Why AI startups are selling the same equity at two different prices.

Source: TechCrunch

Translation: Venture capital has invented a new instrument for manufacturing the appearance of winners and the startups accepting it are trading long-term structural risk for short-term narrative dominance.

A new funding structure has emerged in competitive AI rounds: lead investors split their capital across two valuation tiers within a single round.

In Aaru’s case, Redpoint invested the bulk of its check at a $450 million valuation, then a smaller portion at $1 billion; the price at which all other investors joined. The company then announced itself as a unicorn, valued at over $1 billion, despite the fact that the majority of equity in that round changed hands at less than half that figure.

Serval ran a similar structure: Sequoia’s lowest entry was at a $400 million valuation, while the announced Series B headline was $1 billion.

The mechanism consolidates what would have been two sequential fundraising rounds into one transaction, with built-in tiered pricing. The lead VC gets a discount. Everyone else pays a premium.

The startup gets a headline number. And the market gets a signal it didn’t fully earn.

The distortion is structural, not merely cosmetic.

Valuation, in theory, reflects market consensus on what a company is worth at a given moment. This mechanism deliberately fractures that consensus within a single transaction: producing two prices for the same equity, in the same round, at the same time. The “headline” valuation is not a market-clearing price. It is a marketing output.

The framing offered by VCs (that this protects founders from distraction caused by serial fundraising) is partially true and mostly convenient. It is also true that it allows lead investors to secure favorable entry positions while publicly anchoring the company to a number that serves their portfolio narrative.

What gets obscured is the blended reality: the actual average price paid is materially lower than the announced valuation. The unicorn label, the press release, the talent recruitment pitch are all built on the headline instead of on the blend.

The people reading those announcements are not told the structure. And that gap is not accidental.

The incentive map here has three distinct layers, and none of them are fully aligned.

For lead VCs, the incentive is dual: favorable entry economics and competitive moat-building. A $1 billion headline valuation on a company where you hold the majority of equity at $450 million is an extraordinary position, and, as Jason Shuman notes, it functionally discourages competitors from backing rival startups. Why fund the number two player when the number one is already a unicorn?

For startups, the incentive is narrative capital. Unicorn status attracts talent, signals market leadership to enterprise customers, and generates press. In a crowded AI market where perception of dominance shapes actual dominance through partnerships, hiring, and distribution deals, the headline valuation is a competitive asset, not just a financial one.

For follow-on investors paying the premium tier, the incentive is access. Oversubscribed rounds in hot markets create artificial scarcity. Paying above the true blended price becomes the cost of being on a cap table that might otherwise be closed to them. They are not buying value. They are buying proximity.

The immediate consequence is a cohort of AI startups carrying headline valuations that do not reflect the price at which most of their equity was actually acquired, but that will define the floor for their next round.

That is the trap embedded in the structure. The next fundraise must clear the headline number, not the blend. A down round below $1 billion is punitive regardless of whether the blended entry was $600 million. Employees face dilution. Founders face credibility erosion. Customers and partners who chose the company partly on the basis of its apparent market position feel misled.

The 2022 correction is the relevant precedent. Startups that raised at aggressive headline valuations during the 2021 frenzy spent the following two years doing painful flat or down rounds, renegotiating terms, and losing staff who had priced their options against numbers that no longer held. The mechanism now being deployed is more sophisticated than what produced that correction but the underlying logic is identical: borrow credibility from the future to win competition in the present.

The systemic consequence is subtler. If tiered pricing becomes standard, valuation loses its function as a signal. Announced round sizes and unicorn counts (already imperfect proxies) become even less reliable as market intelligence. Investors, employees, and customers will need new frameworks to assess actual company health, and those frameworks don’t yet exist at scale.

The honest read of this mechanism is that it is a rational response to irrational competitive pressure and that it creates the conditions for the next correction.

When markets are hot, VCs compete for access. When they compete for access, they invent instruments that let them win deals without appearing to overpay. Tiered pricing is that instrument. It is not fraud. It is not illegal. But it is a valuation mechanism designed primarily to produce a number that serves a narrative, rather than one that reflects consensus reality.

For founders, the calibration question is whether the headline is worth the ceiling it creates. Unicorn status is a recruitment and marketing asset. It is also a contractual obligation: the next round must validate or exceed it, regardless of what the market looks like when that moment arrives.

For investors at the premium tier, the question is whether access is worth overpaying for entry in a market where many of these companies will not survive to a liquidity event.

And for the broader ecosystem, the calibration is recognizing that manufactured perception of market winners does not create actual market winners. It just makes the eventual sorting more expensive for everyone downstream.

The high-wire act Jack Selby describes is real. The wire is the gap between what a company announced and what it actually earned. The longer that gap holds, the harder the fall when it closes.​​​​​​​​​​​​​​​​