Flourish Raises $500M to Copy the Brain and Fix AI's Power Crisis

A startup called Flourish has emerged from stealth with $500 million in funding and a wildly ambitious goal: reverse-engineer the human brain to fix artificial intelligence's biggest problem — its enormous appetite for power.

The round values Flourish at $2.5 billion and is backed by an A-list of investors, including Amazon founder Jeff Bezos, Lux Capital, Alphabet's venture arm GV, and Catalio Capital. Bezos nearly doubled an initial $50 million commitment to put in close to $100 million, making him one of the round's largest individual backers.

Why the power problem matters

Today's AI runs on massive clusters of GPUs that consume staggering amounts of electricity. Training and running frontier models has pushed data-center energy demand so high that power availability — not chips or talent — is increasingly the limiting factor on how big AI can get.

The human brain offers a jarring contrast. It runs on roughly 20 watts — about the same as a dim light bulb — yet still outperforms today's AI on many kinds of reasoning. Flourish's entire thesis is built on that gap.

Why this matters: if AI's growth is capped by how much power the grid can supply, then making AI dramatically more energy-efficient isn't a nice-to-have — it's what determines whether the technology can keep scaling at all.

What Flourish is building

The company is building a system it calls Cortex AI. The approach uses connectomics — the detailed mapping of how neurons connect in the brain — to find what the founders describe as the brain's "core algorithm," then replicate that algorithm in silicon.

The power target is audacious: Flourish wants Cortex AI to run on 20 to 50 watts, in the same range as the human brain, versus the far heavier draw of today's GPU clusters. That's not an incremental efficiency gain — it's a bet that a fundamentally different design could cut AI's energy needs by orders of magnitude.

This puts Flourish in the "neuro-AI" camp: a small but growing group of researchers who think the path to better, cheaper AI runs through neuroscience rather than through ever-larger versions of today's architectures.

The people behind it

The founders give the bet credibility. Co-founder Thomas Reardon has an unusual résumé — he created Internet Explorer at Microsoft, then founded the neural-interface company CTRL-labs, which Meta acquired in 2019 for about $1 billion. He later earned a doctorate in neuroscience.

His co-founder, Rob Williams, held senior positions at Amazon before starting Flourish. The company grew out of work tied to Catalio Capital, a firm focused on breakthrough science and healthcare.

The combination — a founder who has both shipped mass-market software and trained as a neuroscientist — is part of what attracted backers willing to fund such a long-horizon idea.

Why investors are betting now

The list of backers tells you how the smart money is reading the AI landscape. Bezos, GV, and Lux Capital are not chasing another chatbot — they're funding a moonshot aimed at the constraint everyone in AI is now talking about: energy.

The competitive implication is significant. If brain-inspired designs can deliver even a fraction of the efficiency Flourish is targeting, they could undercut the current arms race of building bigger and bigger data centers. That's a threat to the assumption — baked into tens of billions of dollars of infrastructure spending — that scaling AI means scaling power consumption.

It's also a high-risk bet. "Reverse-engineer the brain" has been an aspiration in computing for decades, and no one has cracked it. A $2.5 billion valuation for a company just emerging from stealth reflects how much investors want a solution to the power problem — not proof that Flourish has found one.

What's next

Flourish hasn't shipped a product, and turning a neuroscience thesis into working, brain-efficient AI hardware is a years-long undertaking. The near-term signals to watch will be research milestones, any early demonstrations of Cortex AI, and whether the company can recruit the rare blend of neuroscience and chip-design talent the project demands.

The bottom line: AI's next bottleneck is power, and Flourish just raised half a billion dollars on the idea that the brain already solved it. Whether or not the company succeeds, the size of the round — and the names behind it — is a clear sign that energy efficiency is becoming one of the most valuable problems in AI. Keep an eye on the neuro-AI space; it's where some of the biggest bets are now being placed.