Sweden Could Lose AI Investments Unless We Switch Air for Liquid

AI development is driving an explosive increase in power and cooling demands for data centers. The IEA predicts that within four years data centers could account for roughly three percent of global electricity consumption, while the global competition over where next-generation AI capacity will be built is intensifying. In February, the government presented a new AI strategy aiming for Sweden to be among the world’s top ten AI nations. To make that ambition a reality, strategic infrastructure decisions are needed. One urgent choice is to accelerate the transition from air cooling to liquid cooling in Swedish data centers.

A single query to an AI model can use about ten times the energy of a standard web search. That drives demand for servers with much higher power density, creating cooling needs that conventional air cooling struggles to meet efficiently. Energy therefore becomes more than a marginal cost—it is a central investment factor. Sweden’s AI ambitions depend on where investment capital chooses to go. Investors seek markets where the energy question is resolved from the outset. That means cooling must be modernized, and air cooling replaced with liquid cooling.

Demand for high-performance, high-density racks is rising rapidly. Today, rack densities already reach about 40 kW and in some environments exceed 100 kW—levels that are difficult and costly to manage with air cooling alone. McKinsey & Co reports that liquid cooling systems can reduce energy consumption by more than 27 percent compared with traditional air-cooled systems. Unlike traditional CPUs, GPUs and other AI accelerators produce much more concentrated heat loads, which requires different handling. Direct-to-component liquid cooling can significantly improve energy efficiency and also enable heat recovery—for example, to heat buildings.

For investors and operators the issue goes beyond the technology itself: energy is one of the largest ongoing costs in a data center. According to the Uptime Institute, high rack density is the primary driver leading operators to adopt liquid cooling. Liquid cooling has therefore moved from being a competitive advantage to a prerequisite for scaling AI infrastructure over the long term.

Higher upfront investments in more efficient cooling can translate into lower operating costs over time. This allows better capacity utilization and reduces exposure to future electricity price volatility. The alternative—continuing to build with less efficient solutions—risks higher costs across the facility’s lifetime.

If Sweden is to attract world-class AI investment, it needs resource-efficient data centers where liquid cooling is integrated into the core infrastructure.

First, the transition must be phased. Existing data centers cannot be rebuilt from scratch overnight. Hybrid approaches that combine air and liquid cooling allow operators to introduce more efficient cooling gradually as power densities increase.

Second, clearer guidance is needed on how to integrate these hybrid systems. Standards exist for both air and liquid cooling, but combining the technologies introduces interpretation gaps around safety and retrofit work, which can increase project complexity and delay investment decisions.

Third, the entire value chain needs more expertise. The industry needs the right skills and capacity to operate, scale and maintain liquid cooling. That requires new specialist roles and training covering everything from heat-recovery systems to operation of high-density GPU environments.

Building AI capacity without ensuring energy-efficient cooling is not sustainable in the long term—neither economically nor strategically. Sweden has the opportunity to lead in creating competitive, resource-efficient and scalable AI infrastructure, but that will require focused investments and clear priorities that demonstrate commitment.

Hanna Oredsson, Sales Director for Data Center at Schneider Electric Sweden.