Etched raises $700M at $21B valuation led by Jane Street after AI hardware deployment
Etched’s valuation surged from $10.3B to $21B in a month following Jane Street’s deployment of its AI inference cluster, with investors citing speed and cost advantages in the prefill and decode phases.
1 source · cross-referenced
- Etched raised $700 million in a new funding round, doubling its valuation to $21 billion in a month.
- Jane Street led the round after testing and deploying Etched’s AI inference cluster in its data center.
- Etched’s systems use custom chips and cluster-scale memory to accelerate the prefill and decode phases of inference.
- Investors include Kleiner Perkins, Sequoia Capital, Andreessen Horowitz, and others.
Etched, a startup focused on AI inference hardware, announced a $700 million funding round that doubled its valuation to $21 billion in a month. The round was led by Jane Street, a quantitative trading firm, which had previously tested and deployed Etched’s AI inference cluster in its own data center.
Etched’s systems are designed to accelerate the two critical phases of AI inference: prefill and decode. The prefill phase processes user prompts, while the decode phase generates the output tokens. Etched claims its custom prefill chip operates at low voltage, enabling higher transistor density without overheating, and its cluster-scale memory interconnect allows multiple chips to share a fast, low-latency memory pool.
The company’s valuation jumped from $10.3 billion in July to $21 billion in August, following the deployment by Jane Street. Etched’s original intent was to design chips tailored to specific models, but it now states its systems can run any frontier model.
Investors in the round include Kleiner Perkins, Sequoia Capital, Andreessen Horowitz, Peter Thiel, Tiger Global, Bain Capital Ventures, Neo, Stripes, Primary, Positive Sum, Diffusion, Argo, and Blackstone.
- Aug 18, 2026 · Hugging Face
Hugging Face adds multi-vector embedding models to sentence-transformers library
Trust84 - Aug 18, 2026 · Hugging Face
Constraint-aware GPU allocator raises utilization by up to 33 percentage points in enterprise AI workloads
Trust79 - Aug 17, 2026 · TechCrunch — AI
Amazon buys rare books, destroys spines to scan for AI training
Trust72