Etched, an AI chip startup founded by former Cerebras and Groq engineers, has reached a $5 billion valuation after securing $1 billion in contracted sales for its inference-focused chip systems. The company announced the milestone in a funding round that positions it as one of the few serious challengers to Nvidia's dominance in AI accelerators.

Etched's chip targets a specific market segment: inference workloads, where AI models run predictions on data rather than training on it. This narrow focus differs from Nvidia's broad approach to both training and inference. The company's systems are designed around specialized architecture optimized for transformer inference, reducing latency and power consumption compared to general-purpose GPUs.

The $1 billion in booked contracts represents actual customer commitments, not mere interest. This matters because it demonstrates real market demand from enterprises willing to spend substantial capital on alternative hardware. Customers reportedly include major cloud providers and AI companies seeking to reduce inference costs and dependency on Nvidia.

Etched joins a growing list of companies chipping away at Nvidia's market share. Groq, another inference specialist, has also gained traction with similar positioning. Meanwhile, custom silicon efforts from Amazon (Trainium, Inferentia), Google (TPUs), and others reflect broader industry movement away from relying exclusively on Nvidia GPUs.

The $5 billion valuation occurs in a competitive landscape where chip startups need significant capital to manufacture and scale. Etched's path from founding to $1 billion in bookings happened faster than most hardware companies achieve. Manufacturing partnerships and a clear product roadmap appear central to the company's trajectory.

However, Nvidia retains advantages in software ecosystem maturity, installed base, and total addressable market. Etched's success depends on execution: delivering systems on schedule, proving performance claims in production environments, and maintaining customer satisfaction as