Hark's $6 Billion Valuation With No Product Actually Makes Sense

1 week ago 8

12th Breakthrough Prize Ceremony - Arrivals

Last week, Brett Adcock's new startup, Hark, raised a $700 million Series A at a $6 billion valuation. Parkway Venture Capital led the round, while Nvidia, AMD, Intel, and Qualcomm all invested — putting virtually every major AI-chip maker on a single cap table. Adcock himself had already put in $100 million of his own.

Hark has been notably secretive about the rest. It’s described a concept — a "universal" agentic AI assistant, part software and part dedicated hardware, meant to be a single interface to everything you do digitally. Beyond that it has stayed deliberately quiet, with the first models promised this summer and hardware to follow. Stealth is pretty ordinary for a young company, but raising $700 million in stealth, pre-product, much less so.

The Round Skips The Company

A Series A used to be a stage marker. Typically, it suggested that a company had built a product, found early customers and needed capital to scale. The number attached to it, usually something in the range of $10 - $30 million, was a rough measure of how far along that company was.

Hark’s Series A is an entirely different story. There’s no product to scale, no customers to name, no traction to underwrite. So what is the $700 million for?

Cheaper To Build, Bigger To Fund

Building a company has never been cheaper or faster. AI writes much of the code, small teams ship what once took large ones, and the cost of frontier models keeps falling. Two engineers can now deliver products that used to require 50-person teams. Every intuition says capital requirements should be dropping. Instead rounds are at sizes that read as typos.

Hark is not an outlier, a pattern is becoming clear. Ilya Sutskever, OpenAI's former chief scientist, raised $2 billion for Safe Superintelligence at a $32 billion valuation, with roughly 20 employees and no product. Mira Murati, OpenAI's former chief technology officer, closed $2 billion for Thinking Machines Lab at a $12 billion valuation, the largest seed round on record. OpenAI and Anthropic are each raising in the tens of billions on top of that. Amounts that would have looked imaginary three years ago are now the entry ticket.

So why does easier building come with bigger rounds?

The Capital Moat

Because the money is not really for building. It’s for the moat.

When a product is cheap and fast to build, it’s also cheap and fast to copy. The technical edge that used to defend a company — proprietary code, a hard-won architecture — erodes almost as soon as it appears. What doesn’t erode is the scale of capital behind you. A competitor can rebuild your feature in a weekend, but it can’t rebuild a $700 million balance sheet as quickly.

That balance sheet buys the things that are still genuinely scarce: frontier compute, the short list of researchers who can do this work, data, distribution. Hark has been almost explicit about it — the company says the new money will go toward recruiting talent and securing compute and components. In a world of limited supply, buying those things does double duty: it equips you and it denies them to everyone else. The round has stopped being fuel and became the moat.

This is where a founder's record comes back in. Adcock built Vettery and sold it to the Adecco Group for $110 million, co-founded Archer Aviation and took it public, then founded Figure AI, now valued at $39 billion. A track record like that is itself a scarce, hard-to-replicate asset — which is why investors will price it before there’s a product to inspect.

There is a fair objection to all this. Frontier AI genuinely is expensive — training runs cost billions, and a hardware company like Hark has real bills to pay. Some of these rounds fund actual costs. But the raise and the valuation are two different signals. A $32 billion valuation on a 20-person company with no product is not a compute-cost story.

Nobody Has The New Playbook

None of this means the bets are wrong. Adcock, Sutskever and Murati may each deliver, and if they do the returns will justify the prices. The harder problem is for everyone trying to read these companies from the outside. When the moat is the product, you can evaluate it — test it, measure it, watch whether customers adopt it. When the moat is the balance sheet, the thing being underwritten is the capital itself and the founder holding it. That is a different kind of judgment, and the honest truth is that nobody — founders, investors, the analysts who will price the next round — has settled on how to make it yet. At the moment, the rounds are racing ahead of the playbook.

Read Entire Article