AI chip startup Etched reaches $10.3 billion valuation

Etched, a start‑up founded by three Harvard dropouts, has announced a line of chips and memory components that promise to accelerate inference on any AI model without the need for traditional GPUs. The company claims its architecture, which blends custom silicon with advanced memory‑bandwidth optimizations, can deliver faster inference times while reducing power consumption compared to conventional GPU‑based solutions.

The new hardware targets a broad spectrum of AI workloads, from natural language processing to computer vision, and is designed to be compatible with existing software stacks. By offloading inference tasks to specialized chips, Etched aims to lower the barrier for deploying AI models in edge devices and data centers alike, potentially cutting both latency and operational costs.

Industry analysts note that if Etched’s performance claims hold up in real‑world deployments, the company could disrupt the current GPU‑centric AI infrastructure. The firm is currently seeking partnerships with cloud providers and OEMs to integrate its technology into next‑generation AI platforms.