Etched announced on Tuesday that it has successfully secured an additional $700 million in funding, elevating its valuation to an impressive $21 billion. This latest round was spearheaded by Jane Street, a renowned quantitative trading firm, which made the investment after rigorously testing and subsequently acquiring the startup’s advanced AI hardware.
The speed of this valuation surge is remarkable, even within the rapidly evolving AI sector. Etched's valuation stood at $5 billion just in December. This follows a $300 million Series C funding round in July, which valued the company at $10.3 billion. Now, in merely one month, investors have effectively doubled its valuation, adding nearly $11 billion to reach $21 billion.
Etched provides its artificial intelligence technology as comprehensive systems, which the company terms “frontier inference clusters.” This nomenclature draws a parallel to competitor Nvidia, which refers to its full AI systems as "AI factories."
According to Robert Wachen, Etched's Co-founder and COO, who spoke with TechCrunch, the exceptional enthusiasm from investors stems from the company's development of two innovative components, engineered from the ground up to accelerate inference — the crucial computing process initiated after a user submits a prompt.
Wachen elaborated that “Inference is built in two stages: prefill and decode.” The "prefill phase" is characterized by its mathematical and compute-intensive nature, requiring the system to comprehensively understand the prompt, including its full context. Following this, the "decode phase" is memory-intensive, during which the system generates output tokens, ultimately forming the user's desired answer.
To optimize these processes, Etched has engineered a prefill chip designed for low-voltage operation. This design allows for a greater density of transistors without encountering the typical heat dissipation issues associated with other high-end AI chips, thereby enabling faster token processing. For the decode process, Etched developed a novel type of memory and an interconnect system, which it 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 stated. The company guarantees that this innovation will lead to both enhanced speeds and reduced operational costs.
Despite its technological advancements, Etched continues to address an early misconception that its chips are "etched" with a specific model, implying each chip is custom-designed for a single frontier model. While this was an initial concept, it is no longer the case; Etched’s current systems are capable of running any frontier model.
In a blog post announcing the latest funding round, lead investor Jane Street affirmed their confidence, stating, “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.”
A distinguished roster of other investors in Etched includes Kleiner Perkins, Sequoia Capital, Andreessen Horowitz, Peter Thiel, Tiger Global, Bain Capital Ventures, Neo, Stripes, Primary, Positive Sum, Diffusion, Argo, and Blackstone.
The Editorial Staff at AIChief is a team of professional content writers with extensive experience in AI and marketing. Founded in 2025, AIChief has quickly grown into the largest free AI resource hub in the industry.