Google Releases Two New AI Chips

2026-04-22 · Source: AutoGPT · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cloud Computing & IT Infrastructure · Depth: Novice, short

Summary

Google released two new custom AI chips, the TPU 8t and TPU 8i, on April 22, 2026, marking the eighth generation of its Tensor Processing Units. The TPU 8t is designed for AI model training, while the TPU 8i handles inference, a specialized approach intended to optimize performance and reduce costs. Google claims these new TPUs offer up to 3x faster AI model training and 80% better performance per dollar compared to the previous generation. Furthermore, over one million TPUs can now operate together in a single cluster, providing immense computing power. Google is also collaborating with Nvidia to enhance networking software, specifically upgrading the Falcon tool, to improve the efficiency of Nvidia-based systems within Google's cloud infrastructure.

Key takeaway

For CTOs and VPs of Engineering evaluating cloud AI infrastructure, Google's new TPU 8t and 8i chips present a compelling option for optimizing AI training and inference workloads. You should assess the performance-per-dollar improvements and the scalability of these specialized TPUs, especially for large-scale model deployments. Consider how this focused chip architecture could reduce your operational costs and accelerate development cycles compared to general-purpose hardware.

Key insights

Google's new TPU 8t and 8i chips specialize in AI training and inference, aiming for efficiency and cost reduction.

Principles

• Specialized hardware optimizes AI workloads.
• Cloud providers are developing custom AI chips.

Method

Google splits AI tasks into dedicated chips: TPU 8t for training and TPU 8i for inference, allowing for fine-tuning and improved performance per dollar.

In practice

• Utilize specialized chips for training/inference.
• Consider cloud provider custom silicon options.

Topics

• TPU 8t
• TPU 8i
• AI Chips
• AI Model Training
• AI Inference

Best for: CTO, VP of Engineering/Data, MLOps Engineer, AI Hardware Engineer, AI Architect, Director of AI/ML

Related on AIssential

• Our eighth generation TPUs: two chips for the agentic era — Google's new TPU 8t and 8i chips specialize in AI training and inference for the agentic era.
• Google Cloud launches two new AI chips to compete with Nvidia — Google Cloud's new specialized TPUs enhance AI training and inference while complementing, not replacing, Nvidia's hardware.
• Google Introduces Specialized Chip for New Wave of AI Computing — Google's new TPU 8t processor targets AI inference, optimizing query operations over model training.
• Google Cloud Debuts New AI Chips — AI compute capacity and proprietary silicon are critical differentiators in the competitive cloud and AI model landscape.
• Google to Release New AI Chips, Challenging Nvidia — Google is specializing its TPU program for AI inference to meet surging demand and maintain its competitive edge.
• Google unveils two new TPUs designed for the "agentic era" — Google's new dual-chip TPU architecture optimizes AI training and inference for the emerging "agentic era."

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AutoGPT.