Etched's Valuation Doubles to $21B in a Month, Fueled by Jane Street and New Inference Tech

Etched, an AI hardware startup, announced on Tuesday that it has raised an additional $700 million at a $21 billion valuation, in a round led by Jane Street, the renowned quantitative trading firm. Jane Street's investment follows its testing and purchase of Etched's AI hardware. Even by AI standards, this valuation step-up is remarkably rapid. Etched was valued at $5 billion in December, raised a $300 million Series C at a $10.3 billion valuation in July, and has now seen its valuation nearly double to $21 billion in just a month—an increase of almost $11 billion. Etched delivers its AI technology as complete systems, which it terms "frontier inference clusters." (For context, competitor Nvidia refers to its full systems as "AI factories.") Co-founder and COO Robert Wachen told TechCrunch that investor enthusiasm stems from Etched's design of two new components from scratch to accelerate inference—the computing process that occurs after a user submits a prompt. "Inference is built in two stages," Wachen explained, "prefill and decode." In the mathematically and compute-intensive "prefill phase," the system must understand the prompt, including its context. In the memory-intensive "decode phase," the system generates output tokens, i.e., the actual answer the user sees. Etched has created a prefill chip that operates at low voltage, allowing it to pack in more transistors without the typical heat problems of other high-end AI chips. This enables it to process more tokens at faster speeds. For the decode process, Etched developed a new type of memory and interconnect, which the company calls "cluster-scale memory." "It allows many chips to connect together and use a shared memory pool at a very, very fast, low latency," Wachen said. The result, Etched promises, is higher speeds and lower costs. Etched is still overcoming the perception from its early days that it etches a particular model into its chips, implying that each chip is custom-designed to run one frontier model. That was its original intention, but this is no longer the case. Etched's systems can now run any frontier model. In the blog post announcing the new round, investment firm Jane Street stated, "We tested the chip and are pleased with the early results. Etched's unique approach to inference delivers the precision we will need to support our most demanding workloads. We're excited to now have our own rack running in our datacenter." Other investors in Etched include Kleiner Perkins, Sequoia Capital, Andreessen Horowitz, Peter Thiel, Tiger Global, Bain Capital Ventures, Neo, Stripes, Primary, Positive Sum, Diffusion, Argo, and Blackstone.

via TechCrunch AI