UBS Equities-Global Equity Strategy _How much of a bubble are we in What...-118531224

ab29 October 2025Global ResearchGlobal Equity StrategyHow much of a bubble are we in? What are the warning signals of a peak?We have long argued that there are 7 preconditions for a bubble and if the Fed cut in line with the UBS forecast, then we have all 7. The rationale for a bubble in our mind is much clearer than it was in the previous bubbles I have witnessed (Japan in late 80s, TMT, Red Chip to name but a few). If Gen AI is perceived to increase productivity from 2028 as much as TMT temporarily did (c2%), then it is easy to justify c20-25% upside- the speed of adoption makes Gen AI unique. The additional justification this time is that government balance sheets, relative to their norm, are much risker relative to corporate/bank balance sheets (in the TMT period, the US government was running a budget surplus) and there is at least a chance in our opinion that governments eventually print to resolve the issue (leading to a switch from nominal into real assets). This would be accelerated if the Fed was perceived to lose its independence. What marks the peak of a bubble:1) Very clear overvaluation. There are 5 different approaches to this: i) The P/E rises to c45X to 73X for at least 30% of market cap on a 10 year bond yield of, at least, 5.5% in previous bubbles. Currently, the Mag 6 trade on 35X 12 month trailing; ii) The ERP falls to c1% (2000, 1929); iii) Investors move away from conventional valuations to unconventional (land value in Japan in 1989 or eyeballs in TMT) or just focus relative value; iv) TAMs (Total Addressable Market) required to justify valuations become very unrealistic. At the peak of TMT, around a fifth of household income needed to be spent on telephony to justify valuations. We calculate if by 2030 1.3% of GDP is spent on semis, then valuations are reasonable (semis and software are the new oil and today amount to c3% of GDP, oil averaged 3% of GDP but peaked at 10%); v) the overvaluation is so acute when the bubble burst, investors lose 80% (very hard to see being the case currently).2) Long term catalysts of peak: i) clear cut over-investment; ICT investment as % of GDP is well below 2000 levels and close to normal levels; ii) excessive debt financed spending (currently the top 11 hyperscalers in aggregate can increase capex by 40% before they fund out of debt using 2025 revenue numbers)- leverage of tech today is much better than that of tech or telecoms during the TMT bubble despite Open AI's ‘vendor financing’; iii) much larger loss of breadth (in 1999 nearly twice as many stocks fell as rose despite the Nasdaq nearly doubling); iv) overall national account profits come under much more pressure (during TMT national accounts profits fell, unlike now and this explains the narrow breadth point above); v) corrections, the Nasdaq had 5 corrections of more than 10% in 1999, since April there has been none; vi) earnings momentum peaked a year before the market peak; vii) credit spreads troughed 10 months before the