Etched
Category: AI Infrastructure
AI hardware company building frontier inference clusters—co-designed chips, racks, and software—for high-throughput, low-latency, power-efficient serving of frontier AI models. Etched was founded in 2022. The company is led by Gavin Uberti. Based in San Jose, USA. Team size: 101-500. Total funding raised: ~$1.9B total (all rounds). Latest round: Series D. Key investors include Jane Street, Sequoia Capital, Andreessen Horowitz, Kleiner Perkins, Tiger Global, Bain Capital Ventures, Stripes, Primary Venture Partners, Positive Sum, Blackstone, SK Hynix, VentureTech Alliance.
AMW Analysis
Etched develops a Transformer-specific ASIC for AI inference, hard-coding the architecture into silicon to achieve higher throughput and lower latency than general-purpose GPUs. The company’s recent news flow shows rapid scaling of capital and commercial traction. In early 2026, it raised a $500 million Series A at a $5 billion valuation, followed by a cumulative $800 million in funding that included backing from Jane Street and a TSMC-linked venture firm, alongside a $1 billion chip order for its Sohu ASIC. TSMC subsequently manufactured the chip, and Etched booked $1 billion in contract orders for its inference systems. By July 2026, a $300 million Series C led by Sequoia at a $10.3 billion valuation was reported, with participation from a16z and SK Hynix, and later reports indicated the company was raising capital at a valuation near $20 billion. The funding rounds signal investor demand for custom silicon alternatives to GPUs, positioning Etched within the broader shift toward architecture-specific hardware for AI workloads.
AMW analysis, generated from 8 tracked news signals.
- Founded
- 2022
- Headquarters
- San Jose, USA
- Team size
- 101-500
- Total funding
- ~$1.9B total (all rounds)
Value proposition
Delivers up to 20x higher throughput and significantly lower latency compared to general-purpose GPUs (like Nvidia's H100/B200) by hard-coding the Transformer architecture directly into the silicon.
Products and solutions
Frontier inference clusters (rack-scale systems co-designed chips, racks, software), Sohu / inference ASIC with Low Voltage Inference (LVI) for high-throughput prefill, Cluster Scale Memory (CSM) interconnect for low-latency decode, Software stack and manufacturing methods for frontier-model inference
Unique value
Unlike general-purpose GPUs that are designed to handle various mathematical tasks, Etched 'burns' the Transformer architecture into the hardware logic, eliminating the overhead of programmability to maximize compute density for LLMs.
Target customer
AI labs (OpenAI, Anthropic, Google), hyperscale cloud providers, and enterprises deploying massive-scale Transformer models.
Industries served
Artificial Intelligence, Semiconductors, Cloud Infrastructure, Enterprise Software
Technology advantage
By focusing exclusively on Transformers, the Sohu chip can dedicate nearly all its transistors to the specific matrix multiplications and attention mechanisms used in modern AI, achieving performance-per-watt and cost-efficiencies that general-purpose chips cannot match.
How they differentiate
Etched develops the world's first Transformer-specific ASIC (Sohu), which 'burns' the Transformer architecture directly into the silicon. Unlike Nvidia's general-purpose GPUs or Groq's LPU, Etched's hardware is non-programmable for other architectures, allowing it to achieve 20x higher throughput for LLMs by dedicating nearly all transistors to Transformer-specific matrix multiplications.
Main competitors
Nvidia, Groq, Cerebras Systems, SambaNova Systems
Key partnerships
TSMC (N4P manufacturing; A0 silicon), VentureTech Alliance (strategic / TSMC-linked), SK Hynix (Series C investor/partner), Jane Street (lead investor and first deploying customer)
Notable customers
Jane Street (first customer; first rack shipped/deployed), $1B+ signed customer contracts across public/private AI companies and cloud providers
Major milestones
A0 silicon returned from TSMC N4P; working chip demonstrated, Emerged from stealth Jun 30, 2026 with $800M raised and $1B+ customer contracts, Series C $300M at $10.3B valuation led by Sequoia (Jul 2026), Shipped first rack to Jane Street; Series D $700M at $21B valuation led by Jane Street (Aug 2026), Opened Taiwan factory and 80k sq ft / 10MW Milpitas NPI facility
Growth metrics
Scaled to 400+ engineers from NVIDIA, Google TPU, Broadcom, SK Hynix, and TSMC; working A0 silicon; first customer rack shipped to Jane Street; $1B+ contracts; Taiwan factory plus San Jose/Milpitas production and 10MW NPI lab; ~$1.9B raised at $21B valuation.
Market positioning
High-performance AI infrastructure challenger targeting the LLM inference market.
Geographic focus
Global (Headquartered in Cupertino, USA; targeting global hyperscalers and AI labs).
Patents and IP
Proprietary ASIC architecture and specialized matmul kernel designs; specific patent filings are generally kept confidential during the pre-launch phase but involve custom interconnects and memory management for Transformer weights.
About Gavin Uberti
Gavin Uberti is the Co-founder and CEO of Etched. He was previously a compiler engineer at OctoML, where he specialized in high-performance computing and matmul kernels for the Apache TVM project. He is a 2024 Thiel Fellow and a former student at Harvard University, where he studied Mathematics and Computer Science before dropping out to focus on building specialized AI hardware.
Latest news about Etched
- Etched valuation doubles to $21B in a month after Jane Street leads $700M round
- Sequoia Capital Doubles Down on AI, Backs Etched's Specialized AI Chips
- Etched raises $300M Series C at $10.3B valuation for AI inference chips
- Etched raises $300M Series C at $10.3B valuation led by Sequoia for AI inference chips
- Etched's Funding Rounds Signal Shift in AI Investment Toward Custom Silicon
More AI Infrastructure companies
Official website: https://www.etched.com
