Financial Bubbles – Are we in one?

AI = maybe <---> Sovereign Debt = definitely

Almost a generation ago, we saw the biggest investment bubble in history go pop, the internet bubble. To this day, that experience has affected a whole generation of investors, particularly Continental European investors who had only just started investing in equities. They quickly got more than just a bloody nose.

As we wrap-up 2025 with the third bull market year in a row and the NASDAQ index up over 130% since its 2022 low, we need to ask if we are again in a technology bubble; specifically in an AI bubble.

The following indicators are helpful in determining whether we are in fact in a bubble:

• Valuation: three long-term measures are at extremes in the US = Shiller CAPE (10-year cyclically adjusted P/E), Buffett indicator (market cap to GDP) and Tobin’s Q-Ratio (market cap to replacement cost). While this is an absolutely necessary condition for a bubble, it never works for timing.

• Sentiment (1): what used to worked very well was the “Frontpage” indicator (Major newspapers calling to invest into stocks on page 1, rather than in the business section) and the “Taxi-driver” indicator (taxi drivers bragging about what stocks they are buying). Nowadays, due to the prevalence of social media and their fast paced nature, both indicators work less well. The rally in meme stocks in January 2021, when retail investors on Reddit and trading through Robinhood actually bankrupted two hedge funds, reveals that the opposite can be true.

• Sentiment (2): CNN’s Fear & Greed Index is a more useful gauge of investors’ sentiment. However, its track record is asymmetrical. It is useful as a contrarian buy signal when fear is high, but less good at calling tops.

• Financial fundamentals: This is the most reliable indicator. When over-investment happens and companies cannot deliver an adequate return on their investment dollars, stock prices fall. The challenge of forecasting future earnings becomes even more difficult in very dynamic times such as the race for AI now.

Thesis - We DO have an AI bubble:

• The amount of AI capex by Big Tech in 2025 is estimated at $400bn and forecast to reach $7 trillion cumulatively by 2030.

• In Big Tech it seems everybody wants to do everything: chip manufacturers are going into infrastructure and software, and vice versa, software companies are producing chips; social network providers, e-commerce platforms and search engines are doing both. Too much competition in too many areas diminishes the planned rate of return.

• Circular investment: in 2025, competitors have increasingly started to invest into each other.

• These latter 2 points especially reminds us of the internet bubble.

• China’s strategic investment is not entirely profit driven, but part of global power politics (also see Ray Dalio "The changing world order"). This in turn forces American hyperscalers (n.b. the creation of this word could in itself be an indicator of a bubble) to invest with little regard for future returns.

Antithesis - We DON’T have an AI bubble:

• Unlike the year 2000, Big Tech companies investing into AI are hugely profitable and can sustain potential poor returns for long time periods.

• AI implementation is already leading to productivity gains. Chat-bots were only the beginning. Now, many mundane business tasks can be reliably executed by AI agents. While it is difficult to foresee in which areas these gains will materialize, jumps in productivity have always lead to wider wealth creation.

• Big potential exists in areas that have seen unprecedented growth in administration (health care, education, law/regulations).

• On top of this, additional technologies are breaking through: space, robotics, quantum. Particularly, AI and robotics feed off each other and will accelerate productivity.

Synthesis - In conclusion, while I cannot exclude it, I currently do not believe this is an AI bubble. Without a doubt we are facing disruptive times and high unemployment will initially follow, but productivity gains are huge wealth creators and the chances are good that we will experience a new level of broad economic growth into the 2030s. Also see: “Zeitenwende” – the turning of an era

Sovereign Debt

When the Euro was introduced, the participating states agreed that debt over 60% of GDP was too high and needed to be sanctioned. In the USA, UK, France, we are at double(!) that level and Germany would also be that high if it included provisions for liabilities as companies are complied to do. Japan is world champion with 4x that level.

On one hand side, one could argue that Western democracies have room to expand to Japanese levels. On the other hand side, the current levels exceed anything we have seen before, even in WW2.

Superficially, MMT seemed to work as large QE post-GFC and Covid did not cause run-away inflation. Nonetheless, a closer analysis of inflation statistics versus affordability reveal large gaps. Housing, education and health care affordability have declined in both the US and in most of Europe.

Solutions for investors

Five years ago, we wrote about the consequences of endless money printing and how we invest. This is still valid: The CUB and the Hunter

Some form of financial reset seems inevitable, but it is close to impossible to time due to the nature of such a political decision. Classic hedges of currency debasement are hard assets such as precious metals including digital gold (BTC) and productive assets such as agriculture or such equities that are largely immune to inflation.

In terms of AI, we prefer to ride the trend but are switching to areas where competition is less intense and a winner-take-all scenario is less likely. There is one financial indicator we used with good success in the internet bubble. We closely watch revenue and earnings momentum, i.e. declining growth %. Generally, double-digit % growth can be considered as above average. However, when fast growing tech companies decelerate from 50% to 30% and then 15%, it usually has dramatic consequences for the share price due to multiple compression.

We have entered the phase of the market to select the winners of the AI race. There are many to choose from and they are not only in technology. Happy picking!

The HanseaticHunter

#AI bubble #debt #crackupboom #zeitenwende