In 2023, Etched nearly ran out of cash. The founders could not get investors interested. Two years later, the startup is worth $5 billion and has $1 billion in contract orders. How did a small company that makes AI chips become a serious rival to Nvidia?
Etched was founded in 2022 by Gavin Uberti and Robert Wachen. Both founders dropped out of Harvard and earned Thiel Fellowships, a program that backs young inventors who skip higher education to start companies.
The round was led by Stripes, with participation from Two Sigma, Jane Street, Hudson River Trading, VentureTech Alliance, and Ribbit Capital.
But as demand for AI inference chips surged, Etched's technology - built specifically to run models after they are trained - began to attract serious attention. The successful production of their chip by TSMC in early 2026 validated the approach and triggered a wave of pre-orders.
Get your free investing masterclass bonus when you join Market Briefs, our free daily newsletter
Etched's angel investors include billionaires Peter Thiel and Stanley Druckenmiller, along with AI leaders like Arthur Mensch, Fei-Fei Li, Geoffrey Hinton, Andrej Karpathy, and Scott Wu.
In total, Etched has raised $800 million since its founding.
The rapid turnaround from near collapse to a $5 billion valuation highlights the intense demand for specialized AI hardware and the high-stakes nature of the chip industry. Etched's focus on inference rather than general-purpose training allowed it to capture a growing market segment as AI deployment scales.
Etched's chips are designed for inference - the process of running a trained AI model to make predictions or generate answers. Inference is a huge bottleneck for AI companies today. It costs a lot of money and uses a lot of power. An Etched spokesperson stated, "Our systems are faster, cheaper, and more energy efficient than competing chips, including Nvidia's."
That kind of advantage is why customers have already placed $1 billion in orders for Etched's full chip systems.
The company announced its progress after TSMC, the world's largest chip manufacturer, successfully produced Etched's chip earlier in 2026. Now Etched is testing its first product with customers. It expects to deliver the pre-ordered systems, which it calls "frontier inference clusters," to those customers soon.
Etched's rapid rise reflects a larger shift in the AI industry. Training large models once dominated hardware spending, but as companies deploy chatbots, image generators, and recommendation engines, inference workloads now account for the majority of compute costs. Nvidia's general-purpose GPUs handle both training and inference, but specialized inference chips can slash energy use and latency.
Etched's bet on a single-purpose architecture - optimized for transformer models - positions it to capture a slice of a market that analysts expect to exceed $100 billion annually by 2030. The company's $800 million in total funding and $1 billion in pre-orders suggest investors and customers alike see that future arriving sooner than many predicted.
Subscribe to Market Briefs, our free daily newsletter, and claim your bonus investing masterclass
