Etched Raises $800 Million to Build Low-Latency AI Inference Clusters, Secures Over $1 Billion in Orders

CoinFeed July 1 news, AI hardware company Etched announced it is building 'frontier inference clusters', integrating chip, rack, software, and manufacturing design to improve large model inference throughput, latency, and energy efficiency. Etched stated its first A0 chip has been taped out and back on TSMC's N4P process, and is currently verifying first rack-level products with customers to fulfill over $1 billion in inference cluster orders.