Etched on Tuesday announced a $700 million raise at a $21 billion valuation, led by Jane Street afte...
The AMW Read
Etched's valuation doubling to $21B with a $700M round is a significant capital event in AI infrastructure, and the custom silicon design updates the hardware landscape, warranting high significance.
Etched on Tuesday announced a $700 million raise at a $21 billion valuation, led by Jane Street after the quant firm tested and bought the startup's AI hardware. This comes just a month after a $300 million Series C at $10.3 billion, and follows a December round at $5 billion. Etched sells full systems it calls 'frontier inference clusters,' competing with Nvidia's AI factories.
The rapid valuation jump reflects investor enthusiasm for Etched's custom inference chips, which split processing into prefill and decode stages. The company designed a low-voltage prefill chip and a new cluster-scale memory interconnect for decode, aiming to cut costs and boost speed. Jane Street noted it tested the chip and is now running an Etched rack in its datacenter, signaling early customer validation beyond the startup's own claims.
For the AI market, this signals that specialized inference silicon is gaining traction against general-purpose GPUs, especially as inference becomes a larger share of AI compute. Builders should watch whether Etched's approach can deliver on its cost and performance promises at scale. Investors may see this as a sign that dedicated hardware startups can command premium valuations amid the broader compute buildout, though Etched must still prove it can move beyond early adopters.
