During a wide ranging speech at the annual World Economic Forum (WEF) conference in Davos, Switzerland today, President Donald Trump claimed that the U.S. at this juncture has the market for artificial intelligence (AI) “pretty well locked up.”
Much of that perceived control over AI appears to be tied to economic policies based on an ability to impose tariffs and investments in energy. President Trump claimed he personally came up with the idea to give data center providers approval to build their own plants and private grids in as little as two to three weeks, as part of a goal to more than double the energy available just to handle AI requirements. Microsoft, for example, with the help of a $1 billion loan from the U.S., is investing resources to restart the Three Mile Island nuclear energy plant to help meet its energy requirements.
In total, President Trump noted $18 trillion in investment commitments that have been made in the U.S. in the last year, much of which is being allocated to fund the building of massive data centers.
President Trump, however, also made it clear that the administration has a marked preference for using oil, gas, and coal along with nuclear energy over other sources such as windmills, which he said is part of a larger “green new scam.”
Additionally, Trump chided European countries for not being able to keep pace with the level of energy investment that will be required. European countries are far too focused on green energy projects such as windmills rather than maximizing, for example, oil drilling in the North Sea that is being hampered by environmental regulations. Specifically, windmills in addition to losing a thousand dollars with every time a blade goes around also destroy land values, he claimed.
The only country making money on windmills is China, which dominates the manufacturing of windmills today. However, the president also incorrectly claimed that China is not building windmill farms when, in fact, there are numerous offshore windmills operating off its coasts.
Of course, it’s not clear how any one country is going to lock up an AI market. While NVIDIA, OpenAI, Microsoft and Google are based in the U.S., there are already signs that countries around the world are starting to focus more on AI sovereignty and the sovereign clouds required to run them. Additionally, China is already making an effort to rely more on processors made by, for example, Huawei to run AI workloads.
In general, almost every element of the AI supply chain, starting with energy, is currently constrained, says Futurum Group CEO Daniel Newman. Despite those issues, however, AI remains the “fastest technological gigacycle in history,” he notes.
The challenge, as always, is finding the best way to gain and maintain a competitive advantage when much of the underlying infrastructure being used to drive it remains in a state of flux that makes it difficult to determine what is simply new table stakes versus meaningful differentiation.
