Etched just made $TSM leading-edge capacity even scarcer. $1B in contracts means real wafer demand hitting a queue that is already full. Pricing power follows.
Inference scales with usage, not capex cycles. More AI adoption = more inference chips = more N4P/N3 wafer starts at TSMC. TSMC has been raising advanced node pricing 5-10% annually. That’s about to accelerate. The street is not modeling this. When those margins inflect, $TSM re-rates.
Etched working silicon is on TSMC N4P. First-pass (A0) success, no respin needed. Already running DeepSeek, Qwen, Mamba, and Llama on their inference systems.
Etched secured a trategic investment from TSMC’s affiliated venture arm. This is a capacity commitment signal. TSMC backing a chip startup means they see real wafer volume coming.
We're coming out of stealth.
We've built our first racks after a successful A0 tapeout, $1B+ in customer contracts, and $800m raised.
Early customer tests show us achieving SOTA throughput, latency, and power efficiency on inference workloads.
Our first racks ship this summer.
