AI Startup Etched Surges to $10.3B Valuation with Innovative Chip Technology

TL;DR

• Etched, the AI chip startup founded by three Harvard dropouts, has raised a $300 million Series C at a $10.3 billion valuation, according to reports published today.
• The company’s pitch centers on inference-only AI hardware: a custom prefill chip and a proprietary memory/interconnect system designed to speed up model execution without relying on general-purpose GPUs.
• The round underscores both strong investor enthusiasm and ongoing skepticism, as Etched says it has already manufactured chips, begun customer testing, and booked $1 billion in orders.

Etched’s valuation surges again

Etched, the AI chip startup founded in 2022 by three Harvard dropouts, has climbed to a $10.3 billion valuation after closing a $300 million Series C led by Sequoia, with participation from Andreessen Horowitz, SK Hynix, Jane Street, Diffusion Capital, and other backers. The company says this is the highest valuation ever for a Sequoia-led Series C.

The new round comes about seven months after Etched was valued at $5 billion in a prior financing, meaning the company has roughly doubled its valuation in a short span.

What Etched is building

Etched’s core bet is that AI infrastructure does not need one chip that does everything. Instead, the company is focused on inference, the stage where an AI model processes a user prompt and generates an answer.

To support that goal, Etched says it designed two major components from scratch:

• A prefill chip that it says runs dramatically faster by operating at much lower voltage than other AI chips.
• A new memory and interconnect system called cluster-scale memory, which lets many chips share memory across a cluster with very low latency.

This approach is intended to make inference cheaper and faster than using general-purpose GPU systems.

A direct challenge to Nvidia’s model

Etched’s pitch is unusually narrow by semiconductor standards: Nvidia builds chips that do everything, while Etched claims to build chips optimized for one job, inference, and to do it better.

That specialization is central to the company’s strategy. Rather than competing head-on across the full AI chip stack, Etched is betting that the explosive growth in AI inference workloads will create room for purpose-built hardware that can outperform flexible but less specialized GPUs.

Signs of momentum

The company says it has already manufactured its homegrown chips, is testing its first full systems with clients, and has booked $1 billion in orders. Etched also says its system is compatible with models including DeepSeek and Qwen, suggesting it is positioning itself for broad model support rather than a single proprietary ecosystem.

Those claims matter because they suggest the company is moving beyond concept-stage fundraising and into early commercialization, even though it remains in a highly competitive and technically demanding market.

Skepticism remains

Despite the fast-rising valuation, Etched has also faced skepticism from industry observers about whether a highly specialized inference chip can scale into a durable business and whether the company can execute against entrenched incumbents.

That skepticism is typical in AI hardware, where ambitious chip startups often promise major performance gains before proving they can deliver volume production, customer adoption, and consistent economics. In Etched’s case, the stakes are even higher because the company is trying to redefine part of the AI compute stack rather than simply improve it.

Why investors are still betting big

The investment frenzy around Etched reflects a broader wave of capital chasing AI infrastructure. One report said the company was also in talks for a $20 billion valuation in a separate financing structure, with another round led by Sequoia reportedly valuing it around $10 billion, although those talks were not finalized.

Even without those additional rounds, the current financing shows strong conviction from major investors that AI inference will become one of the most valuable layers in the AI economy. If Etched’s hardware works as promised, the payoff could be substantial because inference demand grows every time an AI model is queried in production.

The bigger implication for AI hardware

Etched’s rise highlights a deeper shift in the AI market: training large models still gets most of the headlines, but the long-term economics may depend just as much on inference efficiency. If purpose-built chips and memory systems can reduce cost and latency at scale, they could reshape how AI services are deployed and priced.

At the same time, the company’s success is far from guaranteed. Etched still has to prove that its chips can perform reliably, ship at scale, and win over customers in a market dominated by Nvidia and other established players. For now, its soaring valuation shows that investors are willing to fund that gamble.

AndroGuider Team

Articles written by the AndroGuider team. We try to make them thorough and informational while being easy to read.