Etched, an AI inference chip startup, raised $700 million at a $21 billion valuation in a round led by quantitative trading firm Jane Street. The company's valuation doubled in a month after closing a $300 million Series C at $10.3 billion in July 2026. Additional investors include Kleiner Perkins, Sequoia, Andreessen Horowitz, Tiger Global, Bain Capital Ventures, Peter Thiel, and Blackstone.
Jane Street moved from testing to customer to lead investor, having shipped its first rack to the trading firm last month and actively deploying Etched's hardware into its workloads. Jane Street stated: 'We tested the chip and are pleased with the early results. Etched's unique approach to inference delivers the precision we will need to support our most demanding workloads.' The firm now has its own rack running in its data center.
Etched's inference clusters use two core technologies: Low Voltage Inference (LVI) delivers higher compute density within the same power envelope as existing hardware, and Cluster Scale Memory creates a shared memory pool across an entire cluster rather than a single chip. The company has booked over $1 billion in customer contracts across public and private AI companies and cloud providers, with systems already running large mixture-of-experts and non-transformer models.
For architects, the signal matters: a customer with latency-critical trading workloads became convinced enough to lead the funding round, suggesting the hardware can survive real production traffic. Etched represents a rare successful hardware startup path—from demo to paying customer to co-investor—proving demand in the widening inference gap between NVIDIA capacity and what enterprises need.