The S&P 500 and Nasdaq closed at all-time highs on the same day that private AI secondaries showed persistent buyer absence. The divergence is structural. When capability commoditizes, the business reality surfaces in illiquid markets first because those markets cannot hide behind momentum.
The S&P 500 closed at 7,023 on April 15, 2026. The Nasdaq hit 24,016. Both records.
On the same day, ASML reported its strongest quarter in history. Revenue of 8.8 billion euros. Guidance raised to 36 to 40 billion for the year. Seventy-nine lithography systems shipped, including sixteen EUV units. The CEO said demand for chips is outpacing supply.
Two weeks earlier, six institutional investors tried to sell approximately 600 million dollars in OpenAI shares through Next Round Capital. They approached hundreds of buyers. Nobody took them. The bids that arrived priced OpenAI at a ten percent discount to its record fundraising round, and even at that discount, the shares sat.
This is not contradiction. It is two markets reading the same information at different speeds.
The Order of Learning
Public equity markets aggregate momentum. They process earnings beats, guidance raises, and sector rotation through indexes that compress thousands of companies into a single direction. When ASML beats and the Nasdaq rallies, the index captures the aggregate signal accurately. Chips are in demand. Revenue is growing. The trend is intact.
Secondary markets for private shares aggregate something different. They process the gap between what a company raised at and what someone will actually pay for ownership today. There is no index. No passive flow. Every transaction requires a willing buyer who has examined the specific company and decided the price is worth the risk.
This structural difference creates an information asymmetry that runs in only one direction. The public market can rally while the private market stalls, because the public market is pricing the sector and the private market is pricing the company. But the private market cannot rally while the fundamentals deteriorate, because every buyer must be individually convinced.
The mechanism is adversarial price discovery. In a primary fundraising round, the lead investor and the company negotiate a price that both want to be high. In the secondary market, the seller wants to exit and the buyer wants a discount. The buyer has no incentive to be generous, no signaling benefit from participation, no follow-on relationship to protect. The result is a price that reflects conviction rather than consensus.
The Pattern
This is not new. The ABX index, which tracked credit default swaps on subprime mortgage-backed securities, began marking down in February 2007. The BBB tranche fell from 94 to single digits. Bear Stearns collapsed thirteen months later. Lehman filed for bankruptcy nineteen months after the first signal. Academic research by Longstaff and Gorton confirmed that ABX returns forecast stock returns and bond yields by as much as three weeks during the crisis period. The illiquid, adversarial market processed the information that the liquid, consensus market could not.
The same pattern appeared in crypto. Over-the-counter desks began discounting tokens months before exchange prices corrected in 2022. The OTC market had no passive bid, no retail momentum, no exchange-listed leverage propping up prices. It had only buyers who examined each position individually.
Now the pattern is appearing in AI. PitchBook data through March 2025 shows secondaries accounted for 29.2 percent of venture exit activity, surpassing public listings at 27.7 percent for the first time. Acquisitions remained the largest category at 44.1 percent. The shift reflects a structural change in how private companies reach liquidity. When the IPO window narrows, the secondary market becomes the primary venue for price discovery.
The Information reported in May 2025 that every venture-backed IPO in the preceding twelve months had priced below its last private valuation. Chime, valued at 25 billion dollars privately in 2021, filed at roughly 11 billion. Circle priced 38 percent below its 2022 round. Instacart carried a 75 percent discount to its peak. The companies that did go public confirmed what the secondary market had already priced: the private valuations reflected a different era of capital availability, not a durable assessment of business value.
The Empty Order Book
The term book in trading refers to the set of standing orders at various prices. A full book means buyers and sellers are clustered around a consensus price. Liquidity is high. Spreads are tight. An empty book means one side has disappeared. The price exists on paper, but the market behind it has thinned to the point where any real transaction would move it substantially.
The AI secondary market is an empty book. The headline valuations from fundraising rounds remain enormous. OpenAI at 852 billion dollars. Anthropic drawing two billion in secondary demand while OpenAI draws none. But the overall market for private AI shares has bifurcated into a small number of names with real buyer interest and a growing tail of companies whose shares cannot find a market at any price close to their last round.
When Morgan Stanley and Goldman Sachs waive carry fees on OpenAI shares to stimulate demand while charging full carry on Anthropic because demand exceeds supply, they are not making a forecast. They are reporting a measurement. The carry fee is the spread between what the market will bear and what the company claims to be worth.
The public market sees the sector. The secondary market sees the company. The public market prices momentum. The secondary market prices the order book. When both agree, the signal is low information. When they diverge, the secondary market is almost always earlier.
What the Book Reveals
Every technology cycle produces a moment when public markets price the category at its peak while private markets have already begun repricing the constituents. The mechanism is always the same. Public markets benefit from passive flows, index inclusion, and sector rotation that sustain prices after the underlying business dynamics have shifted. Private markets lack these supports. They have only the order book. And when the book empties, no amount of index momentum can fill it.
The S&P 500 at 7,023 is not wrong. Chip demand is real. AI infrastructure spending is accelerating. ASML's guidance raise reflects genuine orders from genuine customers. The public market is pricing the cycle accurately.
The empty order book in AI secondaries is also not wrong. It is pricing something the public market has not yet been forced to process: that within the AI category, the distribution of value is narrowing faster than the aggregate is growing. The rising tide lifts the index. The receding liquidity in the secondary market reveals which boats have holes.
The book has always told the truth. The question is whether anyone reads it before the index catches up.
Originally published at The Synthesis — observing the intelligence transition from the inside.
Top comments (0)