Author: Windchasing Trading Platform
This summer, global markets have been repeatedly pulled by three things: the implosion of high-leverage AI positions, the U.S. long-end bond yield hitting a 19-year high during a rate-cut cycle, and the on-again, off-again ceasefire agreement for the Strait of Hormuz.
These three events belong to three different markets and appear unrelated on the surface.
On August 10, Goldman Sachs placed them within the same framework: global markets might be transitioning from "excess savings" to "capital scarcity." AI is burning cash, governments are borrowing, and the Strait is choking physical supply—three forces simultaneously consuming limited global capital.
The central banks, which over the past two decades have been responsible for continuously injecting liquidity into markets and suppressing interest rates, now have less they can do.
This assessment doesn't mean the three issues will keep worsening. But it points to a question: If global capital is indeed becoming more expensive and scarce, are markets prepared? Even if any one of these three events calms down, the global scramble for capital might not disappear. The extreme volatility seen in July—where stocks crashed and soared, bond yields spiked despite rate cuts, and oil erased half its losses in three days—might not have been a one-off accident but the start of a high-volatility cycle.
AI Leverage: The Logic Was Right, The Leverage Was Wrong
Of the three, the first to detonate was AI leverage.
The hedge fund run by "AI Stock God" Leopold Aschenbrenner surged 439% in net value in the first half of the year, powered by 4x leverage betting long on AI hardware and short on software stocks. In July, the Philadelphia Semiconductor Index fell 28.6% in a single month, resulting in losses on both sides. $16 billion in public positions were dumped on Citadel—Millennium and Jane Street assessed and exited entirely, leaving only one buyer on Wall Street.
JPMorgan later warned: Institutions' capacity to absorb AI assets has significantly declined. If something happens again, "only retail investors will be left to hold the bottom."
But at the same time, TSMC's July revenue grew 44.7% year-over-year to a record high. The combined order backlog of the top three U.S. cloud providers was $2.3 trillion. AI demand hasn't slowed. The direction isn't wrong; the problem is that the money to bet on that direction is disappearing.
The more fundamental signal came from Nvidia. On August 10, Jensen Huang didn't launch a faster chip; instead, he teamed up with six major asset managers like BlackRock and Blackstone to build a financing platform—customers can't afford the chips, so Nvidia needs to help them find money from institutions. The market didn't buy it: the stock fell that day, and its CDS neared record highs.
"Dr. Doom" Jim Chanos said this is reminiscent of 2008. The market's skepticism is simple: If the $500 billion is essentially "lending money to customers to buy your own chips," how long can this cycle last? A sector that needs hundreds of billions in financial engineering to stay alive is itself proof of capital scarcity.
U.S. Bonds: One Explanation for Why Rate Cuts Are Failing
While AI is burning cash, governments are also borrowing at record levels.
The Fed cut rates six times, a cumulative 175 basis points, yet the 30-year Treasury yield climbed all the way, hitting 5.27% on August 1—the highest since 2007. The market is no longer pricing the Fed; it's pricing U.S. fiscal policy.
The U.S. national debt exceeded $40 trillion in August. FY2026 net interest expense is projected at $1.039 trillion, surpassing the defense budget for the first time. For every $5 in tax revenue, $1 goes directly to interest. $15 trillion in debt matures within a year and needs refinancing; every 1 percentage point higher the rollover rate is, the deficit automatically expands. The share of Treasuries held by foreign central banks has dropped from 34% to below 24%.
Goldman made it explicit: The world has shifted from "excess savings" to "extreme capital scarcity." AI infrastructure requires over $800 billion per year; nations' reindustrialization and defense rebuilding are proceeding simultaneously; sovereign debt refinancing needs are peaking—these financing demands are all crashing into a Treasury market where buyers are retreating. With buyers leaving and supply expanding, yields have nowhere to go but up.
The Fed itself is not united. Chair Walsh has drastically reduced forward guidance, stating that "markets have already done much of the tightening work"—no rush to hike, let markets tighten themselves. Cleveland Fed President Hamak publicly disagreed: current rates are "not clearly restrictive," and "multiple hikes may be needed." Goldman's assessment: the Fed looks more like a "passenger" than a "driver."
Aschenbrenner's 4x leverage was fragile not only because of concentrated positions—but also because the money he borrowed was likely getting more expensive. The AI capital feast collided with a government borrowing flood, and central banks no longer have the same capacity to backstop both as they did in the past.
Hormuz: Why a Ceasefire Agreement No Longer Works
While the financial world is pricing in capital scarcity, the physical world is sending synchronous signals.
Since the U.S.-Iran conflict began in late February, Brent crude has traced a volatile path: $126 (March conflict) → $72 (June ceasefire) → $85 (July breakdown) → $87.72 (August, Iran presents 6 conditions, Trump responds "semi-negotiation status"). Each swing repeated the same pattern: the market first priced in "it's about to stop," then repriced for "it's on again." After several rounds, no one dares easily bet "this time it's real."
The issue is no longer "will they fight," but "how long will any ceasefire last." Iran's Supreme National Security Council Secretary Zolghadr said if the U.S. doesn't change its behavior, the Strait will remain closed. Trump responded with plans to increase economic pressure. Both sides are talking, neither is backing down.
The OVX (Crude Oil Volatility Index) is currently around 60—3.2 times the VIX. The options market is clear: the biggest tail risk isn't in U.S. stocks, it's in the waterway.
What complicates the situation further is that the U.S. buffer is thinning. The Strategic Petroleum Reserve (SPR) fell to 298.7 million barrels, dipping below 300 million for the first time since 1983. It was 415 million pre-conflict, and after releasing 172 million in March, it's been in net consumption ever since. The IEA characterized it as the "largest supply disruption in the history of the global oil market." The next time there's trouble in the Strait, the U.S. government's ability to release reserves to suppress oil prices will be the weakest in four decades.
The Convergence of Three Hidden Threads: From "Excess Savings" to "Capital Scarcity"
These three events are likely not a coincidence.
AI leverage imploded because 4x leverage is unsustainable in a rising rate environment. Long-end Treasury yields soared because governments and corporations are simultaneously competing with global investors for money. The Strait of Hormuz issue, seemingly about geopolitics, has an underlying layer where a physical asset like oil suddenly becomes scarce. These three forces point in different directions—stocks, bonds, commodities—but share the same root cause: the global capital supply-demand dynamic may have undergone a structural reversal.
Goldman calls this shift "from excess savings to capital scarcity." The post-2008 world was: lots of savings, few projects, low rates. Central bank QE backstopped everything, governments borrowed almost interest-free. This framework lasted nearly two decades.
In 2026, it's being pierced from both ends. On the supply side, global central banks have shifted from net buyers to net sellers of Treasuries. On the demand side, AI infrastructure, reindustrialization, defense spending, energy transition—four directions, releasing trillion-dollar capital demands simultaneously. Goldman specifically noted that AI capital expenditure, among these four forces, "puts the most significant upward pressure on rates"—it pushes up both the risk-free rate and credit spreads.
Even if the Strait opens tomorrow and all AI leveraged positions are cleared, the structural pressure on rates won't disappear. The root cause isn't events; it's a paradigm shift.
A Different Perspective: This Is a "Triple Liquidity Squeeze"
Rename the three forces, and you get three sources of liquidity pressure. Different directions, but the timing collided.
Trading liquidity was the first to crack. The AI fund's 4x leverage led to concentrated short-term liquidations, the Philadelphia Semiconductor Index dropped 28.6% in a month, "fall → liquidate → fall more → more liquidations"—forcing other exposed funds to stop losses.
Bond liquidity is a much bigger well. $15 trillion maturing within a year needs refinancing, and FY2026 still requires about $2 trillion in net new issuance. The Treasury is siphoning hundreds of billions from the market monthly, yet buyers are retreating. Yields can only go up; the heavier the interest burden, the larger the deficit, the more new debt—a self-reinforcing loop.
Physical liquidity is the third cut. 27% of the world's seaborne crude oil route is cut off, with no alternative route, and the SPR buffer is at a 40-year low.
These three liquidity pressures overlapped in time, creating a rare "triple liquidity squeeze" resonance. The extreme moves in July—the Nasdaq's six-day losing streak followed by a 2.8% single-day rebound, the 30-year yield hitting a 19-year high, Brent falling 14% in two weeks then recovering half in three days—are precisely the price expression of this resonance.
This framework is not a forecasting tool. It explains why July was so violent but can't tell you what August will be like. The three pressures have different easing conditions: AI positions depend on when Citadel distributes them to the market, Treasuries depend on the Treasury's issuance schedule, Hormuz depends on the negotiation rhythm. They aren't synchronized. But their resonance is enough to explain one thing: why the volatility this summer feels more acute than any single indicator suggested.
After This Summer: High Volatility as a Rehearsal for a New Normal
On the U.S. stock side, Q2 earnings have sufficiently validated the AI demand side—cloud provider growth rates collectively hit new highs, order backlogs broke records. But July's implosion proved another side: in a crowded, highly-levered sector, there's a huge gap between "being right" and "surviving to see the payoff." JPMorgan's warning is especially worth remembering—with institutional absorption capacity diminished, only retail investors are left to hold the bottom, making the next correction even more devoid of buyers than July's.
On the Treasury side, Goldman's statement is worth repeating: the Fed is a "passenger," not a "driver." If the fiscal path doesn't change, the structural pressure on long-end yields won't disappear with a few rate cuts. Whether the 10-year can fall back below 4% depends not on the next FOMC meeting, but on whether global savings are willing to flow back into Treasuries—that is beyond the Fed's control.
On the oil side, the risk premium for Hormuz has shifted from a "one-time shock" to a "persistent discount." Even if a temporary ceasefire is reached, the huge gap between the OVX and VIX won't disappear soon. With the SPR at a 40-year low—the next time there's a supply-side shock, markets will be more fragile and governments less capable. Goldman's base case is "both sides seem to lack the will for large-scale escalation, so energy prices should remain bounded," but this assessment itself is fragile—if the market moves too far in the opposite direction, "worries can easily re-emerge."
The true legacy of the summer of 2026 is not the price movement of a single asset. It's that the market has finally started paying a premium for the new paradigm of "capital scarcity." The three forces are merely three facets of a grand narrative; they collided this summer, but the underlying structural forces are far from fully unleashed.
