AI - How wisely does the market act?

AI applications are currently in the process of structurally changing (not only) the entire software market. Brief announcements are now enough to cause sharp price drops on the stock market. On February 05, 2026, AI developer Anthropic announced “improvements” to its large language model Claude Opus 4.6. Essentially, these improvements involved more efficient input processing while reducing the model's own sources of error. Two trading days later, the value of software stocks worldwide had fallen by nearly USD 250 billion.

At the same time, individual pieces of information that fall short of market expectations are currently being viewed as a fundamental business risk in otherwise positive annual reports from the software industry, resulting in significant valuation discounts. At the end of January, for example, SAP saw its share price immediately plummet by over 15% after presenting its preliminary 2025 financial results, which were nothing than bad. A similar situation occurred shortly afterwards with the shares of the successful US chip designer Arm, which lost 8% in one day. AI is currently considered an “existential risk” according to the reasoning of rating agencies. However, this does not prevent these agencies from simultaneously issuing buy recommendations for software stocks with projected value increases of sometimes more than 50%. And this despite the fact that no AI provider has a sustainably profitable business model yet.

Rational explanations alone have their limits here. But how to judge then on actual market behavior?

Market behavior is usually the result of actions taken by individuals or organizations using tools and other resources. In transparent markets in particular, participants are constantly engaged in close coordination processes, guided by certain expectations (e.g., about anticipated rewards) and changing guiding principles (e.g., about appropriate behavior). This corresponds to an organizational pattern known from swarms.

But what determines the behavior of an individual member, an individual person?

The Interdependence Theory (Kelley and Thibaut) on fundamental processes of social exchange offers further explanatory approaches here. This is because interactions also take place continuously between participants in the software market, which in turn influence others in their actions. The basic expectation of every market participant is clear: in the end, the benefits achieved must exceed the costs incurred – in speculative markets such as that of AI applications, even significantly so.

After two years of an increasing offer of AI tools, many market participants now seem to have changed their perception of the situation. Not only has the cost incurred so far been higher than expected, but it is also set to grow even more significantly in the future. At the same time, the benefits achieved have not yet led to lasting financial gains. Additionally, the competitive behavior of large market participants leads to the conviction that in the end some participants will be losers.

This not only leads to new considerations, such as outsourcing risks to third parties, but also changes the motivation and emotions of the actors themselves. Whereas it was previously important not to miss the departing train, the priority now is to minimize potential future losses in order to avoid ending up on the sidelines. However, according to Expectation Theory (Tversky and Kahneman), expected losses are valued significantly higher than expected gains. People literally see their chances slip away. Fear spreads. This is the affective base for irrational actions – which is currently reflected in the stock prices of the software industry.

From a psychological point of view, what can an individual do in such a situation?

For example, becoming clear about the underlying behavioral patterns. Not to jump on the next train, just because you hear a whistle. Examining under what conditions you can afford which losses. Setting narrowly defined limits. Seeking the perspective of other persons. In short: becoming aware of own affects and then (inter)acting wisely.