Zack Polanski wants Britain to slam the brakes on new data centres, warning about their impact on water supplies and the electricity grid. He is right about the pressure. Data centres already consume around 2% of Britain’s grid electricity, and demand will rise sharply as AI develops. Some proposed sites are in entirely the wrong places. Others may end up relying on gas-fired generation because the grid cannot connect them quickly enough. Water use is not yet measured or reported nearly well enough either. But identifying a real problem does not make a moratorium the right answer. Data centres are no longer simply warehouses for storing photographs and emails. They provide the computing power behind AI, financial services, medical research, defence systems and much of the modern economy. That is why they are now classified as critical national infrastructure. If Britain chooses not to build them, the demand will not disappear. The investment - and quite possibly the carbon emissions - will simply move overseas. We would then become more dependent on foreign computing capacity at precisely the moment when data, energy and technology are becoming inseparable from national security. That would be an extraordinary act of economic self-harm. The Government believes its AI Growth Zones could unlock as much as £100 billion of private investment and more than 10,000 jobs. Those figures matter. But the bigger prize is ensuring that British businesses, researchers and public services have access to the computing power on which the next generation of economic growth will depend. The answer is not to stop building. It is to build properly. New data centres should be directed towards places where power, water and network capacity can genuinely support them. Developers should be expected to invest in additional electricity generation and grid infrastructure. They should disclose their water consumption, use recycled or closed-loop cooling wherever possible and design facilities capable of reducing demand when the electricity system is under pressure. Speculative projects should not be allowed to clog up the grid connection queue ahead of developments that are financed and ready to proceed. That all needs sorting out. But none of it is an argument for Britain to step out of the global race for AI. One lesson I learnt from overseeing Britain’s energy security was very clear: when new industries need more power, government has a choice. It can build the capacity required to capture the investment, or ration growth and watch it go elsewhere. Polanski’s answer to scarcity is the familiar socialist one: stop, ban and ration. It is the politics of managed decline, dressed up as environmental virtue. Britain cannot afford it. A country that refuses to build the infrastructure of the future will inevitably become poorer, weaker and more dependent on those that do. That vision is not simply outdated. In the age of AI, it is dangerous.