IC No. 44 - Human intelligence
In Investment Commentary No. 44 – Human Intelligence, we examine artificial intelligence, a topic that is shaping capital markets and society. We provide an overview of the market situation, the energy and infrastructure sectors and derive options for investors.
On 13 October 2025, Joel Mokyr, Philippe Aghion and Peter Howitt were awarded the Nobel Prize in Economic Sciences for their work on growth through technology and disruption.[1] New products and services create a self-reinforcing cycle of economic growth with rising living standards, a healthier society and a higher quality of life. Old business models are displaced by new ones. Misallocations and overinvestment are an inevitable part of this process. There are losers, but overall, the effects are positive for society. The process is unavoidable, and any alternative would block progress and lead to stagnation. Joseph Schumpeter made this essential developmental process famous under the term «creative destruction»[2].
This process of steady progress has gained new momentum with the development of artificial intelligence (AI). We are still at the beginning of this development. What is certain is that it cannot be stopped. What is unclear is whether humanity will ultimately benefit from it or not. Opinions diverge widely.
Apocalypse
Yudkowssky and Soares, the authors of the book «If Anyone Builds It, Everyone Dies», sketch a scenario in which, after the emergence of artificial general intelligence (AGI), humanity would, with high probability, have only a few days left to live.[3] The term AGI refers to the hypothetical intelligence of a machine that is capable of understanding or learning any intellectual task a human can perform and of improving itself autonomously. It would thus be able to solve problems for which it was never trained.[4] Such a so-called superintelligence would no longer need humanity and, in order to optimise itself and secure its survival, could wipe us out.
A new dawn
At the other end stands Yann LeCun, Chief AI Scientist at Meta, who foresees major advances in productivity, health and science, and sees in AGI the potential to save humanity.[5]
We flag this tension without delving into it further and shift to economic investment analysis.
In everyday language "Let's ask ChatGPT" is replacing the previous habit of "Google that".
AI Boom or Bust
Investments in AI over the past ten years are estimated at more than USD 1.5 trillion and, in scale, are comparable only to the rise of the railways in the 1850s, electrification in the 1920s, or the build-out of internet infrastructure in the 1990s.[6]
Global corporate investment in AI
Cumulative, USD billions, 2015–2024
The United States is investing by far the largest sums. China is attempting an overtaking manoeuvre with its own methods (DeepSeek), and Europe is trying to achieve the impossible within regulatory constraints and limited resources, in line with the familiar pattern: «The US innovates, China replicates, Europe regulates».
So far, most investment has gone into securing the prerequisites. Beyond the dominant themes of data centres and chips, the energy supply must be ensured. AI is both a productivity engine and an electricity story. According to the IEA base case, datacentre electricity consumption will double by 2030 to around 945 TWh. That equates to just under 3 per cent of global electricity demand, or 1.5 times all private households in the EU.[7] On the supply side, almost half of the additional demand over the next five years will be covered by renewables. Gas and coal will provide the rest, while nuclear energy will gain in importance towards the end of the decade. In the United States, nuclear power (particularly small modular reactors, SMRs) is likely to play a pivotal role.[8]
What comes next?
The next stage will be applications used by consumers and companies. That is also where investors' great hope (and bet) lies: they are banking on large revenues and commensurate profits. Superlatives have followed in quick succession:
On 29 October 2025, Nvidia became the first company in human history to reach a market capitalisation above USD 5 trillion.
Nvidia's path to the top
Share price performance in USD, 2015–2025
OpenAI is planning an initial public offering at a valuation of up to USD 1 trillion[9], with first-half 2025 revenue of USD 4.3 billion.[10]
Palantir recently reported third-quarter 2025 results. Over the past twelve months, the company generated around USD 1 billion in profit. Its current market capitalisation is USD 450 billion. Valuing a company at 450 times its earnings is ambitious.
As investors, we want to know whether and where investment in AI still makes sense today - and whether we are in a bubble.
Are we in a bubble?
According to our advisory-board member Prof. Dr Thorsten Hens, the answer is yes. That need not be negative: speculative bubbles and misallocation are part of the innovation process. What benefits societal progress can nonetheless lead to great disappointment for the individual investor. He therefore advises caution in the current euphoric market phase. At the same time, Prof. Hens also puts the potential impact of an AI-bubble burst into perspective: unlike the dot-com era, most of today's investment comes from US technology companies, which arguably have the deepest pockets globally. In other words, potential mis-investment or failure would not hit the poorest.[11]
Adrian Locher, General Partner of the early-stage AI investor Merantix Capital, sees developments as still being at the beginning. Beyond ChatGPT, the potential of AI is particularly high in areas such as cancer research or cyber security. Profitable business models have already been developed here that are superior to conventional methods. He sees a lot of activity in venture capital and invests accordingly with focus.
Regardless of the hype, the key question for us as investors is whether we receive an adequate future return for capital invested today. The intrinsic value of an investment equals the total value of all future payments received up to a given point in time, discounted at an appropriate rate.[12] Because this requires estimating future developments, the likelihood that the expected cash flows materialise is critical. In the current environment and given the partially astronomical valuations the margin of safety appears thin and the risk of disappointment high.
What should investors do?
We see signs of excess and anecdotes that argue for tactical caution in cases such as Nvidia or Palantir.
Michael Burry, made famous by «The Big Short», has positioned his portfolio for an AI-bubble burst. With a USD 1.1 billion stake, he is betting on declines in Nvidia and Palantir share prices.[13] Deutsche Bank is currently examining hedging strategies against losses in AI shares to offset potential risks linked to loans extended for the construction of data centres.[14]
Applications will show whether the expected revenues are achievable. Many developments are occurring at the early-stage company level, with all the familiar advantages and disadvantages.
In addition, the long-term prospects for equity-market returns remain intact. The world's population is set to grow to 10 billion by 2060, per-capita income is rising, and technological progress in the form of AI will drive productivity gains for companies. It therefore continues to make sense for investors to remain invested in the broad equity market.
PC
[1] nobelprize.org, The Royal Swedish Academy Of Sciences, Joel Mokyr «for having identified the prerequisites for sustained growth through technological progress», and Philippe Aghion und Peter Howitt «for the theory of sustained growth through creative destruction»
[2] see Schumpeter, J.A. (1942)
[3] Yudkowsky E., Soares N., 2025
[4] Aws.amazon.com, «What is AGI (Artificial General Intelligence)?»
[5] Fortune.com, «A.I. godfather says an existential threat is preposterously ridiculous»
[6] Goldman Sachs Research, «AI: In a Bubble?»
[7] Eurostat, Electricity and heat statistics 2023
[8] International Energy Agency (IEA) Report 2025, «Energy and AI»
[9] Reuters, «OpenAI lays groundwork for juggernaut IPO at up to $1 trillion valuation»
[10] The Information: «OpenAI Forecasts Revenue Topping $125 Billion in 2029 as Agents, New Products Gain»
[11] SRF Audio mit Prof. Dr. Thorsten Hens, «Wie gefährlich ist die KI-Blase?»
[12] Video: Warren Buffet: How to Calculate Intrinsic Value
[13] cash.ch, «Michael Burry schichtet Portfolio um und warnt vor einer Überhitzung an den US-Börsen»
[14] Financial Times, «Deutsche Bank explores hedges for data centre exposure as AI lending booms»
IC No. 44 - Human Intelligence
