Monday Digest

Posted 17 November 2025

Open and shut
Stock markets rallied early last week and then sold off – strangely, just after Trump signed a bill to open the US government. And, just like the previous week, in the last hours of Friday trading, equities rallied.

First, the UK: the Chancellor has reportedly ditched the simple income tax-raising plans, which isn’t good as it means a more complicated system with unintended impacts. We also had disappointing GDP data, though this largely came from Jaguar Land Rover’s factory closures after August’s cyberattack. Services reported stronger than expected. Bond markets reacted somewhat to the budget shifts although are still relatively sanguine. 

The post-shutdown selloff is classic ‘buy the rumour; sell the fact’, and it was notable that US stocks outside tech (more affected by shutdown) had a relatively better week. The resumption of payments out of the US Treasury General Account (TGA) will relieve a liquidity constraint on markets, though it will take time to come through. There wasn’t a big downturn anyway (estimates say an $11bn GDP loss), and the AAII’s bull-bear index suggests investors have turned neutral. Unfortunately, the lack of October data means markets will continue flying blind. We will have to keep an eye on the US consumer through a crucial spending season. 

Liquidity flowing from the TGA should mean less volatility, but that doesn’t mean stocks going up – just like recently high volatility didn’t mean stocks went down. The TGA isn’t the only liquidity factor either; AI capex is also taking money out of the financial system to build infrastructure. Oracle’s sharp rise in its credit spread this week shows that, as the computing firm’s demand for capital forces investors to reassess the credit (and consequently equity) valuations. 

That’s what happens when businesses invest rather than distribute profits. It can be scary for investors, but it’s ultimately good for long-term growth. If we end 2025 merely at the levels reached at the end of October, this would mean a third strong year for portfolio investors. 

Back to the 90s, but what year is it?
If AI is another dotcom bubble and we’re back in the 1990s, what year of the rally is it? Absolute Strategy Research overlayed current trends to the late 90s and found that US stocks are in a similar position to early 1999 – about 16 months before the bubble finally began to deflate in September 2000. We should take simple comparisons like this with a pinch of salt; profits and valuations for the current tech leaders look much healthier than in the 90s, in part because generative AI is prohibitively expensive. But the similarities are undeniable. The last leg of the dotcom rally was significantly more volatile as companies increased capex. We’re seeing record tech capex and bumpier markets now too. 

Tech companies aren’t the only ones benefitting from the AI theme. European energy stocks have picked up thanks to plans to raise capital and build infrastructure. Investors normally wouldn’t like utilities companies spending big, but many explicitly reference the need to keep up with the energy demands of AI – showing that AI investment is spreading to other areas too. In the 90s, telecoms companies also benefitted, as new internet companies required greater communications infrastructure. 

Back then, dotcom startups were concentrated in the US, while telecoms were prominent in Europe. Now, AI companies are overwhelmingly American, while second-round energy beneficiaries are more European (also related to Europe’s drive to build energy independence). 

AI capex going into background infrastructure is good, and should multiply growth. But strong capex can mean lower liquidity in markets (money spent on datacentres is money not with shareholders now). That can make markets volatile, as they have been recently and as they were in the late 90s. Most expect the AI rally to continue for a while, even if the ride is bumpy. And if there is a pullback later, the current investment spend will be good for long-term growth. 

The Precautionary Tale of Tony Dye
If we’re going to take the AI-dotcom analogy seriously, we should take lessons from the 90s dotcom doubters – like the infamous late fund manager Tony Dye. Dye managed UBS’ UK funds Phillips and Drew and argued as early as 1995 that equities (particularly US tech) were overvalued. Phillips and Drew underperformed for years as the bubble kept inflating and UBS eventually fired Dye – a week before the Nasdaq peaked. The standard narrative is that Dye was proved right, only too late. His son Jon Dye, now research director at Ruffer, recently wrote to investors about Tony’s commendable detail-oriented approach and not getting carried away with optimistic buzzwords. 

However, we don’t think Dye’s badly timed dismissal means he was right to be bearish for so long. In the three-year bear market after dotcom, the S&P 500 never reached its 1995 lows – so you were better off riding the bubble than going bearish too early. Many thought that tech was overvalued in the late 90s but knowing what to do with that information is a different skill. The dotcom bubble might have burst earlier if central banks hadn’t soothed crises in 1997 and 1998, but these external factors always affect the timeline. 

We’re not saying there is an AI bubble (big tech is profitable, but there are areas of concern). The point is that even if there is, the rally could run for another two years, at which point investors would likely be better off riding whatever downturn comes than jumping ship too early. Bubbles burst when liquidity runs out and, while liquidity has tightened recently, upcoming interest rate cuts weaken the bearish case for tech. Don’t get too excited but don’t get too gloomy either.  As ever, it’s about time in the market, not timing the market.