Earlier I read a post by someone who is holding back on diversified investments worldwide due to the overvaluation caused by AI.
In my opinion, we are only at the beginning of the AI transformation. The potential for productivity gains for companies outside the AI supply chain is huge. Not through simple assistance systems, which are not even in widespread use yet. But through AI agents. Their use in companies is still in its infancy.
Companies that cleverly integrate agents into their business model can grow significantly and displace the competition. This is a huge opportunity, the likes of which we will not see very often.
At the same time, however, there is also a risk if we do not find social/corporate solutions to cushion the potential loss of purchasing power due to mass redundancies.
What do you think about this?
Of course, I also asked KI (Gemini) what she thinks:
While the first wave of AI euphoria primarily boosted the "shovel sellers" - chip manufacturers and cloud giants - we are now looking at the much bigger second wave. Over the next 10 years, the AI transformation will no longer just be a tech sector issue, but will redefine the fundamentals of every listed company from consumer goods to heavy engineering.
The key question for investors is: What happens to the companies that don't build AI, but use it?
1. from assistance to autonomy: the new margin arithmetic
In the first few years (2024-2026), we primarily saw assistance systems (copilots). These made knowledge workers more efficient, but hardly changed the business model. However, the next 5 to 10 years will belong to agentic AI.
Agents not only work for humans, they act autonomously within target specifications. This means a fundamental shift for the profit and loss account (P&L):
- Fixed cost degression: Human labor is a variable cost factor. AI agent infrastructure is largely a fixed cost factor. As soon as the system is up and running, the marginal costs for the next unit (service, customer support, design) fall to almost zero.
Valuation shift: Companies that master this transition will no longer be valued on the stock market as traditional industrial or service stocks, but will receive multiples (P/E ratios) that we have previously only seen in the software industry.
2. the K-shaped destiny: the gap between adopters and laggards
We are heading for an extreme divergence on the stock market. Historical data on digitalization shows that early adopters increase their productivity by around 3% annually, while laggards stagnate.
This effect will be magnified in the AI era. A company that shortens its R&D cycles from two years to six months through AI agents will simply absorb the market share of its competitors. Investors therefore need to shift their analysis from "industry trends" to "company-specific AI excellence".
3 The macro paradox: productivity vs. purchasing power
This is the biggest risk to long-term equity growth. AI promises a massive expansion of the supply side through efficiency. But who will form the demand side if the transformation leads to structural unemployment in the field of knowledge work?
- The demand gap: If redundancies in middle management and the service sector are not absorbed by new job profiles, there is a threat of a decline in purchasing power. This would particularly cyclical consumer goods stocks (automotive, luxury, travel) under pressure.
Deflationary pressure: If production costs fall as a result of AI, but at the same time demand weakens due to a loss of income, we are heading for a deflationary phase. In such an environment, companies with extreme brand power (pricing power) are the only safe havens, as they can defend their margins despite falling overall demand.
4 The "asset-light" revolution and new valuation benchmarks
In 10 years' time, the size of a company will no longer be measured by the number of employees. We will see "one billion dollar companies" operating with less than 50 full-time employees and thousands of AI agents.
For stock analysis, this means:
Data sovereignty is the new moat: Patents are ephemeral, but proprietary data sets used to train proprietary AI models form an impregnable fortress.
Decline in CAPEX intensity: Many physical processes are optimized by precise AI simulations (digital twins), reducing the need for expensive physical prototypes and miscapacity.
Conclusion for the community
The next 10 years offer enormous opportunities for returns, but require a new look at the portfolio. Share value will decouple from companies that "use AI" and flow towards those that build their rebuild their entire organizational structure around autonomous agents.
The biggest risk remains systemic: if the AI dividend does not find its way back into the consumer cycle, the productivity explosion could paradoxically end in a sales crisis. Investors should therefore increasingly focus on companies that either offer essential goods or have such strong emotional brand loyalty that they are less affected by a volatile labor market.
