
France’s 10-year government bond yield is above 5%. In absolute terms, that exceeds levels recorded during the European sovereign debt crisis, while the spread over German 10-year yields is at its widest since the euro’s introduction. This is one of three apparently unrelated developments linked by Goldman Sachs FICC trader Jonathon Chryssafis in a report dated 23 October. The others centre on technology. Earnings announcements from AI giants have renewed questions about the boom’s sustainability: Tesla has scaled back capital expenditure, Meta’s AI spending has lifted its credit default swap (CDS) spreads to record highs, and Lam Research’s guidance has fallen short of expectations. Hedge funds are also speeding up their exit from technology stocks.

Approximately 40% is the yen’s undervaluation against the dollar under Goldman’s GSDEER long-term fair value model, the largest gap among the currencies it monitors. Goldman links the risks above to a shared outcome: less demand for US dollar assets, capital repatriation to Japan and yen strength. Options pricing presents a different picture. USD/JPY two-year at-the-money implied volatility is only 9.9%, versus a 20-year average of 13.8% and a historical minimum of 8.4%.
Where would a sharp correction in AI valuations hit hardest? Goldman identifies credit, especially bonds from lower-rated borrowers, rather than equities. This distinction is critical to the report’s reasoning, yet is also among its easiest points to miss.
By 13 October, approximately $105 billion in US dollar-denominated project finance bonds had been issued in 2026. Almost all of the funding was for data centres and semiconductor or computing infrastructure. Issuance is roughly four times the pre-AI record of approximately $25 billion set in 2020, and nearly three times the record total achieved in 2025. Against that backdrop, Goldman’s credit strategy team remains underweight US dollar investment-grade technology bonds and neutral on high-yield debt. Its explicit expectation, however, is for an increasingly wide performance gap between AI-related issuers, particularly in project finance bonds.
Wider credit spreads would weaken the foundations of the “US exceptionalism” narrative if they pushed AI infrastructure financing costs higher across the board. The transmission is direct: reduced appetite for US dollar assets starts the sequence behind Goldman’s broader yen trade.
At 173 basis points, the French–German 10-year yield spread is wider than during the 2011 European sovereign debt crisis and at any time since the euro’s introduction. Goldman examines this French bond market crisis in the same report. France’s 10-year government bond yield stands at 5.08%. Italian yields are now approximately 50 basis points below France’s, a reversal that would have been almost unthinkable during the 2011 crisis.
Energy prices are the key potential source of relief in Goldman’s assessment. Its strategists see a fall as the clearest way to bring spreads down. Cheaper energy could improve the balance of risks to growth and fiscal deficits, restore limited European Central Bank rate increases as the market’s baseline expectation and lower interest rate volatility. That could encourage a return to carry trades. Without lower energy prices, Goldman expects the opposite: wider spreads and more pressure on the euro.
The capital-flow implication extends beyond France. Its risks could speed up European outflows into safe havens outside the United States. Japan’s recovering domestic yields and position as the world’s largest net external creditor make it structurally well placed to capture some of that capital.
Estimates of identifiable capital flows anchor Goldman’s bullish yen position. The argument does not depend on a broad, undefined improvement in Japan’s economy.
The role of GPIF portfolio rebalancing is underappreciated. Goldman’s September report put assets managed by Japan’s Government Pension Investment Fund (GPIF) at approximately $2 trillion. Increasing its domestic fixed-income weighting by five percentage points would direct roughly $100 billion into Japanese assets and generate USD/JPY selling of a corresponding size.
Measured against Japan’s external flows, that would be nearly half its approximately $210 billion annual current account surplus. It would also comfortably exceed the roughly $35 billion in net current and financial account flows recorded in 2025. GPIF’s allocation decision could become a template for other Japanese pension funds, insurers and long-term institutional investors. Financial institutions and corporations could consequently unwind yen-funded carry trades accumulated over the past several years.
The earnings backdrop also favours additional rate increases. A 1.5% year-on-year gain in Japanese real cash earnings in August marked the eighth consecutive monthly rise, the longest run in nearly a decade. Base wages advanced 3.8%. Sustained increases in real wages add weight to the case for further Bank of Japan tightening after its latest rate rise.
A gap remains between the policy path embedded in markets and its treatment in options prices. Late-October overnight index swaps implied only roughly a 6%–10% chance of a Bank of Japan increase in October. For a hike by December, the probability was 80%–89%.
September saw the Bank of Japan lift its policy rate to 1.25%, completing its second increase in three months. This was its fastest tightening cycle since 1990. Market expectations broadly envisage 1.50% by year-end or early next year, then 1.75% in March or April.
December tightening is therefore a likely outcome rather than a tail-risk scenario. But at 9.9%, two-year implied volatility still describes a calm environment with annualised volatility below 10%. The inconsistency is central to Goldman’s trade.
Recent trading helps explain why implied volatility is low. USD/JPY has held to a narrow band around 158, while realised volatility has undershot implied volatility. Ongoing options selling has compressed the volatility premium further.
The process reinforces itself. Low realised volatility increases sellers’ confidence, allowing their activity to push implied volatility down again. It depends, though, on the absence of fresh market information. Goldman argues that new information is now arriving.
First, the timing of further Bank of Japan tightening is increasingly the issue, rather than whether it will occur. An increase in December appears to be the market’s baseline, with its probability above 80%. Options prices have yet to register a matching rise in volatility.
Consider a December Bank of Japan increase with the Federal Reserve on hold: a narrower US–Japan interest rate differential would directly undermine the economics of carry trades. Survey results show that 94% of economists expect the Fed to keep rates unchanged in October, although all respondents expect a December hike.
The direction of a surprise is even more important. Markets that have not fully discounted a December Bank of Japan move could face a sharp yen repricing if the increase takes place.
Second, a warning is visible in options skew but not yet in the overall volatility level. Risk reversals continue to price protection against a USD/JPY decline at a premium. Traders are thus already paying to insure against the tail risk of yen appreciation, while aggregate implied volatility remains low.
High skew paired with low overall volatility implies recognition of the tail risk’s direction, but incomplete pricing of the possible size of the move. This is the gap Goldman’s recommendation seeks to capture. It favours purchasing convexity at low implied volatility over taking a directional position that depends on precise timing.
Third, revisions to Goldman’s forecasts are already in place. The bank cut its three-, six- and 12-month USD/JPY projections on 25 September: the respective forecasts moved from 162, 163 and 165 to 158, 155 and 150.
Moving from bearish on the yen in July to bullish in September points to a reassessment of Goldman’s macroeconomic framework. Its 23 October report turns that revised assessment into an actionable trade structure carrying a defined cost.
Taken together, the bank’s yen appreciation thesis has three drivers. AI credit strains are eroding the foundations of demand for US dollar assets. French risks are speeding up capital reallocation. Meanwhile, structural repatriation to Japan is creating appreciation pressure, backed by GPIF rebalancing, the wage–price cycle and gradual Bank of Japan tightening.
Options markets still charge an “insurance premium” tied to the low-volatility equilibrium established during the recent spell of narrow trading ranges. For this combination of risks, 9.9% implied volatility continues to price an environment without new information.