BlockBeats News, July 1st, AI inference chip company Etched announced the end of stealth mode and disclosed the completion of its first chip A0 Tapeout, the construction of the first batch of rack systems, with over $1 billion in customer contracts and $800 million in funding secured. The first batch of racks will start shipping this summer.

Etched stated that early customer tests have shown that its inference system has reached the state-of-the-art (SOTA) level in throughput, latency, and energy efficiency. The company currently has over 400 engineers, with team members from NVIDIA, Google TPU, Broadcom, SK Hynix, TSMC, and other enterprises.

In addition, Etched also released two new technologies: Low-Voltage Inference (LVI) technology for high-throughput inference scenarios, and Cluster-Scale Memory (CSM) architecture for low-latency scenarios, aimed at improving the inference performance of ultra-large-scale MoE models, long-context, and AI Agent workloads.

Etched mentioned that the company has built a 2MW data center and established a factory in Taiwan, China, to accelerate the mass production process. In the coming months, more performance tests and product roadmap updates will be announced.

---------------------------------
Click the original text link below to join the BlockBeats Beating · Feishu AI News Channel, monitoring global AI hotspots and news 24/7.