When world powers risk war to slow down AI competition, AI is no longer a bubble

It's about more than energy

It's about industrial precursors, about critical infrastructure, about data centers, about cooling, about material flows and thus ultimately about the technical foundation of modern AI.

One example that hardly anyone has on their radar: Qatar supplies nearly a third of the world's helium. Helium currently has practically no viable alternative for central steps in semiconductor manufacturing. It is used, among other things, for thermal management and sensitive production processes in chip manufacturing. If production facilities fail there and sea routes are blocked, it doesn't just hit the energy market. It directly hits the ability to build AI hardware.

Supply chains as a strategic weapon

This is exactly where my thesis begins. Perhaps we are seeing not just a military escalation in the Middle East, but a conflict that simultaneously slows down technological competition. Not openly. Not cleanly provable. But plausible as a direction. Because whoever puts supply chains under pressure that are crucial for chips, hardware, and industrial scaling doesn't just hit states, but entire innovation pathways. You don't have to completely stop the competition. Often it's enough to make it more expensive, slower, and less secure.

Who benefits from the instability?

For me, therefore, the more exciting question is not who dominates which headline, but who strategically benefits from this development. When a power bloc makes its own infrastructure more robust, secures its energy supply for AI, and simultaneously increases global dependencies elsewhere, a dangerous imbalance emerges. In my view, that is exactly where the real lever lies. Not in open confession, but in effect. Whoever becomes more independent can scale better. Whoever depends on hardware, cloud, raw materials, or unstable supply chains loses time.

Slowing down is enough

China will find solutions. Of that I am convinced. But even that doesn't change the core of my consideration. It is enough to slow down China and other catching-up actors. It is enough to drive up the costs for local infrastructure. It is enough to bring companies worldwide into a situation where their own systems become harder to finance and they increasingly fall back on large, already dominant cloud providers. Whoever controls computing power will in the future control not only technology, but also dependency, speed, and value creation.

AI is infrastructure. AI is power.

That is exactly why I consider the discussion about an alleged AI bubble to be increasingly superficial. When even world powers are willing to accept massive economic shocks, geopolitical tensions, and potentially global side effects, while simultaneously struggling for technological supremacy, then that shows above all one thing: AI is no longer a marginal phenomenon. AI is infrastructure. AI is power. AI is competition at the hardest conceivable level.

What this means for companies

For companies, this results in an uncomfortable but necessary consequence. The question is no longer whether AI will change one's own industry. The question is how dependent one wants to be on the infrastructures of others in the future. Whoever does not develop a clear line of their own today runs the risk of tomorrow being merely a customer in a game whose rules were written by others long ago. That is exactly why companies should concern themselves not only with tools, but with sovereignty. With their own data base. With their own process intelligence. With the question of which parts of their own AI future must be built locally, controllably, and resiliently.

Recognizing the pattern

Whoever pretends that this is only about another regional war overlooks the pattern. Perhaps we are currently experiencing the phase in which geopolitical conflicts and AI competition finally collapse into each other. And it is precisely this that makes visible how serious this topic really is.