Google is advancing its chip manufacturing capabilities to decrease reliance on external vendors for its cloud services. On Monday, the company introduced two new variants of its Tensor Processing Units (TPUs), which are custom-built for AI tasks. These chips expand Google's offerings to new scenarios, specifically addressing confidential inference and the increased memory needs required for extensive workloads.
The TPU 8tx and TPU 8i Max represent the company's eighth generation of chips, which debuted in April. They enhance the previous generation's capabilities by increasing inference capacity and supporting large-scale training requirements.
The TPU 8tx, developed in collaboration with AMD, is specifically tailored for secure training and execution of AI models. According to a blog post, the chip integrates security directly into the hardware layer. This ensures that sensitive data and model weights remain protected and isolated even while the models are actively in use.
The chip features a custom inferencing network that allows it to perform reasoning tasks simultaneously with training. This function enables the distillation and training of smaller models in real-time. Google claims this is particularly useful for organizations that wish to train models on proprietary data without moving that data between different devices.
Conversely, the TPU 8i Max integrates two of the TPU 8i inference chips launched earlier this year. By combining them, the chip doubles the memory capacity, processing power, and bandwidth within a single package. This allows AI models to process a significantly larger volume of information during queries.
This release occurs as major cloud providers like Microsoft and Amazon, as well as AI research labs such as OpenAI and Anthropic, are developing their own semiconductor lines and associated software. These efforts aim to lessen their dependency on Nvidia's current market dominance.
The timing of this announcement highlights a challenging period for Nvidia, following CEO Jensen Huang's admission over the weekend regarding an alleged affair with a Chinese spy.
The Editorial Staff at AIChief is a team of professional content writers with extensive experience in AI and marketing. Founded in 2025, AIChief has quickly grown into the largest free AI resource hub in the industry.