Etched doubles valuation to $21 billion in a month — investors bet on custom AI inference chips

AI hardware startup Etched announced on Tuesday that it has raised an additional $700 million at a $21 billion valuation, led by quantitative hedge fund Jane Street, which tested and purchased the company's systems.

The valuation jump is exceptionally fast even by AI industry standards. In December, Etched was valued at $5 billion. In July, it closed a $300 million Series C at a $10.3 billion valuation. Now, just one month later, investors have doubled the figure to $21 billion — an increase of nearly $11 billion.

Etched delivers its technology as complete systems, which it calls "frontier inference clusters." Its competitor, NVIDIA, markets similar offerings as "AI factories."

Co-founder and COO Robert Wachen told TechCrunch that investor enthusiasm stems from two new components designed from scratch to accelerate inference — the computing process that takes place after a user submits a prompt.

Inference is divided into two stages: prefill and decode. The prefill phase is mathematically and computationally intensive, as the system must understand the entire prompt, including context. The decode phase is memory-intensive, generating the output tokens that form the user's answer.

Etched created a prefill chip that operates at low voltage, allowing it to pack more transistors without the typical heat issues of high-end AI chips. This enables faster processing of more tokens. For the decode phase, the company developed a new type of memory and an interconnect technology it calls cluster-scale memory. This system allows many chips to connect and share a common memory pool with very low latency, significantly speeding up the generation of responses.

The new funding round reflects growing investor confidence in specialized AI hardware designed to outperform general-purpose GPUs in specific inference tasks.