AI chip startup Etched has reached a valuation of $21 billion, more than doubling its value in less than a month as investors continue to put large amounts of money into companies building hardware for artificial intelligence.
The San Jose, California-based company raised $700 million in fresh funding. Trading firm Jane Street led the investment round, while other investors included Kleiner Perkins, Sequoia, Andreessen Horowitz and Tiger Global. Etched was valued at about $10.3 billion in a Series C funding round in July 2026, making its jump to $21 billion especially notable.
Etched is developing specialized computing systems designed mainly for AI inference. Inference is the stage where an already-trained AI model processes a request and generates an answer, image, code or other output. As companies deploy more AI applications to millions of users, the cost and speed of inference are becoming increasingly important.
This is where Etched hopes to compete with Nvidia. Nvidia remains a major force in AI computing, but Etched is taking a more specialized approach. Instead of building general-purpose processors for many different workloads, the startup is designing chips, racks and supporting technology specifically for running advanced AI models with higher speed and better power efficiency.
Etched already has a working AI Chip and employs more than 400 people. The company says its first rack-scale systems are being tested and deployed with customers. Jane Street is also Etched’s first customer and has started deploying the company’s hardware in its own workloads.
Demand appears to be growing before Etched has reached large-scale commercial production. The startup says it has secured more than $1 billion in customer contracts, including demand from AI companies and cloud providers. It has now raised about $1.9 billion in total funding.
The rapid increase in Etched’s valuation also shows how strongly investors are betting on alternatives to traditional GPUs. AI companies are looking for ways to reduce the huge computing and electricity costs involved in operating increasingly powerful models. Specialized inference hardware could become an important part of that shift if it can deliver better performance at a lower cost.
For Etched, however, a $21 billion valuation also brings much higher expectations. The company will need to prove that its technology can perform reliably at large scale, attract more major customers and compete in a semiconductor market where Nvidia has a powerful software and hardware ecosystem.
Etched’s latest funding round makes it one of the most closely watched private AI hardware companies. If demand for AI inference continues to rise, specialized chip makers could gain a bigger role in data centers and create stronger competition in a market that has so far been heavily dominated by Nvidia.
