KMD Economic & Market Update – Autumn 2025
An AI bubble?
Despite the concerns over higher tariffs, inflation and government finances the financial markets have enjoyed another robust quarter since our last update. The equity market rally which began during the summer has continued, the major government bond markets have been stable and the dollar has firmed. Yet again the bull market has climbed the wall of worry.
Continued growth in the US economy and increases in global trade (despite the tariffs) have been important in sustaining investor optimism and this has been reinforced over the past three months by an easing of monetary policy with rate cuts from the US Federal Reserve (Fed) and Bank of England. Overall, the worst fears have not been realised and at their recent meetings the IMF raised their forecasts for global growth.
On monetary policy we see the Fed cutting interest rates further in coming months as the US economy cools and the conditions for lower inflation fall into place. For the UK though the rate call is more difficult as inflation has proved to be more persistent than expected despite weak activity. This is adding to the Chancellor’s woes ahead of the budget where a lack of growth means taxes are almost certain to rise to meet the government’s fiscal rules.
We will not add to the speculation over which taxes will rise, but simply add that with almost every part of the economy within scope, the increase in uncertainty has curtailed decision making and spending in the UK thus making the growth problem worse.
Macro factors are important in supporting markets, but they are not the whole story. The key feature of the recent rally in equity markets has been its dependence on firms linked to the development of Artificial Intelligence (AI) particularly the M7 or Magnificent 7, made up of Microsoft, Amazon, Alphabet, Nvidia, Tesla, Meta and Apple.
As such the S&P500 index rally of nearly 8% over the last three months has been highly concentrated: the M7 / AI-exposed mega-caps (?35% of S&P by weight) explain roughly three-quarters of the gain. These figures are from ChatGPT, a manifestation of AI, but I have no reason to suspect bias in matters relating to performance attribution.
Consequently, equity market strength has been driven by a relatively small set of names, increasing concentration risk to the point where investors are comparing the rise in the market with the tech bubble of the late 1990s. At that time, the top 10 stocks accounted for 26% of the market whereas today they comprise 38%. As we know, that ended in tears, so are we heading for another major collapse in markets?
There is no easy answer, but having lived through the tech bubble I have vivid memories of that period and would highlight a few points.
In contrast to the 1990s when many tech companies had no revenue and were valued on “clicks”, today the AI sector is being driven by real spending and profits as firms invest to capitalize on AI. That spending looks set to continue given proposed investment and will support earnings.
However, this also raises questions about how realistic future projections can be. For example, to facilitate the rise in computing power there will need to be a massive increase in the supply of energy, data centres and semi-conductor chips. Data centres alone already use more electricity than France – a level of consumption which is expected to triple by 2030.
On the chip side there is a scramble to secure access to supplies and the area has already become a battleground between the US and China. Alongside the risk that the demand for AI may not materialise, constraints on the supply of energy and materials may also limit progress. If so, there is significant scope for disappointment given equity valuations, with potential for an adverse impact on the world economy in the event of a crash.
That doesn’t mean there will be a crash, but we could see a compression in valuations as the market becomes more discerning over the winners and losers from AI. Furthermore, as we have seen in previous episodes the winners might well come from the suppliers to the gold rush rather than the AI firms themselves.
The real difficulty for investors though is the length of time that the market can run on before reaching a top. During the late 1990s there were many calls for a bursting of the bubble. Perhaps the most famous from the then chair of the Fed, Alan Greenspan, who spoke of “irrational exuberance” in markets in 1996. Those investors who took that as a cue to avoid tech stocks had to endure four years of underperformance before they were ultimately proved correct. As the economist John Maynard Keynes said, “the market can be irrational for longer than you can remain solvent”.