Virgin Unicorns – GeekWire

Twelve AI labs have a combined value greater than Ford and GM. None of them sell anything. I call them Virgin Unicorns – worth over a billion dollars, but with no product liability or revenue.

OpenAI has proven that an AI research lab with the right product can become one of the most important companies in the world. A dozen other AI labs are trying to replicate this trick. They have raised more than $29 billion for a combined value of close to $130 billion, without sending anything the customer can buy.

Two questions to ask:

• Why do savvy investors write growth-stage checks to early-stage companies?
• What does history say about how this story ends?

Top Virgin Unicorns

To answer these questions, let’s identify four patterns across a group of companies.

Pattern 1: Pedigree premium. Every founder is a recognized leader in their field, and most come from a small set of institutions. About four-fifths hold PhDs, mostly in computer science from a few universities – Berkeley, Stanford, MIT, Toronto, Alberta, Cambridge, UCL – and most of the rest left their PhDs in one of those programs to start their own companies.

On the employer side, concentration remains strong. Four of the twelve companies are equipped with DeepMind alumni (Ineffable, Reflection, Recursive, Recursive Superintelligence). Two are powered by OpenAI alumni (Thinking Machines, Safe Superintelligence). AMI Labs traces back to Meta’s FAIR team, and Humans& draws its founders from across Anthropic, xAI, and Google. Stanford and Berkeley faculty appointments account for most of the rest (World Labs, Physical Intelligence, and Noah Goodman of Humans&).

Four institutions – DeepMind, OpenAI, Berkeley, and Stanford – produced the founders of all the Virgin Unicorn on the table. Investors value CVs, not products.

Pattern 2: Nvidia as king. Nine of the twelve companies at the table have Nvidia as an investor. The supplier of picks and shovels is also a shareholder in the inspectors. Nvidia gets early visibility into the most ambitious AI bets, a key to paying commitments, and achieves multiples of capital invested at near-zero costs. Selling shovels was a good business. Mining has also never happened.

Pattern 3: Tables of unusually wide caps. Each round at the table includes a group of ten to twenty investors – corporate firms, corporate strategies, private equity funds, and individuals. Sequoia and a16z still lead. But the rounds are big enough that they need a lot of equity capital — from JPMorgan, BlackRock, Alphabet, the UK Sovereign AI Fund, Samsung, Temasek, ADIA, and Bezos himself — to fill them. That makes these cycles structurally different from classic corporate finance.

Pattern 4: Post-LLM thesis. Every company argues, in one way or another, that the current paradigm is not enough – that scaling LLMs will not reach AGI, and that something else (world models, reinforced learning, agent systems, AI scientists, novel chips, formal mathematical reasoning) is needed. The thesis is the product. A product is a promise.

Others have classified these unicorns as:

• Howard Marks, in his December 2025 Oaktree memo Is It a Bubble?, described investor behavior as a “lottery ticket mentality” – investors backing startups without a product with the dream of huge profits despite high odds of failure.
• Derek Thompson, writing in October, did the same by reporting that the Thinking Machines meeting was described by one investor as “a very stupid meeting” because Mira Murati “couldn’t answer any questions” about what he was building.
• GeekWire’s year-end survey of regional investors found similar skepticism closer to home: the bubble, they say, is most pronounced in the early stages, when AI storytelling can replace real traction.

The draft lottery ticket is now common wisdom. But will the lottery pay off? Another form of disability is looking at the past.

What history teaches us

The closest of history is not the dot-com era. Webvan, Pets.com, and Boo.com failed not because they were pre-owned, but because they had bad products and business models. Those companies burned money on infrastructure and advertising, not on research.

The cautionary tales that are close to the leading celebrity brand founder of the past fifteen years.

• Magic Leap raised $3.5 billion in nine years on the strength of Rony Abovitz’s previous exit and posted a flop.
• Quibi raised $1.75 billion through Katzenberg and Whitman and held for six months.
• Inflection AI raised $1.5 billion from Mustafa Suleyman and Reid Hoffman and was successfully incorporated into Microsoft in 2024 – its team was hired, its technology was licensed, its company was drilled into the shell.

In each case, the founder’s guarantees raised the money. The product did not justify the rating.

The closest analogy, however, is biotech. About 80% of 2021 biotech IPOs were pre-revenue. The probability of a pre-clinical drug reaching commercialization is less than 10%. Development takes a decade and costs $1 billion. Yet a Bentley University study of 319 biotech IPOs from 1997 to 2016 found that the cluster generated more than $100 billion in shareholder value despite a failure rate of more than 50%. The winners were numerous enough to carry a portfolio. And many of the most successful biotechs are acquired before reaching profitability.

Virgin Unicorns are biotech-shaped businesses. Past income, science-driven, ten-year timelines, binary outcomes, acquisitions as a common exit method. But they are not funded like biotechs. Biotech investors shell out big bucks for milestones tied to specific scientific results, and expect many candidates to fail. Virgin Unicorn investors cash out in one big round on the strength of CV, and the price of success. It’s the same business orientation, as opposed to a financial concept. That conflict is where disappointment will come from.

Why is Sequoia investing anyway?

The story of OpenAI defies the biotech analogy. From its inception in 2015 to the launch of ChatGPT in late 2022, OpenAI looked like a Virgin Unicorn – a pre-consumer product for seven years, billions of money, and only research to prove it. Then ChatGPT was launched and revenue went from zero to over $10 billion in three years. No biotech has ever scaled like that.

Sequoia and other investors writing checks to today’s Virgin Unicorns aren’t pricing in biotech results. They have prices for the second coming of OpenAI.

The table above makes the size of that bet readable. Early stage investors aim for 10x returns. Most of these twelve will return zero, so the winner must carry eleven alone. With a marked-up price of 127 billion, that means the winner alone has to generate something like 1.3 billion dollars worth.

That’s not a prediction – it’s a bet VCs have already placed. Sequoia and a16z made this kind of bet on OpenAI and Anthropic, and the returns on paper have already proven many times over. Anthropic itself looked like a Virgin Unicorn in 2022 – then it shipped Claude and built up revenue.

Recorded history suggests some skepticism. But bubbles have a way of producing Amazon or Google from time to time in the middle of a crisis. Identifying which Virgin Unicorn will become a trillion-dollar company — a “kilocorn,” a thousand unicorns in one — is difficult. Which one would you bet on?