Author: Long Yue,Wall Street View
Two seasoned macro investors sat down together and reached nearly identical conclusions: this AI-driven upward cycle is nearing its end, and the upcoming downturn will not be a single-digit correction, but a major bear market of 30 to 50%.
On June 22, DoubleLine Capital, a U.S. asset management firm, published a deep-dive interview on its blog, in which “New Bond King” Jeffrey Gundlach and Swiss hedge fund manager and “Stock Market Prophet” Felix Zulauf discussed how the world is shifting from unipolar to multipolar, and how geopolitical conflicts and sanctions will drive structural inflation. Against the backdrop of the old order crumbling, both the U.S. tech stock frenzy and the abyssal U.S. fiscal deficit have reached critically dangerous tipping points.
From left to right: Felix Zulauf, host Grant Williams, Jeffrey Gundlach
The AI frenzy is coming to an end, and the U.S. stock market is poised for a 30% to 50% correction.
“This is not a 20% pullback—it’s a bear market driven by economic recession and valuation contraction, with declines ranging between 30% and 50%,” Zulauf stated bluntly, adding that U.S. equities could peak as early as this third quarter or as late as next first quarter.
The logic chain he presented is clear: the capital expenditure as a percentage of revenue for hyperscale cloud companies (super scalers) has surged from 10% to 30%, semiconductor memory chip prices have risen 200%-300%, and free cash flow has turned negative—Oracle is already negative, and others will follow. “When these companies start raising capital in the market and their free cash flow begins to shrink, the entire AI cycle will begin to slow down.”
To precisely exit at the top, you must closely monitor the price movements of semiconductor stocks that are “selling shovels to gold miners.”
Gundlach fully agrees. In the current S&P 500, the top ten AI-related stocks account for a staggering 41% of the index weight—a level of extreme concentration that closely mirrors historical market cycle peaks.
"I recommend that people not hold any momentum-driven or market-cap-weighted U.S. stocks," Gundlach offers a direct hedging strategy.
He also mentioned his "famous misjudgment" on September 30, 1999—when he turned maximally bearish on the Nasdaq, only for the index to rise another 80% in the fourth quarter. "But 18 months later, from that point, the Nasdaq dropped from 100 to around 20. So the most dangerous moment is when fundamentals are deteriorating but prices are still rising. We are right there now."
Recession is coming, and U.S. bond yields won't fall—U.S.-style YCC and a "major U.S. debt restructuring" are unavoidable.
This is one of Gundlach’s most core judgments and also his greatest point of divergence from traditional economic logic.
The usual logic is: recession → Fed rate cuts → long-term rates decline → bond prices rise. But Gundlach believes this time is different. Even if the U.S. economy enters a recession in 2027, U.S. long-term Treasury yields will not experience a meaningful decline.
The reason is that fiscal issues have reached a level of structural失控: U.S. interest expenditures have surged from approximately $300 billion seven years ago to nearly $1.4 trillion per year today. Meanwhile, the fiscal deficit is expanding at a rate of $2 trillion per year, accounting for about 6% of GDP.
“When a recession hits, the deficit won’t be 6% of GDP—it’ll be 10% or even higher. That will trigger a bond buyer strike,” he said. “We’ve already seen this in developed countries—even Japan’s long-term interest rates are rising, something many thought would never happen.”
Gundlach believes that policy responses at that time will take two directions:
Option A: Yield Curve Control (YCC). Minister Besant may choose to suppress long-term interest rates, as the U.S. did after World War II—allowing inflation to rise while artificially keeping long-term rates low, resulting in sustained negative real interest rates and a 40-year bond bear market.
Option B: U.S. debt restructuring. Guggenheim revealed that two years ago, he reduced the coupon rate of U.S. Treasuries with maturities over 10 years in his managed fund from 4.75% to 1.5% to hedge against restructuring risk. After he publicly discussed this idea in an interview last year, media inquiries reached Kevin Hassett, Director of the White House National Economic Council, who responded, “It’s absolutely not going to happen.”
Gundlach's response was: "In the investment world, the synonym for 'Never' is 'Imminent'."
Zulauf has a slight disagreement on long-term interest rates: he believes that during a recession, the 10-year Treasury yield could still decline from its high of around 5.25% to approximately 3.75%—but this window would last only about six months, not a full 12 months. He added that short-term rates will be kept very low by central banks.
Private credit crisis: "It feels like 2006 right now," "Everyone is lying"
Private credit, hidden beneath the surface compared to public markets, has sparked greater concern, rife with rating manipulation, illusions of liquidity, and accounting maneuvers to conceal losses.
Gundlach said:
It gave me a strong feeling, exactly like the one I had in 2005 and 2006: everyone was lying—lying about credit quality, lying about software exposure—they said it was 15%, but it was actually 28%—creating a completely illusory liquidity that has now shattered.
Ratings are bought. “These private rating agencies have only 30 employees, yet they’re rating hundreds of loans, each with 200 to 250 pages of documentation. I don’t think they’re actually analyzing anything—I think they’re selling price lists. Want a CCC rating? That’ll be $1. Want a single B? That’s $10. In the end, everyone ended up with a BBB-.”
Credit quality has been severely overstated. A major private credit fund claimed in its marketing materials that “investment-grade corporate bonds are the backbone of the portfolio,” but in the private markets, securities rated B+ or higher account for only 2% of all securities. “Less than 2% are above B+, so what are you basing your backbone on?”
Software asset risk is underreported. One fund claimed a software exposure of 15%, while the actual figure is 28%.
The illusion of liquidity has burst. Many investors who purchased interval funds through financial intermediaries believed they could redeem their full investment quarterly, but in reality, the fund-level redemption cap is only 5%.
Valuation markings are inconsistent. Gundlach cited an example where the same loan is held by eight different private equity firms, yet its price ranges from 95 to 8—same asset, one marked at 95, another at 8. Another case: a $100 million PIK bond, despite the underlying private equity having been written down by 98% to $800,000, is still marked at par value of 100.
Offshore reinsurance is the last black box. A closed loop has formed between private equity, private credit, and the insurers they control, with risks transferred to offshore reinsurers in Barbados, the Cayman Islands, Bermuda, and other jurisdictions—without regulation or transparency. “I’m not sure those risks have actually been hedged. When a downturn hits and fixed annuities and life insurance policies come due, those assets simply don’t have adequate reserves.”
Zulauf added: "All issues will surface when the market turns and the tide goes out."
AI funding chains and private credit are essentially the same line.
AI and private credit may seem like two separate markets—one on the equity side, the other on the credit side—but within this framework, they are connected through the cost of capital.
AI capital expenditures continue to rise, putting downward pressure on free cash flow. As free cash flow declines, companies must either issue equity or take on debt. When taking on debt, if long-term interest rates do not decline, financing costs will not automatically ease as they did in previous cycles.
Lower-rated companies are more problematic. In the past, when the economy weakened, spreads widened, but risk-free rates fell, sometimes offsetting some of the pressure and allowing struggling companies to refinance and survive. Now, if risk-free rates rise instead of fall, the window for refinancing will narrow.
This will directly transmit to bank loans, CCC-rated loans, and private credit. Gundlach noted that these markets have already begun to show signs of strain. The core reason is not that any single industry has suddenly deteriorated, but rather that the model relying on low interest rates and refinancing is no longer functioning smoothly.
So, AI trading isn't just about NVIDIA, cloud providers, or data center orders—it ultimately depends on whether financing markets can continue to provide funding and whether credit markets can withstand higher interest rates.
The dollar is weakening, U.S. stocks are underperforming—the second round has just begun.
Gundlach mentioned a historical pattern: in the previous 13 U.S. stock market crashes, the dollar rose in the first 12, by approximately 8%-10%. However, during the 2025 tariff turmoil, the dollar fell by 8%-10%.
This confirms my assessment—that during this interest rate hiking cycle, the market's reaction function has changed.
He believes that the long-term outperformance of U.S. stocks relative to global markets has ended, and emerging markets are now outperforming the S&P 500. "We're in the second inning, not the eighth or ninth."
Zulauf added a risk point: Over the past 12 months, Asian sovereign funds have purchased large amounts of dollar-denominated assets, but no longer U.S. Treasuries—instead, they’ve bought AI stocks. “Once the market turns, they will sell their stocks and simultaneously sell dollars. This is completely different from holding U.S. Treasuries and will accelerate the dollar’s decline.”