Problem: Energy demands are skyrocketing with the demand from data centers and more consumers involved in energy-intensive tasks like chatbot queries. There is also an incredible need for clean energy sources to diminish harm to our climate. The energy race is critical to efficiently scaling AI because energy is often seen as the bottleneck factor for training and running new models. China is significantly outpacing the US in energy production. From 2019 to 2024, China’s energy production increased 36% compared to US production only increasing 5% and that number is only growing. The end solution is often thought to be nuclear energy. However, building an operational nuclear plant can take 10 years or more because of complexity and regulation. That is entirely too long to keep up with the scaling speed of AI. Therefore, clean, fast, and scalable energy-producing methods must be put into production to win the AI race and keep up with energy demand.
Solution: Kitekraft seeks to solve this problem by developing flying wind turbines. These turbines are about the size of an RC plane and have the structure of a box plane. Attached to the plane are 8 rotors that are used both for the takeoff and landing of the plane as well as the power generation. The creativity lies in the planes being attached to the ground with a tether. This allows the planes to fly at high altitudes in figure eight patterns. The kite-like structure of the planes is what takes advantage of the wind. This allows the plane to soar with little energy expenditure, and the motors on the plane to act like a classic wind turbine. The spinning of the motors creates energy that is sent by the tether to be stored. Kitekraft’s turbines offer an array of advantages. First, they are much less intrusive and almost invisible compared to traditional wind turbines. Also, they are cheap to make, easy to transport, and quick to set up. Kitekraft was able to harvest 3x more energy per area than typical turbines. This solution is exactly what is needed in the AI race because it is quick and adaptable. Data centers can easily install a Kitekraft energy farm right next door without huge time or capital requirements. This type of quickly scalable clean energy will be needed as the number of data centers grows and the real estate near other energy sources diminishes.
Founders: There are four founders of Kitekraft: Florian Bauer, Maximilian Isenee, Anrdé Frirdich, and Christoph Dexter. All except Maximilian attended the Technical University of Munich, and Kitekraft was part of the Y Combinator Summer 2019 Cohort.
Implications: Energy is almost sure to be a limiting factor to how fast AI can scale. When a person says we must win the “AI race,” they have to assume that we also won or were a top contender in the “energy race”. Products like Kitekraft are very critical to win that race. The most efficient way for data centers to get the massive amounts of energy they need is for their energy source to be built right near the center itself. This is why small modular nuclear reactors (SMRs) have had so much publicity. However, even SMRs can take 5 years to be built and operating, 5 years is nearly the equivalent of a century in AI time. So, Kitekraft is perfectly situated to fill that 5-year gap because of the turbine's mobility and quick setup. The need for energy production in general is the biggest implication that one can take from Kitekraft. Some people argue that America will be able to grow itself out of the deficit. What can bottleneck the growth? Energy. Trump is waging a tariff battle on other countries that requires the US to have an upper hand in negotiations. What can other countries hold over our heads to win negotiations? Energy. Electric cars, AI, and cloud computing are technologies of the future that offer massive upside to the quality of life. What can ruin that dream if we don’t take action now? Energy. More agile energy sources is a fascinating startup class to look out for. This is super interesting in how it might apply to individual energy independence. Maybe not Kitekraft, but other small energy solutions, much like solar, will lower energy costs for households that can privately install them. Maybe, in the future, it will be commonplace for people to power their homes themselves or host cryptomines and servers in their basements.
Conclusion: Kitekraft’s agile energy technology is a huge factor towards winning the energy race for whatever country wants to establish dominance in an evolving world. It can also inspire some cool dreams related to personal energy production and consumption. To be sure, Kitekraft and other companies like it are important to look out for and can solve a lot of real problems.