@Etched
Etched is building custom AI inference chips and rack-scale systems for frontier models. They've raised $800M+ (including a recent $300M Series C at $10.3B valuation), have 400+ engineers from NVIDIA/Google/TSMC, completed their A0 silicon tapeout on TSMC N4P, and claim $1B in customer contracts with first racks shipping summer 2026. While technically impressive with innovations like Low Voltage Inference architecture, this is NOT an early-stage project—it's a well-funded, established AI hardware company with massive backing from Sequoia, a16z, and SK Hynix.
AI Analysisneutral
Etched is building custom AI inference chips and rack-scale systems for frontier models.
They've raised $800M+ (including a recent $300M Series C at $10.3B valuation), have 400+ engineers from NVIDIA/Google/TSMC, completed their A0 silicon tapeout on TSMC N4P, and claim $1B in customer contracts with first racks shipping summer 2026.
While technically impressive with innovations like Low Voltage Inference architecture, this is NOT an early-stage project—it's a well-funded, established AI hardware company with massive backing from Sequoia, a16z, and SK Hynix.
Green flags: Real technical differentiation with custom silicon tapeout and novel Low Voltage Inference architecture · Strong team of 400+ engineers from top semiconductor companies · Substantial customer validation with $1B+ in contracts and shipping product · Major institutional backing and strategic partnerships
Red flags: Already established—$10.3B valuation, $800M raised, 400+ person team, partnerships with Sequoia/a16z · Not early-stage by any definition—massive scale operation with 80,000 sqft facility · Account only 2 years old but company is far beyond 'gem' status with institutional recognition · This is a well-known AI infrastructure company in Silicon Valley circles, not an undiscovered project
Recent tweets
Tweets unavailable right now.
Signal Timeline
Score breakdown0–100
Watching for additional signals.