AI is not a bubble vs AI is a bubble

It sounds contradictory, but it hits the nail on the head in the current debate. Artificial intelligence is considered both a fundamental technology of the future and an overheated market. If you take a closer look, you will see that these two assessments are not mutually exclusive. They describe different levels of the same development.

AI is not a bubble – structural change in work and everyday life

Artificial intelligence is already changing the way companies work, make decisions and develop products. In industry, learning systems optimise production processes; in medicine, they support diagnoses; in retail, they analyse customer behaviour in real time. In journalism, marketing and software development, too, more and more processes are relying on automated evaluation and generation. The impact on the world of work is profound. Certain routine tasks will disappear, especially where data is structured and pattern recognition is sufficient. At the same time, new job profiles are emerging: specialists in AI training, data ethicists, prompt designers, system architects and experts in AI-supported process integration. Historically, technological upheavals have rarely been pure job destroyers – they shift job profiles. The influence is also noticeable in everyday private life. Language models assist with research and text work, image recognition software sorts photo collections, and smart assistance systems structure appointments and control household appliances. What is still perceived as innovation today is likely to become commonplace in the medium term – similar to the smartphone, which was still considered a luxury product just a few years ago. Economically, AI opens up enormous potential. Companies can increase efficiency, develop new services and establish data-driven business models. Automated customer communication, personalised product offers and predictive maintenance are just a few areas of application. Those who use AI in a concrete and practical way can reduce costs and tap into new sources of revenue at the same time. In this structural sense, AI is not a bubble, but part of a long-term transformation process.

AI is a bubble – questionable business models and circular transactions

At the same time, the current boom has all the classic characteristics of a bubble. Valuations of AI companies are rising rapidly, investments are reaching record highs, and not every business model is yet economically viable. Many providers are investing heavily in computing power, data centres and development without it being clear how these expenses can be refinanced in the long term. Added to this are constellations that, at the very least, raise questions. When a large chip manufacturer acquires shares in an AI start-up and this start-up in turn purchases hardware from the same manufacturer, sales increase for both sides. At first glance, this seems convincing to investors. However, these are circular cash flows that do not yet provide evidence of sustainable demand. Rising sales do not automatically mean profitable growth. Many applications are also in an early experimental phase. Companies are testing AI tools without having established long-term subscription models or stable revenue structures. Competition is intense, and price pressure is likely. It remains to be seen whether high investments in training, infrastructure and energy consumption will pay off in the long term.

Difference from previous technology bubbles

Comparisons with the dot-com phase at the turn of the millennium are obvious. At that time, speculative valuations and a lack of real business fundamentals led to a massive slump. The difference today is that the AI wave is backed by financially strong corporations with considerable capital reserves. Billions in investments are not being made exclusively by small investors, but by established companies. If some of the expectations are not met, this would primarily affect these companies and their investors. Whether the investments will pay off in the long term for technology corporations such as Apple, Google or Microsoft is impossible to predict. The only certainty is that real products, real users and concrete use cases exist – unlike in some previous speculative phases.

Conclusion: substance and exaggeration at the same time

Artificial intelligence is neither pure illusion nor an infallible promise of salvation. It is a technological revolution with a substantial impact – and at the same time a market in which expectations, valuations and business models have yet to come together. The potential is sustainable and will shape both the economy and society. The decisive factors will be who develops viable business models, who translates efficiency gains into real added value and who prevails in the competition. Whether global technology corporations or specialised start-ups with previously underestimated solutions will ultimately dominate remains to be seen. One thing is clear: the topic will remain – but the market structure is likely to change several times.