Etched, a US startup focused on AI inference hardware, has secured a major $700 million investment round.
Jane Street, a leading quantitative trading firm, led the round and valued the company at $21 billion.
Why Specialized AI Chips Matter Now
This funding comes as demand grows for specialized hardware designed to run AI models during inference, the process of using trained models to generate responses.
Etched delivers its technology as full systems called “frontier inference clusters,” designed specifically for inference. The company has developed new components for the prefill and decode stages of inference, aiming to make AI processing faster and more cost-efficient.
Jane Street has become Etched’s first customer, testing and purchasing its hardware and deploying an initial rack in its data center.
The deal highlights growing interest in specialized AI infrastructure as companies seek faster and more efficient ways to run demanding AI workloads.
Impact on AI Competition and Daily Life
Etched’s approach could help challenge Nvidia’s dominance in AI hardware by offering specialized systems for inference workloads.
With more than 400 employees and a working chip, the company aims to expand its frontier inference systems as demand for AI compute grows.
Such advancements could support faster and more cost-efficient AI applications across industries as inference becomes an increasingly important part of AI infrastructure.
For everyday users, more efficient inference hardware could eventually support faster AI-powered services, including real-time applications and other systems that rely on rapid model responses.
Looking ahead, Etched’s progress signals growing competition in specialized AI hardware and a broader push to develop more efficient infrastructure for running increasingly demanding AI models.
Maria Lourdes
