is based on the TPU 8t chip and is tailored specifically for training and running AI models with confidentiality, Google said in a blog post. The chip "builds security directly into the hardware layer, keeping sensitive data and model weights protected and isolated even while in active use," the company wrote. The chip also features a custom inferencing network that enables it to run reasoning alongside training, allowing it to distill and train small models on the fly. According to Google, this capability could be valuable in scenarios where a company wants to train a model on its proprietary data without having to move that data between devices.

EMBED_PLACEHOLDER_0

Meanwhile, TPU 8i Max essentially combines two of the inference-focused TPU 8i chips that the company launched in April, doubling the memory capacity, processing power, and bandwidth—all within a single package. The result is a chip that enables AI models to handle a broader range of information during queries.

The chip launch comes as hyperscalers such as Microsoft and Amazon, along with AI labs like OpenAI and Anthropic, are developing their own lines of AI-specific semiconductors and accompanying software in an effort to reduce their reliance on Nvidia's dominance. The news also arrives at a particularly vulnerable moment for Nvidia, whose CEO Jensen Huang over the weekend admitted to having an affair with someone alleged to be a Chinese spy.