Etched's AI valuation doubles in a month

Etched, the AI hardware startup that designs systems specifically for running large language models, has closed another major funding round. On Tuesday, the company announced it raised $700 million at a $21 billion valuation, with the round led by Jane Street, the quantitative trading firm known for its rigorous technology evaluations. The news caps a remarkable stretch for the startup, whose valuation has more than doubled in the span of a month.

According to TechCrunch, Etched was valued at $5 billion in December, then raised a $300 million Series C at a $10.3 billion valuation in July. Now, investors have valued the company at $21 billion — an increase of roughly $11 billion in just a few weeks. The dramatic step-up signals intense investor appetite for specialized AI infrastructure, even as the broader tech market remains cautious about high-cost hardware bets.

Etched delivers its AI technology as full systems, which it calls “frontier inference clusters.” The company’s competitor Nvidia refers to its own full systems as “AI factories.”

Why Jane Street led the round

Jane Street, which typically invests in quantitative trading strategies and technology, has a reputation for testing hardware before backing it. In this case, the firm said it did exactly that. In a blog post announcing the new round, 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.”

That hands-on validation appears to have been a key factor in the investment. Etched’s co-founder and COO Robert Wachen told TechCrunch that investor enthusiasm is driven by the company’s two new custom components designed to speed up inference — the computing process that happens after a user submits a prompt.

The two stages of inference

Wachen explained that inference is built in two stages: “prefill and decode.” In the prefill phase, the system must understand the prompt, including its context. This is mathematically and compute-intensive. In the decode phase, the system generates output tokens — the actual answer the user sees. This phase is memory-intensive.

Etched designed 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. The result, according to the company, is that the chip can process more tokens faster. For the decode process, Etched created a new type of memory and an interconnect 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 said. Etched promises that this architecture delivers higher speeds and lower costs for AI inference workloads.

Moving beyond the custom-chip perception

Etched is still contending with the perception from its early days that it “etches” a particular model into its chips — meaning each chip is custom-designed to run one frontier model. That was indeed the company’s original intention, but Wachen says that is no longer the case. Etched’s systems can now run any frontier model, which is a significant shift from its initial focus.

This change is crucial because the AI model landscape evolves rapidly. If a customer invests in hardware tied to one model, they risk being locked out of future, more capable models. By supporting any frontier model, Etched aims to offer flexibility that its early approach lacked.

A growing list of backers

Jane Street is not the only prominent investor backing Etched. The company’s investor roster includes Kleiner Perkins, Sequoia Capital, Andreessen Horowitz, Peter Thiel, Tiger Global, Bain Capital Ventures, Neo, Stripes, Primary, Positive Sum, Diffusion, Argo, and Blackstone. This breadth of support spans venture capital firms, hedge funds, and institutional investors, reflecting broad confidence in the company’s technology and market position.

The latest round brings Etched’s total funding to a level that underscores the capital intensity of the AI hardware sector. Building custom chips and full systems requires significant upfront investment, and Etched has attracted top-tier backers willing to fund that ambition.

Key numbers at a glance

• New funding: $700 million at a $21 billion valuation
• Previous valuation: $10.3 billion in July, $5 billion in December
• Valuation increase: ~$11 billion in one month
• Prior round: $300 million Series C at a $10.3 billion valuation

What this means for AI infrastructure

Etched’s rapid valuation growth suggests that investors are increasingly focused on the inference layer of AI — the part that actually serves models to users. While training gets much of the attention, inference is where the recurring costs and performance bottlenecks lie. Companies that can make inference faster and cheaper could have a significant advantage.

For businesses considering AI adoption, this trend could mean more options for specialized hardware that promises lower costs and higher throughput. But it also highlights the volatility of the AI hardware market, where valuations can swing dramatically based on perceived technical breakthroughs and investor sentiment.

The fact that Jane Street, a firm known for its quantitative rigor, has already installed a rack in its own datacenter is a notable endorsement. It suggests that Etched’s technology is not just theoretical — it’s being used in production by a demanding customer. That could be a positive signal for other potential enterprise buyers who need assurance that the hardware works in real-world environments.

However, the rapid pace of valuation increases also raises questions about sustainability. The AI hardware space is crowded, with Nvidia dominating the market and numerous startups vying for a slice. Etched’s success so far may depend on its ability to deliver on its promises of higher speeds and lower costs at scale. If it can, the current valuation may be justified; if not, the correction could be swift.

For now, Etched appears to have momentum, backed by a strong syndicate of investors and a lead investor that has put its own infrastructure on the line. The next few quarters will be telling as the company ramps up production and faces the realities of the market.