With new technologies, the focus is often initially on the dangers: there is a fear that Artificial Intelligence will replace thinking and destroy jobs. This reflex falsely assumes a fixed economic pie that is merely being redistributed, but economic history proves the opposite. Productivity gains enlarge the pie. Just like with the invention of the light bulb, the automobile, or the internet, old professions disappear while new ones emerge; technological progress has always increased the prosperity of the general population, not just that of a minority.AI takes over the execution, leaving humans with more capacity for the actual purpose. Nothing else lies behind the concept of productivity: the potential to achieve significantly more with the same work-force. If, for example, the efficiency of a software developer multiplies, the price of their service drops accordingly, allowing small businesses to afford technologies that were previously reserved for large corporations. The addressable market thus does not grow linearly — it explodes. High-quality AI support in daily operations frees up capacity for backlogs and innovation, noticeably improves the quality of almost every service, and fuels the rate of innovation — from scientific breakthroughs to the daily routine of our own company.
The Scale of AI Investments
The brightest minds in the technology industry are going uncompromisingly on the offensive regarding data centers. For the year 2026, the announced investments by Alphabet, Microsoft, Amazon, and Meta total around 700 billion US dollars — and even reach 750 billion when including Oracle. This exceeds the entire German federal budget, equals roughly six times the Swiss federal budget, and accounts for around two percent of the US gross domestic product.To finance this, these corporations are now tying up nearly their entire operating cash flow and issuing bonds in the triple-digit billions. What were once capital-light business models are turning into highly capital-intensive constructs. Yet a look at the order books refutes the thesis of a pure speculative bubble: Microsoft’s commercial backlog has doubled to over 600 billion US dollars, and Google Cloud's backlog rose to over 460 billion. Mean-while, the language model company Anthropic took just under four years after its founding to reach its first billion in revenue — and then increased its annualized revenue from 9 billion to a remarkable 47 billion US dollars within just a few months. Never before in economic history has a company grown so rapidly; a comparable pace was achieved only shortly before by OpenAI with ChatGPT. Established technology giants like Alphabet and Amazon required one to two decades to reach a comparable revenue scale. The demand is therefore just as real as it is impressive, even if the valuations of the players are highly heterogeneous.