Cross-Asset Quarterly Outlook
December 2025*
Beyond the Noise
- Markets have been ‘choppy’, but AI-related growth and a dovish Fed stance should support risk assets going into 2026.
- We see further upside for global equities over fixed income, particularly in non-US markets as USD weakness re-emerges.
- We maintain an equity overweight, stay constructive on duration, and hold an underweight in high-yield credit. We remain overweight commodities while reallocating part of our precious metals position into industrial metals.
The past few months could best be described as a period of ‘choppy’, range-bound trading. Chart 1 shows that most asset classes moved higher over the period; however, some areas of the market experienced elevated volatility – notably AI-related stocks and credit, cryptocurrencies, and Japanese government bonds. Over the same period, the VIX index spiked above 25 twice, in October and November.
Some degree of market anxiety is understandable, at least in the short term. Rich valuations in technology stocks have fuelled concerns of an “AI bubble.” Meanwhile, a court ruling on IEEPA tariffs remains pending, and key economic data releases were delayed by the US government shutdown in October and November. A data-dependent Fed has added to uncertainty, yo-yoing its guidance between pausing and cutting with a significant number of dissents.
Taking a step back from the recent noise, our views are largely unchanged from three months ago. Our base case remains that the Fed will maintain a dovish bias heading into 2026 – an environment that has historically favoured risk assets, provided a recession is avoided. The current K-shaped economic backdrop has created a quasi-Goldilocks scenario: pockets of economic weakness (particularly the labour market) have pushed the Fed toward a more dovish stance this year, while AI-related sectors continue to expand at a robust pace (chart 2).
Overall, we believe the AI theme remains durable, and investors should maintain exposure. A wall of worry may persist as hyperscaler capex spending continues to rise (chart 3). However, free cash flows (FCFs) remain elevated and stable (chart 4), suggesting bubble concerns may be premature – at least until debt becomes the primary funding source for AI investment. As with previous technology cycles, there will be winners and losers over time, and bottom-up investors should remain vigilant regarding froth in individual names. From a macro perspective, however, identifying and timing bubbles is notoriously difficult. One must assess when the aggregate market value of AI-linked firms exceeds realistic future cash flows and identify potential catalysts once the conditions of a bubble are met (for example, Fed hikes or an earnings recession). Our base case is the AI economy and future expectations can continue to grow for some time. Major banks and consulting firms will likely continue publishing widely divergent forecasts for the “AI economy” – often with trillion-dollar differences in size estimates. We therefore favour maintaining exposure to strategic AI themes, while ensuring adequate diversification.
Unlike AI-linked equities, the US dollar trading has been less eventful in recent months, with its prior downtrend stalling. Should US rates continue to decline into 2026, the dollar could reach new cyclical lows (chart 5). Additional Fed cuts would likely reinvigorate USD bears, supporting non-US indices relative to the US and bolstering both precious and industrial metals.
Chart 1: Sep-Nov Performance (%)
Source: Bloomberg
*The publication reflects asset performance up to November 28, 2025, and macro events and data releases up to December 4, 2025, unless indicated otherwise.
The information contained herein is obtained from sources believed by City of London Investment Management Company Limited to be accurate and reliable. No responsibility can be accepted under any circumstances for errors of fact or omission. Any forward looking statements or forecasts are based on assumptions and actual results may vary from any such statements or forecasts.