Paul Tudor Jones and Goldman Sachs agree on more than the headlines suggest. Jones calls 2027 "the most negative flow picture for the U.S. stock market that we've had, say, since 2002." Goldman's Ben Snider, writing the bull case, projects "the largest increase in net US public equity supply since 2000." On the supply arithmetic, the macro trader and the underwriter's strategist land in the same place. They split over the demand side. Goldman's offset rests on households and foreign investors, the most pro-cyclical buyers in the market, while the supply arriving next year is built to sell into exactly the strength that keeps those buyers engaged.
That is the flow read the price tape hides. The S&P 500 sat within 1% of its record as of Snider's note, so the index looks calm. Underneath it, a large block of private holders is about to gain the ability to sell, and the price-insensitive corporate bid that has absorbed supply for four years now depends on a financing condition Goldman itself calls a vulnerability.
Float, Not Shares, Is What Unlocks
The $1.7 trillion figure needs careful handling. Lockup expirations do not create new shares. They convert shares that already exist, held by founders, employees and venture investors, into tradable float. Goldman's "net public equity supply" counts float, which is why the 2027 number dwarfs the $600 billion of fresh corporate issuance Snider expects.
The difference matters for how the supply behaves. New issuance is a fixed quantity sold on a known date. Unlocked stock is optional: holders sell when they choose, and they choose based on price, tax position and diversification needs. Holders with a cost basis near zero, as early SpaceX or SK hynix shareholders likely have relative to current valuations (an inference, since individual cost bases are not disclosed), are the least price-sensitive sellers in the market. They do not need a high price. They need a liquid one.
So the $1.7 trillion functions less like a wall that falls on a single day and more like a standing offer above the market. Rallies in recently listed names meet sellers. Supply of this kind often caps upside before it forces downside, which is a different setup from a crash call and closer to what Jones described: years when being short was "easiest" because every bounce met fresh supply.
The 2026 issuance pipeline sets up that calendar. Snider counts $431 billion raised year to date through IPOs, follow-ons, converts and SPACs, up 98% year over year, with full-year dollar volume likely to clear the 2021 record of $540 billion. SpaceX's $75 billion IPO and SK hynix's $26.5 billion US cross-listing are the anchors. Standard lockups often run around six months, so a heavy 2026 issuance year turns into a heavy 2027 unlock year almost mechanically. Lock-up terms vary by deal, and the exact expiration dates for the largest listings are not established here.
Two Ledgers That Both Balance
Goldman's case is not naive, and its arithmetic deserves to be stated fairly.
On the corporate ledger alone, demand overwhelms supply. Snider projects $1.7 trillion of buybacks in 2027, up from $1.5 trillion in 2026, with cash acquisitions pushing total corporate equity demand above $2 trillion. Against $600 billion of new issuance, US corporates in aggregate remain heavy net retirers of stock. Snider also notes that 2027 issuance should equal slightly less than 1% of US market cap, in line with the 10-year average, and that deal counts and issuance scaled to market cap in 2026 are near historical norms. On that denominator, the "tsunami" is a normal tide on a much larger beach.
Add the unlocks and the ledger flips. Gross potential supply becomes roughly $2.3 trillion ($1.7 trillion of unlocks plus $600 billion of issuance) against a corporate bid of a little over $2 trillion. Whether 2027 is a supply problem depends on how much of the unlocked float actually trades, and that number cannot be known in advance.
Both men can therefore be right on their own terms. Snider is right that corporate issuance is ordinary relative to market size. Jones is right that the float coming loose is not ordinary. The disagreement lies in the residual: what absorbs the gap if the corporate bid does not grow fast enough.
There is also a distribution problem the aggregate figures smooth over. Buybacks are concentrated in large, cash-generative incumbents, while unlocks are concentrated in recently listed names. The bid and the offer do not meet in the same securities. Apple repurchasing its own stock does nothing for a SpaceX insider looking for a buyer. At index level the flows may net. At the level of individual names, they collide.
The Bid Goldman Is Counting On
Goldman fills the residual with non-corporate demand. Federal Reserve data cited by Snider show US household equity purchases above $800 billion in H1 2026, with non-US investors adding more than $350 billion, for over $1 trillion in a half-year. Annualize that and the gap closes comfortably.
That is the weak link. Corporate buybacks are the most price-insensitive bid in US equities. Companies execute authorizations largely regardless of the tape, subject to blackout windows and financing. Household flows behave the opposite way: they tend to follow recent returns, and they slow or reverse when returns turn. Extrapolating H1 2026 into 2027 assumes the conditions that produced it continue. The supply side of 2027 is structured to undermine those conditions, because unlock selling caps the price gains that attract household money in the first place.
The breadth data make this concrete. With the index within 1% of its high and the median stock 16% below its own high, Snider's breadth gauge has weakened to its lowest level since the dot-com era. Index-level returns are being carried by a narrow group of AI-heavy megacaps, which are also among the largest buyback programs. Household buyers looking at the index see a bull market. Holders of the median stock have not had one.
The policy push fits this picture. Treasury Secretary Scott Bessent framed the new Trump Accounts as "the beginning of the ownership society." Whatever the program's merits, the flow implication is plain: a new pool of retail demand arrives just as private holders gain the right to sell. It would be a stretch to claim the program was designed for that purpose. It would be naive to ignore the timing.
The pipeline is still lengthening. Anthropic and OpenAI have both confidentially filed draft S-1 registrations. If either lists, the 2028 unlock calendar starts filling before the 2027 one has cleared.
Where 2027 Differs From 2000
Jones's analog is specific: 1998 through 2002, when IPO issuance and rolling unlocks kept adding float. It is worth stating what transfers and what does not.
What transfers is the mechanism. In 2000 through 2002, unlock supply hit a market where the marginal buyer had already turned. Supply did not cause the bear market alone, but it removed the floor every time prices tried to stabilize. That is the trade Jones is describing: a short position held against repeated rallies that fail into supply.
What does not transfer as cleanly is the corporate bid. The dot-com era did not feature a $1.7 trillion annual buyback machine concentrated in highly profitable index leaders. Today's largest companies retire stock on a scale that has no 2000 equivalent, and that bid supports the cap-weighted index directly. A full-year short of the S&P 500 is a bet that the buyback bid shrinks, or that household selling outruns it. A short of recent IPOs and high-unlock names is a narrower bet that requires neither.
Passive mechanics add a less visible channel. If newly listed mega-caps become eligible for major indexes, passive funds must buy them and fund those purchases by trimming existing holdings. That shifts supply onto incumbents without any change in sentiment. Index eligibility depends on committee rules around float and profitability, so whether and when names like SpaceX qualify is unresolved. If they do, the supply wave reaches the stocks that look insulated from it.
What Would Settle It
Goldman named the variable that decides this, and it is not the unlock calendar. It is bond yields. Snider explicitly flags that further rate increases "could slow buyback growth as debt financing becomes more expensive relative to equity," and that weak breadth, AI concentration and record net supply form "a troubling backdrop if bond yields continue rising." The bull case requires declining rates. Goldman's own conditional path runs through energy: easing tensions between Russia and Ukraine or in the Gulf lowering crude and diesel prices, cooling inflation expectations and easing yields.
That gives a clean test. If yields fall, buyback growth holds near Goldman's $1.7 trillion projection, the corporate bid absorbs most of the float, and unlock selling shows up as underperformance in recent listings rather than index damage. If yields rise, the price-insensitive bid weakens just as the price-sensitive bid is asked to carry more of the load, and Jones's flow picture becomes an index-level problem.
Two observable conditions will show which path is underway before the index does. The first is the gap between the cap-weighted index and the median stock: if the S&P 500 holds near highs while that 16% gap widens, the narrow buyback-supported leadership is absorbing the supply and the rest of the market is not. The second is household flow persistence through the first quarters of 2027, as the largest 2026 listings come off lockup. If retail keeps buying through those expirations, Goldman's offset is real. If household purchases slow while unlocks accelerate, the market will be left with a sharply reduced bid and a growing supply of stock.
