
GMO Research Report: Before Earnings Disprove the AI Bubble, US Stocks May First Be Crushed by Equity Supply
Author: Ben Inker, John Pease
Translated by: TechFlow
TechFlow Guide: GMO Asset Management heads Ben Inker and John Pease target equity supply in their Q2 2026 letter: SpaceX unlockings, potential IPOs from Anthropic and OpenAI, plus net new shares from hyperscale cloud providers. The US stock market is shifting from an annual contraction of about 1% in equity supply to an expansion of nearly 5%. They estimate that for every 1 percentage point increase in supply, market capitalization drops by roughly 4% over the following year; this single headwind alone could reduce returns by about 20% compared to normal levels—the market may be pushed down before AI revenue realities are disproven.
Summary
The market pullback appears to be driven by a wave of internet-related stock issuances.
On June 12, 2027, approximately $2 trillion in SpaceX shares will have their contractual "lock-ups" lifted, becoming fully tradable.
After controlling for valuation factors, historically, every 1 percentage point increase in the IPO-to-market-cap ratio has corresponded to roughly a 4% drop in stock prices over the following year.
We are transitioning from a world where equity supply contracts by about 1% annually to one where it expands by nearly 5%, or even more.
We believe the surge in equity supply already flooding the market is sufficient to exert a meaningful suppressive effect on future returns.

Summary
As AI infrastructure attracts massive capital, we believe the catalyst for a potential bubble burst may stem from equity supply rather than disappointing AI demand:
■ Bubbles require a constant influx of fresh capital; rising supply often delivers the final blow that pops them.
■ In this AI episode, supply may come from the unlocking of SpaceX (SPCX) shares, likely upcoming IPOs from Anthropic (ANTHRO) and OpenAI (OPENAI), and follow-on offerings led by hyperscale cloud providers.
■ Today's market is dominated by passive, benchmark-tracking, and heavily constrained strategies, leaving fewer buyers willing to absorb new equity supply.
■ Data shows that historically, a 1% increase in equity supply corresponds to roughly a 4% decline in market capitalization over the next year.
■ Rising supply may weigh on returns well before AI fundamentals visibly deteriorate.
Authors: Ben Inker, John Pease
Introduction
Investment bubbles are inherently dangerous. Like naturally forming Ponzi schemes, they require a continuous inflow of larger volumes of capital to sustain themselves. Nothing attracts capital quite like "apparent success": as bubbles expand and early entrants get rich quick, latecomers inevitably follow. Unfortunately, this means that losses realized when a bubble bursts often exceed the paper wealth created during its ascent.
For professional investors, avoiding bubbles is equally perilous—because exiting too early is often less forgivable than staying invested too long. Staying invested too long at least puts you among sympathetic peers during a bear market; however, clients are most furious watching others profit from bubbles, and the tracking error generated by avoiding bubble assets only widens as the bubble continues to inflate. Thus, "timing the exit" becomes a high-stakes gamble: those who spot warning signs wait for some mysterious catalyst to give them enough time to slip away near the top.
We are generally better at identifying bubbles than pinpointing their peaks; but regarding this AI bubble, we feel we know what will catalyze its inevitable curtain call: supply. Increased supply is often the proximate cause of bubble bursts. The intriguing question—and one only time can answer—is: exactly what kind of supply?
Some bubbles die from an excess of underlying physical supply. For example, the British railway boom was destroyed by overinvestment in railways: when demand failed to keep pace, industry capital returns collapsed. Other bubbles fall differently, brought down not by the physical capacity of the technology that ignited them, but by the frantic rush for financial asset supply. The dot-com bubble of 2000 seemed to belong in the latter category. Fiber optic capacity did eventually outpace traffic growth, but the bubble burst long before this became common knowledge. The market downturn appeared more driven by a wave of internet-related equity issuances—even as disappointing fundamental news about internet growth had yet to arrive, the quasi-Ponzi dynamics were already unsustainable.
If anything has changed in the equity markets since 2000, it is increased sensitivity to supply. This makes "equity supply" highly likely to impact the market before AI overinvestment truly outweighs credible demand. Plenty has been written about how much AI revenue is needed to justify today's colossal investments. We speculate that eventual demand will fail to catch up with supply, and the investment returns on AI infrastructure will prove disappointing. Yet far less has been written about why equity markets have grown increasingly sensitive to shifts in stock supply and demand. The remainder of this piece focuses on that sensitivity, and how it compounds with the monumental equity supply expected over the coming quarters—enough to easily push the market downward even before AI revenues significantly miss investor expectations.
Financial Supply and Demand
On June 12, 2027, approximately $2 trillion in SpaceX shares will have their contractual "lock-ups" lifted, marking the final batch ready for circulation (see Figure 1). This will allow individuals and institutions—who bought or were granted shares while the company remained private, ranging from endowments and foundations to company employees—to sell their stakes to public market buyers. Our best guess is that most non-Musk shareholders will opt to sell. Employees need cash to pay off mortgages or service interest on debt secured by SpaceX stock; foundations face tight cash flows in a "lower dividend" world and want to avoid concentrating risk in a single name; endowments are especially liquidity-constrained. Fortunately, SpaceX does not seem to lack buyers.
Figure 1: Scale of IPO Unlockings
As of June 30, 2026 | Source: GMO, SpaceX filings

To buy SpaceX shares, investors must pay market prices to former private shareholders. Most investors do not hold large pools of idle cash, so they are likely to fund purchases by selling assets—with liquid stocks being the most likely candidates. If these former private shareholders subsequently recycle all the cash back into other liquid equities, the transaction places minimal pressure on the market, essentially functioning as a swap between SpaceX stock and other holdings.
But if they do not merely swap equities, and instead use part of the proceeds to pay capital gains taxes, reduce leverage, fund new AI-themed venture capital, or buy caviar, then it is no longer a straight stock-for-stock swap. Former private shareholders will not automatically become the counterparty to "the pile of stocks that must be sold to buy SpaceX." Consequently, public equity supply will rise above demand; as witnessed after successive waves of large IPOs, the result is lower stock prices. As shown in Figure 2, after controlling for valuations, historically, every 1 percentage point increase in the IPO-to-market-cap ratio has corresponded to roughly a 4% drop in stock prices over the following year.
Figure 2: US IPO Issuance vs Next-Year S&P 500 Returns
As of June 30, 2026 | Source: Jay Ritter, Compustat, Worldscope, Bloomberg, MSCI, GMO
Both IPO size and next-year S&P 500 returns are residualized against contemporaneous forward one-year earnings yields (Frisch–Waugh–Lovell), so the slope represents the IPO coefficient conditional on fixed valuations.

At face value, increased equity supply should not matter. Financial assets—unlike consumer goods—are priced solely based on their claim on corresponding cash flows. The price of a Starbucks (SBUX) coffee depends on consumer preference (demand) and the availability of substitutes (supply); meanwhile, the price of Starbucks stock should depend only on investors' expectations for future cash flows, discounted at a rate compensating for risk. Whether 100 or 1,000 shares trade on the secondary market, or whether other coffee chains go public, ought not to affect Starbucks' intrinsic value—unless such events alter the company's projected cash flows or perceived risk. This implies that in a world characterized by profit maximization and broad asset choice, misalignments in risk-return profiles caused by supply-demand shifts should theoretically be arbitraged away. An idiosyncratic capital flow—"my leverage on AI stocks is too high, I need to cut everything now; alas"—might cause temporary price dislocations, but profitable arbitrageurs would quickly swoop in to restore prices to fair value without requiring much time.
In reality, there are indeed pure profit-seekers specializing in exploiting price deviations. But more commonly, reality features constraint-bound profit maximizers. For instance, a US equity fund manager might see European stocks or Argentine bonds as cheap, but prospectus restrictions forbid her purchase; she might also view the broader US market as expensive, yet struggle to meaningfully reduce her position. While this sounds reasonable—an US equity fund should hold US equities—it remains mostly constrained even within its mandate: she might dislike certain benchmark weightings but be forced to hold them due to relative positioning limits; she might perceive today's benchmark-relative opportunity set as excellent but be unable to take on additional risk due to tracking error caps. Even within her own asset class, this investor cannot simply chase profit maximization.
Why are there so many constraints? Institutions outsource portfolio management, making it easier to compare and track managers within specific "silos"; underperforming managers possess asymmetric incentives to take oversized bets, making clear risk limits crucial; and benchmarks are easily replicated by passive products, serving as convenient yardsticks for evaluating active performance. More interesting than the reasons, however, are the consequences.
The first consequence is market fragmentation. Professionalism tends to anchor pricing across "obviously adjacent" securities while driving wedges between seemingly dissimilar assets. As more managers shift from cross-asset, cross-class comparisons to rotating within narrow sub-sectors of a single asset class, comparative analysis diminishes. If your job is trading US equities, you might still compare Microsoft (MSFT) and Exxon (XOM); for a growth-style manager, this is rarer; for someone on a focused growth-tech team, it is virtually impossible. The fewer people willing to trade between MSFT and XOM, the more likely their prices drift apart. A trend that grows clearer over time is rising correlation among stocks with similar business models, alongside persistent correlations hovering near zero for stocks spanning disparate industries and styles.
Figure 3: Average Pairwise Correlation
Most Similar vs. Least Similar Stocks (Russell 3000 proxy)
As of June 30, 2026 | Source: Compustat, Worldscope, Bloomberg, MSCI, GMO. Similarity is based on GMO Asset Allocation Risk Model characteristics and GICS sectors; calculated as pairwise correlation of trailing 36 months' monthly residual returns (individual stock minus that month's market-cap-weighted universe return). Most similar = highest decile within same-industry pairs; least similar = lowest decile across cross-industry pairs.

The second consequence is that benchmarks matter disproportionately. In the 1980s, of every dollar invested in US active equity mutual funds, roughly 80 cents was allocated to off-benchmark positions. This figure has declined yearly, reaching about 60 cents by 2025, partly due to diminished reliance on fundamental stock picking and partly because the overall approach is more "benchmark-aware." The decline may seem unremarkable until you consider that today over 50% of equity assets flow into passive products (versus less than 10% in 1980), which make zero off-benchmark allocations. We have moved from a world where roughly 75% of capital continuously sought better opportunities to one where "active" capital accounts for less than half. In plain terms: of every dollar flowing into the stock market, only about 30 cents is used opportunistically to trade securities. This means that for anyone wishing to execute a trade, the available counterparties represent a vastly smaller fraction of the market.
Figure 4: Active Share of US Active Equity Funds Benchmarking the S&P 500
As of June 30, 2026 | Source: Cremers & Petajisto, Thomson Reuters, Factset, GMO. Sample consists of US common equity funds holding at least 10 stocks, with AUM of at least $10 million, and active share exceeding 50%.

This is startling when you consider it: the vast majority of US equities held by professional investment managers will never be bought or sold, regardless of price movements. Theoretically, stocks could still be perfectly priced via discounted cash flows—you "merely" need the remaining active capital to remain highly sensitive to price and information—but this clearly diverges from reality. First, even amid extreme price volatility, we observe minimal portfolio turnover; second, academic research on exogenous demand shocks like index inclusions or mutual fund forced liquidations corroborates this price insensitivity. On average, these studies find: a 1% increase in demand for a given stock typically drives its price up by roughly 1%.
Frankly, a 1% capital flow translating to a 1% price move hardly warrants alarm—roughly equivalent to the price elasticity of ordinary consumer goods like cars or seafood. But individual stocks ought not to be traded like commodities. Theoretically, any equity has hundreds of substitutes evaluated by the same metric (risk and return); even if capital flows have only a modest mid-term impact on prices, it proves we are far from a realm of "pure profit maximizers." This disparity grows even wider when scaling from individual stocks with abundant substitutes to the broader US equity market, where substitutes are scarce.
The Equity Market Cares Deeply About Capital Flows
Few investment vehicles actively and significantly adjust US equity allocations. As of May 2026, of roughly $30 trillion in combined active and passive mutual funds and ETFs, approximately 83% resides in vehicles that exclusively trade US equities. In fact, only about 3% of (non-target-date) equity funds can freely move between stocks and other asset classes. This means that even if all asset allocators behaved like GMO—meaningfully rebalancing portfolios as opportunity sets shift—they still could not fully offset the price dislocations originally caused by capital flows.
Ultimate asset owners, such as retirement accounts and sovereign wealth funds, show little inclination to materially change allocations either. Within 401(k) and IRA accounts—which constitute the bulk of approximately $45 trillion in retirement assets—only about 10% adjust their asset mix annually; sovereign funds similarly resist altering target weights. The world's largest sovereign fund adjusted its target equity allocation just twice over the past two decades, the most recent being a shift from 60% to 70% in 2017. All this points to a stark conclusion: very few dollars are genuinely chasing opportunities across asset classes.
The inference mirrors the individual stock level: once capital flows cross asset classes, the impact is substantial. Our best estimation of this effect—observing how various predictable, non-informative flows impact US equity prices—is: a 1% increase in equity supply correlates with roughly a 4% decline in total market capitalization over the subsequent 12 months. Given that few investors adjust equity positions in response to falling prices, this is hardly surprising; nor does it help that US investors carry record-high equity allocations—never before has such a large proportion of liquid assets been locked into publicly traded stocks.
Figure 5: Aggregate Investor Equity Allocation Ratio
As of June 30, 2026 | Source: Jay Ritter, Compustat, Worldscope, Bloomberg, MSCI, GMO. Based on a metric proposed by Philosophical Economics: domestic corporate equity liabilities (non-financial plus financial) divided by equity plus debt securities and loans held by households, non-financial corporations, federal government, state/local governments, and foreign entities. Shaded areas denote NBER recession periods.

Consider what is happening to equity supply now. Potential sales by non-Musk SpaceX shareholders alone exceed 1% of US equity market capitalization. Over the next 12 months, we are highly likely to see IPOs from Anthropic and OpenAI—combined valuations roughly equivalent to 5% of the total investable US stock market. For the first time in two decades (excluding a brief exception during the global financial crisis), the US equity market faces net dilution, meaning new issuances outpace buybacks, as hyperscale cloud providers continue their capex sprint. Collectively, we are transitioning from a world where equity supply contracted by about 1% annually (buybacks exceeded primary or secondary issuance) to one where it expands by nearly 5%, or more.
Figure 6: US Net Primary and Secondary Equity Issuance
As of June 30, 2026 | Source: Jay Ritter, Compustat, Worldscope, Bloomberg, MSCI, GMO

If you subscribe to our 4x multiplier, this single headwind of rising equity supply implies returns over the next 12–18 months could lag normalcy (approximately 6% real returns) by roughly 20%. That 20% discount excludes today's exceedingly rich valuations, elevated profit expectations, and macro/geopolitical issues cheerfully ignored by global markets thus far. During the dot-com bubble, new issuance arguably ignited the first leg of the bear market; once prices began falling, high valuations and margin compression from that cycle's overinvestment turned a typical drawdown into a 50%+ crash—the digital darlings suffered far worse. There is nothing here to suggest today's market will script a different ending.
Conclusion
The rising sensitivity of stock prices to supply shocks carries implications for professional investors far beyond aggregate market levels. Coupled with mechanisms like leveraged single-name ETFs and absurdly leveraged AI hedge funds that generate supply-demand shocks, we have already witnessed how severely this sensitivity can shake prices. Frankly, we suspect this topic has received scant attention because investors capable of exploiting parts of its premise have no incentive to explain their methodology. But when addressing the broader market, we are not discussing arbitrage opportunities; we are describing a shift in the risk-reward tradeoff—a shift significant enough to warrant sharing, harmless to our strategy, and therefore offered openly. We believe the equity supply already flooding the market is sufficient to exert a meaningful suppressive effect on future returns—if history remains a reasonably reliable guide, returns over the next 18 months may trail normal levels by roughly 20%.
Of course, this is far from certain. Demand could grow alongside supply, allowing the market to sidestep stumbles in the coming quarters. But bubbles feature Ponzi-like positive feedback: if this round of supply is calmly absorbed by the market, it merely incentivizes further supply, ultimately overwhelming the market's absorption capacity. Meanwhile, raised capital will almost certainly continue pouring into additional AI capacity, eventually compressing investment returns in the sector. Given enough time, weak returns will terminate the bubble—because supply increments will inevitably outpace demand, and aggregated AI investment returns will prove disappointing. Yet in a market this sensitive to equity supply, prices are highly likely to pivot long before the market truly "knows" all of this.














