AI chip startup Etched defies skeptics, hits $10.3B valuation from big-name investors  | TechCrunch

AI-generated image Image credits to TechCrunch

AI hardware is no longer a speculative bet—it's a battleground, and Etched just signaled it's here to dominate. Three Harvard dropouts launching a chip company in 2022 might’ve sounded naive just a few years ago, especially when betting against GPU giants like Nvidia. But now, after a $300 million Series C at a $10.3 billion valuation—doubling from just seven months prior—it's clear investors aren’t just interested, they’re convinced.

What stands out isn’t the funding alone, but how Etched is reengineering AI inference from the ground up. Most chips treat inference as a monolithic task. Etched splits it intelligently: a dedicated prefill chip optimized for heavy compute at low voltage, and a revolutionary cluster-scale memory system for the memory-hungry decode phase. This isn’t incremental improvement—it’s architectural reinvention. And unlike assumptions, their systems aren’t locked to transformer models. They’re running Mamba, MoE, and more, proving flexibility at scale.

In my experience prepping founders to pitch investors, one thing stands out: traction beats skepticism. Etched’s secret weapon wasn’t just tech—it was private demos. Getting names like Karpathy, Hinton, and Thiel wasn’t about hype; it was about letting elite minds touch and test the hardware. That kind of validation is priceless when you’re challenging the status quo.

And let’s not gloss over the human story. Sleeping on floors, servers in garages, reboot calls to an employee’s wife—this is the raw, unglamorous hustle so many founders endure. Today, they're running a 10MW facility in Milpitas and shipping full rack systems. The scale shift is staggering.

What’s next? Mass production and delivery. The road from prototype to widespread deployment is steep, but if they maintain momentum, Etched could redefine how AI workloads are powered—not just faster, but cheaper and more efficient.

Want to see what the future of AI infrastructure looks like? This is one to watch closely.

This post has originally been written by TechCrunch on Thu, Jul 23, 26. Find the original post here at TechCrunch
Connie Harrell

Working with investors and entrepreneurs to gain the best ROI possible.

All publishers posts
Related Posts
Abliteration.ai is making a business out of rem...

A new service offers uncensored AI models capable of generating exploit code and dangerous biopro...

Own the Language Skills That Help a Business Tr...

Language skills are a strategic advantage in business—here’s how to build them with a proven, lin...

Robert Irvine on How Entrepreneurs Can Overcome...

What if 'impossible' is just a story you're telling yourself? Robert Irvine breaks down how entre...

AI data startup Micro1 reaches $500M gross run ...

A quiet player in AI training data is scaling fast—and taking a stand on ethics in the process.

Konstantin Sintsov: Business Expansion, Philant...

How Konstantin Sintsov turned rail logistics into a legacy of community and sport

ChatGPT Is Now Giving Out Personal Finance Advi...

AI is reshaping how Americans manage money — but blind trust could lead to costly mistakes.

0 comments
Write A Comment As Guest