Etched is reviewing early investment offers that would value the AI-chip startup at $40 billion to $50 billion, TechCrunch reported, citing people familiar with the company. Etched builds complete hardware systems around proprietary chips for customers running AI models, rather than selling only standalone components. One source said prominent investors are proposing $40 billion valuations, while lesser-known backers have floated $50 billion. No deal has been agreed, the amount of any new financing wasn't disclosed and Etched declined to comment.
The product addresses inference: the computing that begins after a trained model receives a prompt. Etched says it designed two components to accelerate that work and claims its processors can generate more tokens, faster and at lower cost than Nvidia's. Speed is particularly valuable to disclosed customer Jane Street, a quantitative-trading firm whose strategies can benefit from small performance advantages.
Jane Street also led Etched's last $700 million round and took delivery of an early system. Etched said in July that it had secured $1 billion in customer orders, including Jane Street's, after TSMC produced its test chip during the summer. The report describes those figures as orders, not recognized revenue.
Building the business requires more than chip design. Etched operates a new 10-megawatt data center in Silicon Valley and a Taiwan facility that coordinates production near TSMC, according to TechCrunch. One person told the publication that another round equal to the prior $700 million financing could provide as much as three and a half years of runway, although that isn't a disclosed fundraising target.
The proposed price is moving faster than any committed financing. Etched announced $300 million at a $10.3 billion valuation in July and another $700 million at $21 billion later in the year. Even the low end of the offers now under review would nearly double that last mark. For now, this is a report about investor interest, not a closed round.
Published versions remain available when an article is updated or corrected.
- Revision 1 · Initial publication
Initial publication