Nvidia rival Etched raises $800 million from Jane Street and a TSMC-partnered fund

Etched, an AI chip startup billed as an Nvidia competitor, has raised $800 million. Its investors include trading firm Jane Street and VentureTech Alliance, a venture fund with a strategic partnership with TSMC. Etched designs chips specifically for inference, meaning running models that have already been trained, and says it has signed contracts worth $1 billion.

AI-processed from TNW; edited by Hamidun News
Nvidia rival Etched raises $800 million from Jane Street and a TSMC-partnered fund
Source: TNW. Collage: Hamidun News.
Listen to article
Article Fact Sheet & Key Figures
ParameterDetails
Primary sourceTNW ↗
Editor-in-Chief
Round size$800 million
Key metric #2Investors include trading firm Jane Street
Another investorVentureTech Alliance, a fund with a strategic partnership with TSMC

Etched, an AI chip startup billed as an Nvidia competitor, has raised $800 million. Its investors include trading firm Jane Street and VentureTech Alliance, a venture fund with a strategic partnership with Taiwanese chipmaker TSMC.

Who invested in Etched

The round stands out for its investor lineup. Alongside dedicated venture funds, trading firm Jane Street joined the deal, as did VentureTech Alliance, a fund with a strategic partnership with TSMC.

  • Round size: $800 million
  • Investors include trading firm Jane Street
  • Another investor: VentureTech Alliance, a fund with a strategic partnership with TSMC
  • Etched says it has signed contracts to sell $1 billion worth of chips

How Etched's chips differ from Nvidia's

Etched doesn't build chips for training AI models. Instead, it designs them specifically for inference: running models and serving responses to end users. That sets the company apart from Nvidia, whose GPUs handle both training and inference, and puts Etched among a growing number of startups betting on specialized hardware built for deploying models in production.

Why the inference chip market is growing

Until recently, the main battleground for AI chips was training large models, where Nvidia holds overwhelming dominance. But companies are shifting from one-off model training to running models continuously: serving millions of user queries, powering agentic workflows, and handling a growing volume of so-called test-time compute, where a model "thinks" longer before answering. As a result, the cost and efficiency of inference have become a separate line item, comparable in importance to training. Against this backdrop, a whole cohort of startups focused specifically on inference chips has emerged over the past couple of years. Judging by the size of its new round and its signed contracts, Etched ranks among the best-capitalized players in this wave.

What it means

The $800 million round and $1 billion in contracts show that the market believes in a distinct niche for AI inference chips, running in parallel with the training chip market, where Nvidia still dominates. The participation of Jane Street and a TSMC-linked fund also signals that players from outside traditional venture capital are entering the inference chip race.

FAQ

How much did Etched raise, and who are the investors?

Etched raised $800 million. Investors include trading firm Jane Street and venture fund VentureTech Alliance, which has a strategic partnership with TSMC.

How do Etched's chips differ from Nvidia's?

Etched designs chips specifically for running (inference on) already-trained AI models, not for training them. Nvidia's GPUs, by contrast, are used at both stages.

How much are Etched's signed contracts worth?

The company says it has signed contracts to sell $1 billion worth of its chips.

Story

Key Questions Answered

How much did Etched raise, and who are the investors?

Etched raised $800 million. Investors include trading firm Jane Street and venture fund VentureTech Alliance, which has a strategic partnership with TSMC.

How do Etched's chips differ from Nvidia's?

Etched designs chips specifically for running (inference on) already-trained AI models, not for training them. Nvidia's GPUs, by contrast, are used at both stages.

How much are Etched's signed contracts worth?

The company says it has signed contracts to sell $1 billion worth of its chips.

ZK
Zhemal Khamidun — CPO at Alpina Digital, product manager of AlpinaGPT. Building AI products since 2019. Hamidun News is a pet project that grew into a production aggregator with 6,000+ readers.
✦ Unified AI Workspace · Nomad

Test Frontier LLMs & Autonomous Agents in One Workspace

Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.

Try Nomad AI →Launch Nomad Agent
What do you think?
Loading comments…