[Financial Recap] Why Has Goldman Sachs Turned Bullish on the Yen? Four Key Reasons as It Trades Around 40% Below Fair Value Against the Dollar

Concerns over the durability of the technology boom have resurfaced following earnings from AI giants, while hedge funds are stepping up their retreat from technology stocks. Tesla has reduced capital expenditure, Meta’s AI spending has driven its credit default swap (CDS) spreads to record highs, and Lam Research has delivered disappointing guidance. At the same time, the gap between French and German 10-year government bond yields has reached its widest since the euro was introduced. France’s 10-year yield has moved above 5%—higher in absolute terms than during the European sovereign debt crisis. In a report dated 23 October, Goldman Sachs FICC trader Jonathon Chryssafis brought these three apparently disconnected market developments together within a single narrative.

A stronger yen is the common outcome Goldman sees emerging from these risks, as demand for US dollar assets weakens and capital flows back to Japan. According to its GSDEER long-term fair value model, the yen trades approximately 40% below fair value against the dollar, the widest valuation gap among the currencies covered by the model. Yet USD/JPY two-year at-the-money implied volatility is only 9.9%, against a 20-year average of 13.8% and a historical trough of 8.4%.

Reason One: Are Credit Markets Revealing the Cracks in AI?

Credit, not equities, is where Goldman locates the real epicentre of a sharp correction in AI valuations, with bonds from lower-rated borrowers particularly exposed. This assessment of where AI-related risk is concentrated is among the report’s most easily missed—yet logically central—arguments.

By 13 October, issuance of US dollar-denominated project finance bonds in 2026 stood at approximately $105 billion. Almost all of it was intended to finance data centres and semiconductor or computing infrastructure. The figure is about four times the pre-AI record of approximately $25 billion established in 2020, and close to three times the 2025 total, itself a record. Within US dollar credit, Goldman’s strategy team remains underweight investment-grade technology bonds and neutral on high-yield debt. It nevertheless explicitly anticipates increasing divergence between AI-related issuers, especially in project finance bonds.

The transmission mechanism is clear. A broad rise in AI infrastructure funding costs, driven by wider credit spreads, would start to erode the foundations of the “US exceptionalism” narrative. For Goldman’s wider yen trade, weaker demand for US dollar assets represents the first link in the chain.

Reason Two: Could French Unrest Redirect Capital Flows?

France’s bond market crisis forms another part of the same Goldman report. At 173 basis points, the French–German 10-year yield differential now exceeds its level during the 2011 European sovereign debt crisis and is the widest recorded since the euro’s introduction. The French 10-year government bond yield has climbed to 5.08%. Italian yields, meanwhile, are approximately 50 basis points below France’s—a reversal that would have been almost inconceivable during the 2011 crisis.

Cheaper energy offers the most direct path to tighter spreads, in the view of Goldman’s strategists. A decline in energy costs could improve the risk balance for growth and fiscal deficits, restore a baseline market expectation of limited European Central Bank rate rises, and dampen interest rate volatility. That, in turn, could draw investors back into carry trades. If energy prices provide no relief, however, Goldman anticipates further spread widening and additional pressure on the euro.

The implication is that risks in France could hasten the movement of European capital into safe havens beyond the United States. With domestic yields recovering and its status as the world’s largest net external creditor, Japan is structurally positioned to receive part of that capital.

Reason Three: How Can Japan’s Structural Capital Repatriation Be Quantified?

Specific estimates of capital flows underpin Goldman’s positive yen outlook; a general account of Japan’s economic improvement does not.

Rebalancing by GPIF deserves more attention than it receives. Goldman highlighted in a September report that Japan’s Government Pension Investment Fund (GPIF) oversees approximately $2 trillion in assets. Raising its allocation to domestic fixed income by five percentage points would translate into roughly $100 billion moving into Japanese assets, with an equivalent amount of USD/JPY selling.

Such a shift would represent nearly half of Japan’s annual current account surplus of approximately $210 billion. It would also far exceed the roughly $35 billion of net current and financial account flows registered in 2025. Other Japanese pension funds, insurers and long-term institutional investors could follow a change in GPIF’s allocation. In turn, this could prompt financial institutions and corporations to unwind yen-funded carry trades built up over the past several years.

Further rate rises also remain supported by wage trends. In August, Japanese real cash earnings increased 1.5% year on year, extending the run of monthly gains to eight—the longest in nearly a decade. Growth in base wages was 3.8%. Continued gains in real wages reinforce the argument for additional Bank of Japan rate increases after its latest move.

While markets have factored in the policy trajectory, options pricing has not captured it. Overnight index swaps in late October assigned a probability of only roughly 6%–10% to a Bank of Japan rate rise in October, compared with an 80%–89% chance of an increase by December.

September’s Bank of Japan increase took the policy rate to 1.25%. As its second rise within three months, the move marked the fastest tightening cycle since 1990. Broad market expectations point to a further increase to 1.50% by year-end or early next year, and then to 1.75% in March or April.

An increase in December is thus a high-probability prospect, not a tail-risk event. Nevertheless, two-year implied volatility at 9.9% describes a subdued setting in which annualised volatility stays below 10%. Goldman’s trade centres on the inconsistency between those signals.

Reason Four: Why Does Goldman See 9.9% Implied Volatility as “Mispricing”?

There are fundamental reasons for subdued implied volatility. Recent USD/JPY trading has been confined to a tight band around 158, and realised volatility has remained below implied volatility. Continued options selling has squeezed the volatility premium.

A feedback loop follows: subdued realised volatility emboldens sellers, whose activity drives implied volatility lower still. For that loop to persist, however, one condition must hold—the market must receive no fresh information. Goldman contends that fresh information is now coming through.

The first point is that the timing of further Bank of Japan tightening increasingly matters more than whether it happens at all. With the probability of a December increase above 80%, markets appear to regard a hike as their base case already. Options valuations, however, have not shown an accompanying increase in volatility.

A Bank of Japan rate rise in December, alongside an unchanged Federal Reserve rate, would narrow the US–Japan interest rate gap and directly reduce the economic appeal of carry trades. A survey finds that 94% of economists anticipate no change from the Fed in October, although every respondent expects a December hike.

The direction of a surprise is more important still. A December move by the Bank of Japan could lead to a sharp repricing of the yen if markets have not fully incorporated the increase.

The second point comes from options skew, which is signalling a risk that the broader volatility level has yet to reflect. Risk reversals continue to show a premium for protection against USD/JPY falling. Options traders are therefore already paying to cover the tail risk of a stronger yen, despite overall implied volatility staying subdued.

High skew alongside low aggregate volatility indicates that markets recognise the direction of the tail risk but have not fully allowed for the possible size of the move. This is the gap Goldman’s recommendation is designed to capture. It favours purchasing convexity while implied volatility is low, rather than depending on correctly timing a directional position.

The third point is the change in Goldman’s own projections. The bank cut its three-, six- and 12-month USD/JPY forecasts on 25 September, replacing 162, 163 and 165 with 158, 155 and 150, respectively.

Moving from a bearish yen position in July to a bullish one in September reflects a reappraisal of Goldman’s macroeconomic framework. Its 23 October report turns that revised assessment into an actionable trade structure with a specified cost.

Taken together, Goldman’s argument for a stronger yen rests on three forces. Strains in AI credit markets are weakening the foundations of demand for US dollar assets, risks in France are speeding up capital reallocation, and structural repatriation into Japan is creating upward pressure on the yen. That repatriation is supported by GPIF rebalancing, the wage–price cycle and gradual tightening by the Bank of Japan.

Despite this combination of risks, the options market’s “insurance premium” remains tied to the low-volatility equilibrium established during the recent spell of range-bound trading. At 9.9%, implied volatility still prices in a world receiving no fresh information.