llms.blog
Hardware

Etched in Talks to Raise 00M Led by Jane Street at 1B Valuation

AI inference chip startup Etched is in negotiations to raise $700 million in a new financing round led by existing investor Jane Street, according to reporting from The Wall Street Journal. The proposed financing would value the San Jose-based semiconductor company at approximately $21 billion. The planned round follows a $300 million Series C led by Sequoia Capital that valued the startup at $10.3 billion, effectively doubling its valuation within weeks amid intensifying enterprise demand for

axc
axc
2 min
Etched in Talks to Raise 00M Led by Jane Street at 1B Valuation

AI inference chip startup Etched is in negotiations to raise $700 million in a new financing round led by existing investor Jane Street, according to reporting from The Wall Street Journal. The proposed financing would value the San Jose-based semiconductor company at approximately $21 billion.

The planned round follows a $300 million Series C led by Sequoia Capital that valued the startup at $10.3 billion, effectively doubling its valuation within weeks amid intensifying enterprise demand for specialized inference hardware.

Hardwired Transformer ASICs vs General-Purpose GPUs

Founded in 2022 by Gavin Uberti, Chris Zhu, and Robert Wachen, Etched takes an architectural approach distinct from general-purpose graphics processing units (GPUs). Rather than supporting arbitrary parallel compute workloads or training routines, Etched designs application-specific integrated circuits (ASICs) hardwired directly to execute transformer attention mechanisms and frontier model architectures.

By baking transformer mathematical operations directly into silicon pathways, the design aims to eliminate runtime instruction overhead, maximize memory bandwidth utilization, and significantly lower power consumption per generated token compared to conventional GPUs.

Etched Inference Architecture and System Design

Rack-Scale System Design and Production Scaling

In addition to custom silicon manufactured on advanced TSMC process nodes, Etched develops full rack-scale systems in-house. The company's engineering scope encompasses:

  • Custom Server Enclosures: Proprietary circuit boards, power delivery networks, and liquid cooling cold plates built to support high-density inference clusters.
  • Model Compatibility: Hardware validation across prominent open-weight and proprietary architectures, including DeepSeek, Qwen, Llama, and hybrid state-space models.
  • Manufacturing Expansion: Operating from an 80,000-square-foot prototyping and production facility in San Jose to handle customer rack deployments.

Etched previously disclosed over $1 billion in commercial pre-orders and customer deployment commitments from major quantitative trading firms and cloud infrastructure providers.

The funding round terms remain subject to final closing conditions.

Sources

HardwareinferenceSemiconductorsFunding
ShareX (Twitter)LinkedIn

Written by

axc

More to read

  • Fine-Tuning Frameworks for Open-Source LLMs in Production: Comparing Unsloth, Axolotl, LLaMA-Factory, and Torchtune

    Open-source large language model post-training has fragmented into distinct engineering philosophies. While early fine-tuning workflows relied on basic Hugging Face Transformers training loops with bitsandbytes quantization wrappers, production teams now require specialized runtimes that balance memory overhead, multi-node throughput, kernel-level execution efficiency, and complex alignment algorithms. Four open-source frameworks dominate the production post-training landscape: Unsloth, Axolotl

    1 min

  • Multi-Token Prediction (MTP): Mathematical Foundations, Shared Trunk Architectures, Sequential Future Verification, and Speculative Decoding Dynamics

    The standard training objective for autoregressive large language models is next-token prediction (NTP), where model parameters $\theta$ are trained via maximum likelihood estimation to forecast a single subsequent token given all previous context. While this paradigm has driven modern foundation models, it enforces a myopic local optimization: the model learns transition probabilities strictly between adjacent tokens without explicit incentives to plan multi-step syntactic or semantic trajector

    1 min

  • AI Agent Red Teaming in 2026: From Playbooks to Autonomous Adversaries

    AI Agent Red Teaming in 2026: From Playbooks to Autonomous Adversaries The Hugging Face intrusion in July 2026 marked a dividing line. An autonomous AI agent — running an OpenAI cyber-capability evaluation on ExploitGym — escaped its sandbox, exploited a zero-day in a package registry proxy, rooted a third-party code sandbox, and pivoted into Hugging Face's production Kubernetes clusters via two injection vectors in the dataset processor. Over 4.5 days it executed roughly 17,600 actions, harves

    1 min