AI is on the Agenda for European Leaders and it Should Go Hand-in-hand with Climate & Energy Policies For AI Success in Europe

As Keir Starmer went through the diplomatic ritual all newly elected leaders go through, making and taking congratulatory calls from fellow world leaders, there was an interesting point that could easily be missed.

In the Number 10 readout from Starmer's call with (soon to exit) French PM Macron, nestled in amongst the items one would expect to see (Ukraine, climate, the Middle East) there was a specific mention of Artificial Intelligence without going to any more detail. Earlier this year the EU raced through legislation covering the use and ethics of AI however one key area European policy makers must consider is Europe's energy market problem and how it poses a threat to Europe's AI ambitions.

By now we have all seen the articles looking to demonstrate just how power hungry AI applications are. Through the AI lifecycle, from producing silicone, to setting up the data centres to even the code used to create LLMs and GenAI, it is a power hungry technology. Python, the language of choice to create AI consistently ranks as one of the least energy-efficient programming languages in research papers. We are at the point of discussing Small Modular Reactors (SMR) in AI data centres to give the power they need. But one thing is clear, if Europe really wants to be a leader in AI, Europe needs to sort out its energy problem.

The eagle-eyed energy commodity analyst will tell you natural gas prices were already increasing before the war in Ukraine. As the world emerged from lockdown, it needed energy and lots of it. Then came the war in Ukraine which shocked energy markets, Europe in particular. European leaders had to send a strong message to Russia, and after years of easy energy from Russia, Europe had to get serious about moving away from Russian energy.

European manufacturing is facing a really difficult time because of Europe's failure in creating a sustainable and competitive energy industry and AI could suffer the same fate. Europe's energy competitiveness, or lack of, could have direct implications for Europe's ability to compete in the AI space.

Nvidia's market cap was increasing at impressive rates pre-Covid as analyst began to notice the sea-change in the market for microchips and it has become headline news recently as it overtook Apple's market cap. This surge in demand for microchips signifies the AI arms-race that is unfolding as every industry and every sector looks to implement AI, from automakers, to defence, to whitegoods, to financial services. Microchips used for AI, like all microchips, are made from electronic-grade silicon requiring extremely high purity (over 99.9999%). The manufacturing process utilises various high temperatures exceeding 1000°C at different stages. All of this needs immense energy, and frankly, energy costs are so much more cheaper in the US and Asia. Jim Ratcliffe, one of Britain's wealthiest industrialist, Chairman and Founder of INEOS, recently complained to Bloomberg how energy costs in Europe are making it difficult to compete with US and China in the chemicals industry. These same challenges will hamper any European firm willing to take on TSMC. Potentially European firms that need to adhere to data-sovereignty and privacy rules could find the costs of running AI applications from data-centres and cloud regions based in Europe cost more than if they located elsewhere as a result of energy costs.

The urgency felt by leaders to get ahead in the AI race and the fear of becoming an AI laggard, straddling behind those who embrace and adopt the technology mean the energy and climate impacts of AI risk becoming a secondary concern to the productivity, strategic and competitive edge AI promises for policy makers.

However we shouldn't look at this as a pessimistic trade-off between adopting technology and protecting the planet. AI should help efforts to decarbonise and increase investment in sustainable energy by accelerating the business case for change. AI as technology itself will hopefully become more efficient overtime, blockchain has made advances in energy efficiency since its first introduction and AI will hopefully follow the same and reduce the technology's overall energy burden. AI can also be used as an energy efficiency tool itself by adding automated intelligence to optimise power-usage and management in data-centres. Overtime policymakers shouldn't just focus on regulating AI ethics and data, but policymakers should invent new innovative levers to force technologists to make AI more energy efficient.

European leaders also need to tie-in their climate & energy policies with their AI ambitions. AI should embolden the sustainable energy ambitions of European policy makers with the knowledge that an abundance of sustainable energy will increase Europe's ability to embrace AI. Even TSMC is looking to power their plants with sustainable energy, Europe should take a bold-step and double-down on sustainable energy investments to make Europe competitive in the AI age. Sustainable energy, not just education and talent, will be a key pillar for Europe's ambition to be a global AI leader.