Etched Raises $700 Million at a $21 Billion Valuation: The AI Chip Race Is Heating Up

The AI investment boom is moving deeper into the technology stack.

Instead of focusing only on AI models and applications, investors are increasingly betting on the hardware needed to run those models.

The latest example is Etched, a San Jose-based AI chip startup that has raised $700 million at a $21 billion valuation. The round was led by Jane Street, with participation from major investors including Sequoia Capital, Andreessen Horowitz and Tiger Global.

The striking part isn't only the amount raised.

Etched more than doubled its valuation in less than a month.

Why AI Chips Matter

Artificial intelligence requires enormous computing power.

Training a large model can require thousands of specialized processors running simultaneously.

But training is only one part of the equation.

Once an AI model is released, millions of users may interact with it.

Every question, image generation or AI-agent task requires inference—the process of running a trained model to generate an output.

As AI adoption grows, inference is becoming an increasingly important part of the computing market.

Etched's Bet on Specialized Chips

Etched is developing specialized chips designed specifically for AI inference.

The company is positioning its technology as an alternative to more general-purpose AI accelerators, particularly in environments where speed and efficiency are critical.

The argument is relatively straightforward.

A specialized chip can potentially be optimized for a specific workload.

That could allow AI systems to generate responses faster while using less energy or reducing the cost of processing each token.

For companies operating AI services at massive scale, even small improvements can translate into significant savings.

Why Jane Street Is Interested

Jane Street is not the typical AI hardware investor.

The trading firm is also reportedly Etched's first customer and has begun deploying its technology in its own data center.

That commercial relationship gives Etched something many early-stage semiconductor companies struggle to obtain: evidence that customers are willing to actually use the hardware.

AI chips are extremely difficult to commercialize.

Building a technically impressive processor is one challenge.

Convincing customers to deploy it inside production infrastructure is another.

More Than $1 Billion in Customer Contracts

Etched has already secured more than $1 billion in customer contracts, according to Reuters.

That figure helps explain the valuation.

Investors aren't simply betting on an idea.

They are betting on a company that has already attracted significant commercial interest.

However, contracts and deployment plans do not guarantee long-term success.

The semiconductor industry is filled with technically impressive companies that failed to compete economically against larger incumbents.

Nvidia Remains the Giant

Any discussion of AI chips inevitably comes back to Nvidia.

Nvidia has built a dominant ecosystem around GPUs, software, networking and AI infrastructure.

Competing against that ecosystem is extremely difficult.

Etched's strategy is therefore not necessarily to replace Nvidia everywhere.

Instead, it is targeting a specific segment of the market: high-volume AI inference.

That could become a large enough market to support specialized competitors.

The Economics of AI Are Changing

The funding round reflects a broader shift.

As AI models become more capable, the cost of operating them becomes increasingly important.

Companies are beginning to focus on metrics such as:

Cost per token

Inference speed

Energy consumption

Data-center efficiency

Latency

Computing density

These factors could determine which AI products can scale profitably.

A model that is slightly better but dramatically more expensive to run may not necessarily win the market.

Why This Funding Matters

Etched's $21 billion valuation suggests investors believe AI infrastructure could become one of the most valuable technology markets of the decade.

The AI industry needs models.

Models need computing.

And computing needs specialized hardware.

That creates a chain of opportunities for startups.

The AI boom therefore isn't only creating billion-dollar software companies.

It is creating an entirely new generation of semiconductor businesses.

Etched's latest funding round is a powerful signal that investors are willing to make enormous bets on that future.

The next battle in AI may not happen inside the chatbot.

It may happen inside the data center.