Etched raises $700 million at a $21 billion valuation in a round led by Jane Street

Etched, an American AI inference chip startup, raised $700 million in a Series D funding round led by quantitative trading firm Jane Street, achieving a market valuation of $21 billion for a company that is only four years old.

This valuation establishes Etched as one of the highest-valued semiconductor startups globally outside of Nvidia, reflecting growing institutional confidence in the idea that specialized silicon for running large language models is the future of this sector.

Etched designs chips optimized exclusively for inference workloads in large language models, in contrast to the general-purpose graphics processing units (GPUs) that currently dominate the market. Investors believe this specialization will generate higher profit margins as AI workloads mature in major data centers.

The round comes at a moment when competition for AI chips is accelerating, alongside a growing conviction that the era of overall GPU dominance will make way for an era of application-specific chips. Analysts believe Etched is betting that the inference market size will surpass the training market over the next three years—a bet that Jane Street appears to have bought into.

What do these terms mean?

Inference: The process of running a language model and using it to respond—every time you ask an AI model a question, that is inference. It differs from training, which happens once when the model is built.

Series D Round: A late-stage funding round for startups that have proven their business model and are seeking major expansion—the higher the letter (A, B, C, D...), the larger the company's size and valuation usually are.

Specialized Chips: Electronic circuits designed for a single task rather than general tasks—like a race car engine compared to a family car engine: less flexible, but faster and more efficient at its primary job.