What exactly is an AI? It’s nothing more than a highly developed computer program. The best-known AIs, such as OpenAI's ChatGPT or Google's Gemini, are based on huge amounts of data. However, even if the term "cloud" often suggests otherwise, all this data is stored on physical servers in large server rooms around the world and does not simply "float in the sky". The hardware requirements for artificial intelligence are correspondingly high. Not for the individual users, but for the developers, providers and suppliers of such technologies. But how will hardware requirements develop in the future and what does this mean for companies that develop programs for medical technology?
Massive power consumption due to AI - the trend is rising
According to John Pettigrew, CEO of the British transmission system operator National Grid, the power consumption of the data centers required for artificial intelligence could increase six-fold over the next ten years. "Future growth in foundational technologies like artificial intelligence and quantum computing will mean larger scale, energy-intensive computing infrastructure," Pettigrew told the BBC.
A study by the Swiss economic research institute Prognos found an increase in electricity consumption for the "battery factories and data centers" sector of 13 TWh (terawatt hours) by 2030 compared to 2018. Interestingly, the consolidation of data centers and the use of more effective hardware and software could even lead to a slight decrease in electricity consumption for data centers if they are considered independently of battery factories.
However, according to experts, many of these estimates are rather speculative. This is because the big tech companies such as Google, Apple and Microsoft keep their technical data secret, including their exact computing power and power consumption. However, some indications give cause for concern: if we analyze the amount of computing power that could be used for AI, which could be achieved by Nvidia microchips alone, we arrive at an energy consumption of 85 to 134 TWh. "You would be talking about the size of a country like the Netherlands in terms of electricity consumption. You're talking about half a per cent of our total global electricity consumption," says Alex de Vries from the Amsterdam School of Business and Economics. His recommendation: AI should only be used when it is really needed.
Kate Crawford, Professor of Artificial Intelligence at the University of Southern California in Los Angeles, adds: "These energy-intensive systems take enormous amounts of electricity and energy, but also enormous amounts of water to cool these gigantic AI supercomputers. So, we are really looking at an enormous extractive industry for the 21st Century."
Another problem is that the energy infrastructure in many European countries dates back to reconstruction after the Second World War and is hardly, or not at all, designed for future requirements. According to an analysis by the Handelsblatt Research Institute, the goal of climate neutrality in Germany alone would require investments of 1.1 trillion euros in the electricity grid.
