AI crashed, France exploded, Goldman Sachs recommends a "hedging strategy": pay attention to the Japanese yen!
Goldman Sachs FICC team weaves together three main threads—AI valuation collapse, French bond crisis, and undervaluation of the yen—into a single macro narrative: the US exceptionalism is cracking, and capital is returning. Their core conclusion is that the yen is undervalued by 40%, while the two-year implied volatility is only 9.9, far below the 20-year average, which means buying protection for the currency with the largest valuation divergence at near historical lows. As the correlation chain of AI decline, lower US Treasury yields, and a softer USD/JPY exchange rate has taken shape, holding yen has become the cheapest diversification tool available at present.
The Goldman Sachs FICC team has linked three seemingly unrelated market threads—AI valuation plunge, French bond crisis, and the undervaluation of the yen—into a single narrative, recommending to institutional clients that going long on the yen currently offers the most cost-effective hedging tool.
According to the latest report by Goldman Sachs FICC trader Jonathon Chryssafis, disappointing earnings from AI giants have reignited concerns about the sustainability of this technology trend. Tesla’s capital expenditure cuts, Meta’s AI spending pushing its credit default swaps (CDS) to historic highs, and Lam Research’s disappointing guidance have combined to accelerate hedge funds’ exit from the tech sector.
Meanwhile, the spread between French and German government bonds has widened to its greatest level since the inception of the euro, with the yield on 10-year French government bonds breaking 5%—surpassing the absolute level seen during the Eurozone debt crisis. Goldman Sachs believes that all these risks ultimately point in the same direction: diminishing appeal of U.S. dollar assets, capital flowing back to Japan, consequently driving yen appreciation. Goldman Sachs’ long-term fair value model shows that the yen is undervalued by about 40% against the U.S. dollar, making it the most undervalued among the tracked currencies.
AI Credit Cracks: Data Center Bonds at the Eye of the Storm
Chryssafis emphasizes in the report that the true epicenter of the AI valuation plunge is not the equity market, but the credit market—especially the bonds of lower-rated borrowers.
Data shows that, as of October 13, 2026, dollar-denominated project finance bond issuances have reached about $105 billion, almost entirely flowing into data center and chip/computing power financing projects. This scale is about four times the highest pre-AI record set in 2020 (about $25 billion) and nearly three times the previous record set in 2025.
The Goldman Sachs credit strategy team maintains an underweight rating on U.S. dollar-denominated investment-grade technology bonds, pointing out in their report “AI, Credit, and Higher Capital Costs” that, amid a backdrop of “higher for longer” rates and potential energy cost pressures, the market is reassessing the economic logic of AI infrastructure. In the high-yield bond segment, Goldman remains neutral but expects a continued divergence among AI-related issuers—“especially in the project finance bond sector.” Bernstein currently estimates a $76 billion funding gap for CoreWeave, and this so-called “high divergence” may mean some bonds will get paid while others will not.
French Bonds “Off the Rails”: Pressure on the Euro
In Europe, the French bond crisis continues to escalate. The France-Germany 10-year yield spread has widened to 173 basis points—not only surpassing levels seen during the 2011 Eurozone crisis but setting a historic record since the euro’s birth. The yield on 10-year French government bonds reached 5.08%, while Italian bond yields are now about 50 basis points lower than France’s—a reversal that was almost unimaginable during the 2011 crisis.
Goldman Sachs strategists believe that declining energy prices are the clearest path to narrowing the spread, because: firstly, this would improve the risk bias for growth and fiscal deficits; secondly, it would help the market return to the ECB’s limited rate-hike baseline, bringing some core yield relief; and thirdly, it would lower rate volatility, increasing investor willingness to return to arbitrage trades.
Chryssafis states, if such energy price relief does not materialize, “we expect to see even wider spreads and increased pressure on the euro’s exchange rate.” Goldman’s trading desk has recently observed rising client demand for euro/USD downside options, though implied volatility remains low by historical standards. From a valuation perspective, the five-year euro/USD real yield spread has narrowed to -136 basis points, comparable to where the euro traded near 1.03 in early 2025, while the current rate remains near 1.12—suggesting further downside for the euro.
Yen: The Most Undervalued Hedge, Priced 40% Below Fair Value
The most notable part of the report is Goldman Sachs directly tying AI trades to the direction of the yen.
Goldman’s GSDEER long-term fair value model shows the yen is undervalued by about 40% against the dollar—its largest divergence among tracked currencies. The logic: For decades, Japan has deployed its excess savings into U.S. financial assets; once the “U.S. exceptionalism” story driven by AI falters and the U.S.-Japan yield gap narrows, capital will flow back to Japan, resulting in dollar selling and yen buying.
The Goldman trading desk points out that three structural factors naturally support a stronger yen: Japanese officials continue to express support for a strong yen; Prime Minister Takaichi is pushing aggressively for greater domestic allocation to local financial assets; the Bank of Japan maintains a gradual rate hike trajectory, and Japan’s largest labor union is reportedly seeking a 6% increase as the median target for 2027 spring wage talks (according to Reuters), giving the BOJ ample reason to keep raising rates.
More importantly, the cost to hedge is extremely low. The two-year USD/JPY at-the-money option implied volatility is currently 9.9, much lower than the 20-year average of 13.8, and near the 20-year low of 8.4. Goldman believes that buying protection on the most undervalued currency in its coverage list at near historic lows offers outstanding value.
Correlation Reversal: AI Decline Equals Yen Strengthening
Since October 14 (when Goldman’s AI leaders’ basket fell 4.7% in a single day), the correlation between this basket and the S&P 500 jumped from 0.73 to 0.91; correlation with changes in the 10-year U.S. Treasury yield flipped from -0.13 to +0.19; and the correlation with USD/JPY flipped from -0.34 to +0.46.
This means that the current market logic chain is now clear: AI declines → S&P declines → U.S. Treasury yields drop → yen appreciates. Goldman’s trading desk says it continues to favor “equity downside/rates downside” twin binary options as hedges against growth shocks and specifically notes that correlations have reverted to the mode of “AI declines, 10-year yields fall, USD/JPY weakens.”
The only potential break in this logic would be a tail risk Goldman itself highlights—Brent crude surging above $140. In such a case, yields would rise with oil prices, equities would also fall, and the twin-hedge strategy would fail. But until then, with S&P 500’s correlation to the standalone AI basket now at 0.91, “diversification” has a new meaning: holding yen.
Disclaimer: The content of this article solely reflects the author's opinion and does not represent the platform in any capacity. This article is not intended to serve as a reference for making investment decisions.
You may also like
October Non-farm Payrolls Reignite December Rate Hike Expectations! Wall Street Bets Probability Over 80%, "New Fed News Agency": AI Sell-off May Become a Policy Variable
After the non-farm payroll report, traders' bets on a December rate hike surged to as high as 85%. Wall Street analysts believe the report provides room for further rate hikes, bringing renewed pressure for a hike at a time when post-election market concerns over slowing AI development might dampen expectations. Some remain cautious about a December hike, as stock market volatility and AI investment prospects could restrain the Federal Reserve. "New Fed News Agency" notes the report has not shaken the Fed's outlook on economic growth and inflation, but whether AI asset sell-offs change financial conditions could become a new variable for the December decision. Attention should be paid to whether Waller values lagging employment data or forward-looking market signals.
U.S. non-farm payrolls surge by 165,000 in October, far exceeding expectations! Unemployment rate unexpectedly drops to 4.1%, December rate hike expectations intensify
In October, US non-farm payrolls increased by 165,000, which is 2.5 times the Wall Street median forecast of 65,000 and represents an exceptionally strong deviation of about 5 standard deviations, exceeding the upper limit of all institutional forecast ranges. The unemployment rate unexpectedly fell from 4.2% to 4.1%, while the market had expected it to remain unchanged; average hourly wages also accelerated year-on-year growth to 3.1%. Meanwhile, the previous two months' non-farm payrolls were revised upwards by a total of 84,000. Following the release of the data, the yield on 2-year US Treasury bonds rose by 7 basis points, and the pricing of the probability of a Fed rate hike in December climbed to 70%.
AI Sell-off Hits Anthropic IPO, Potential Valuation Expected to Drop to $1.2–1.3 Trillion
The valuation range discussed by potential investors has been lowered from the previous high of $2 trillion to between $1.2 trillion and $1.3 trillion, and some strategists have even suggested going below $1 trillion to ensure safety. In the face of significant market volatility and financial concerns such as quarterly losses, Anthropic is facing tough decisions, including a discounted valuation, reducing the size of the offering, or postponing the listing.
Crypto prices
More
