Is artificial intelligence shifting into the geopolitical arena?

For a long time, we have been discussing artificial intelligence primarily through the lens of technology companies. What did OpenAI do, which model did Google release, how much did Anthropic grow, which model did China develop… However, a development that has occurred in recent days has completely changed the direction of this debate.

The access restrictions imposed on Anthropic’s most powerful models, Mythos and Fable, show that artificial intelligence is no longer just a software product. We are entering a strategic field where states are directly intervening.

In fact, this transformation has been coming for a while. First, restrictions were placed on the export of advanced artificial intelligence chips. Then, high-performance graphics cards were placed at the center of national security debates. Now, it is the turn of the models themselves. It is no longer just hardware, but information that is beginning to be controlled.

This situation reminds me of other transformations in history. At one time, uranium was not just an ordinary mineral; it was a strategic element of power. Later, supercomputers were placed under export controls. Powerful cryptography software was classified as munitions for many years. Today, large language models are being added to that same list. Artificial intelligence has entered the ranks of technologies that states will either release or restrict.

Perhaps the most important result of this development is that the artificial intelligence race will no longer take place between companies, but between countries. OpenAI, Anthropic, Google, xAI, or Meta are no longer just private sector companies. They are increasingly moving into a position similar to defense industry companies. The models they produce carry as much strategic value as they do economic value.

Another important change is that we are moving from chip wars to model wars. For the last three years, everyone has been talking about NVIDIA’s graphics cards. However, in the near future, the topic we will be discussing will not be graphics cards, but which country can access which model. The real competition will now take place over algorithms.

This also shows that the internet is beginning to fragment. For many years, the internet was a common platform. Today, on the artificial intelligence side, ecosystems are forming in the US, China, and likely Europe that are increasingly diverging from one another. Are the new borders of the digital world being drawn not by cables, but by models?!?

This is precisely why the importance of open-source models is increasing. You can shut down an API. You can prevent a company from providing services. However, you cannot stop open-source models that have spread across the world with the same ease. In the coming years, the struggle between open-source artificial intelligence and closed commercial models will be one of the most important topics in the technology world. Of course, that is a separate article topic... Perhaps a video....

When viewed from Turkey’s perspective, we still have not asked the real question. The question that needs to be asked is: What will we do if access to these models is restricted for any reason tomorrow? What investments are we making today to build our own models, our own data, and our own information infrastructure?

At this point, I would like to recall the national artificial intelligence strategy proposal I prepared a few weeks ago.

( https://www.linkedin.com/pulse/t%C3%BCrkiyenin-yazay-zeka-eylem-plan%C4%B1-ve-stratejisi-nas%C4%B1l-atif-unaldi-zfeif )

In that article, I proposed a long-term roadmap extending from data sets that would ensure the protection of the Turkish language to quantum computers, not just model development. I shared the article on my LinkedIn account. The developments taking place today show more clearly why that strategy is not only a technological necessity but also a geopolitical one.

Because artificial intelligence is no longer just a tool that writes better text or produces more beautiful images. It is shaping the economy. It is affecting defense. It is generating intelligence. It is accelerating scientific research. It is guiding societies. It was unthinkable that such a powerful technology would not enter the radar of states.

Perhaps historians will write years later that 2026 was the year artificial intelligence ceased to be software and became an element of national power. When that day comes, the debates surrounding Mythos and Fable will be remembered not just as a corporate decision, but as one of the first major signs of the new artificial intelligence geopolitics.