The rapid expansion of artificial intelligence (AI) is sparking a significant shift in data center infrastructure. One notable development is the growing adoption of liquid cooling systems, which are becoming essential for managing the heat generated by powerful AI applications. This transition addresses energy consumption challenges and signifies a burgeoning market for companies specializing in cooling technologies.
As businesses enhance their AI capabilities, the heat output from servers has become a pressing concern. Traditional air-cooling methods are often energy-intensive, consuming nearly as much power as the computing hardware itself. In contrast, liquid cooling systems can dramatically reduce energy usage, with estimates suggesting cuts of up to tenfold in electricity required for cooling. Shawn Tuteja from Goldman Sachs stated, “With liquid cooling technology, you can significantly reduce by as much as 10X the amount of electricity that goes towards cooling. And you can shift that more towards the power and distribution of actually fueling the AI trade.”
The momentum behind liquid cooling is already evident, with industry analysts reporting a 30% rise in interest this year. This trend is driven by the need for data centers to optimize their operations and enhance the return on investment (ROI) for hyperscalers. Dauvin Peterson of 22V Research emphasizes that advanced thermal management has become critical for maximizing AI token production, highlighting the economic incentives for adopting these cooling solutions.
The Rise of Liquid Cooling in AI
The shift towards liquid cooling is expected to accelerate with the introduction of new AI server architectures, particularly those being developed by NVIDIA. The upcoming Vera Rubin configurations, scheduled for delivery in 2026, will feature integrated liquid cooling systems, setting a new standard for AI infrastructure projects. Peterson points out, “As a standard for AI factories, it presents a clear and credible demand pull.” This indicates a broader trend where liquid cooling is not just an alternative but a necessity for modern data centers supporting next-generation AI workloads.
The surge in demand for liquid cooling technologies is benefiting a range of companies across the industrial and infrastructure sectors. Major players in the HVAC industry, such as Carrier and Trane Technologies, are reporting significant growth linked to data center requirements. Carrier recently revealed a staggering 500% increase in data-center-related orders, with CEO David Gitlin confirming that the company's backlog is set to meet its ambitious $1.5 billion sales target for the segment this year. Trane Technologies also noted a substantial uptick in its commercial HVAC backlog, bolstered by its acquisition of cooling specialist Stellar Energy.
Infrastructure Providers Thrive
Infrastructure providers are also thriving in this evolving landscape. Vertiv, which offers power and cooling systems for data centers, recently surpassed analyst expectations with first-quarter earnings of $1.17 per share. Their revenue increased by 30% to $2.7 billion, prompting the company to raise its full-year earnings forecast. Vertiv’s stock has soared by 84% this year alone, reflecting the market's confidence in its growth potential.
Eaton, another key player, has reported better-than-expected quarterly earnings, driven by stable demand from data center clients focusing on AI technologies. Meanwhile, Ecolab is making significant strides with its planned $4.8 billion acquisition of CoolIT, a company specializing in liquid cooling hardware, further emphasizing the trend's momentum.
Looking Ahead: Opportunities and Challenges
As the space continues to evolve, the importance of energy-efficient cooling solutions cannot be overstated. Companies like Madison Air, Dover, Chemours, and Solstice Advanced Materials are also positioned to benefit from this trend, supplying essential refrigerants and thermal management systems for next-generation cooling infrastructure.
The growing adoption of liquid cooling technologies is not merely a response to immediate operational challenges; it signifies a strategic pivot towards sustainable energy practices in the tech industry. As AI applications demand more from data centers, the focus on efficient thermal management will likely intensify, creating fertile ground for innovation and investment. The future of AI infrastructure is becoming clearer, and it is increasingly water-cooled.
