A 141-year-old Swedish company that built its reputation cooling crude oil tankers and pasteurizing milk just showed up at Data Center World in London with a product nobody in the data center industry saw coming. On March 4, Alfa Laval launched FreeWaterLoop, a fully integrated liquid-based facility cooling system designed from scratch for data centers. This is their first purpose-built product for the sector. And it tells you everything about where the cooling market is headed.
Alfa Laval is not a startup chasing hype. The company was founded in 1883 by Gustaf de Laval and Oscar Lamm, originally building centrifugal cream separators for Swedish dairies. Over the next century and a half, they became the dominant name in industrial heat exchange, selling plate heat exchangers into marine shipping, offshore oil and gas, food processing, pharmaceuticals, and power generation. Their 2025 annual revenue hit SEK 69.6 billion, roughly $6.6 billion. They employ more than 23,000 people across 100 countries. When a company this size and this old decides to build its first product for a new vertical, they are not guessing. They have done the math.
FreeWaterLoop is an external cooling system that sits on the facility loop, not at the chip or rack level. It combines Alfa Laval's plate heat exchanger technology with pump engineering from Framo, their Norwegian subsidiary that has spent more than 50 years building submerged cargo pumps used on nearly every liquid-cargo vessel afloat. Add integrated filtration and you get a single packaged system that pulls heat from the facility water loop and rejects it into a natural water source: a lake, a river, seawater, whatever is locally available. The water goes back to its origin. Net consumption approaches zero.
Thomas Moller, president of Alfa Laval's Energy Division, framed the product as a response to AI-driven thermal loads. Martijn Bergink, who runs the Ocean Division and Framo, called FreeWaterLoop "one of the fastest development cycles for a solution of this scale," a nod to the urgency the company felt once they saw the demand signals.
The competitive pitch is straightforward. Air-based cooling eats floor space. FreeWaterLoop shrinks the physical plant footprint because water carries heat roughly 3,500 times more efficiently by volume than air. That ratio matters enormously when operators are trying to pack 50 to 100 kW per rack into facilities originally designed for 5 to 10 kW.
Here is where Alfa Laval's industrial pedigree becomes a genuine moat. Their gasketed plate-and-frame heat exchangers use corrugated stainless steel plates pressed together in a stack, creating thin channels that force turbulent flow on both the hot and cold sides simultaneously. That turbulence dramatically increases the heat transfer coefficient, often reaching 3 to 5 times the values achievable in conventional shell-and-tube designs. The thin plate walls also reduce thermal resistance compared to thick tube walls. The result: a plate heat exchanger can deliver equivalent thermal performance in roughly one-fifth the physical volume of a shell-and-tube unit. Alfa Laval has been refining this technology since 1938. That is 88 years of metallurgical optimization, fouling resistance testing, and gasket engineering that no data center cooling startup can replicate by hiring a few thermal engineers and raising a Series B.
The timing here is not accidental. Data centers are under siege over water. In 2023, U.S. data centers consumed 17 billion gallons of water directly through cooling, according to Lawrence Berkeley National Laboratory. That number could quadruple by 2028. Google's data centers alone consumed 6.1 billion gallons in 2023. Data centers in Texas are projected to use 49 billion gallons in 2025, with estimates reaching 399 billion gallons by 2030.
Most evaporative cooling towers consume water and do not return it. That is the core problem. FreeWaterLoop uses a closed-loop approach where the natural water source acts as a heat sink but is returned to its origin after absorbing thermal energy. The water is not consumed, evaporated, or chemically treated. It passes through the heat exchanger and goes back. For operators facing permitting battles and community opposition over water draw, this distinction between consumption and use is the difference between getting a project approved and watching it die in review.
The data center cooling market is projected to grow from roughly $21 billion in 2026 to over $54 billion by 2034. The liquid cooling segment alone hit an estimated $6 billion in 2026. Vertiv holds about 11% market share in precision cooling. Schneider Electric, Rittal, Stulz, and Boyd round out the top five, which collectively control around 35% of the market. The rest is fragmented across dozens of smaller players and startups.
Alfa Laval is entering from a completely different direction than any of them. Vertiv and Schneider built their cooling businesses as extensions of their electrical and critical infrastructure portfolios. Alfa Laval built theirs on thermal physics. They are not learning heat exchange for the first time. They invented much of the modern plate heat exchanger industry. That gives them manufacturing scale, materials science depth, and a supply chain that already operates in 100 countries.
The question is whether they can learn the data center customer. Hyperscalers and colocation operators have specific procurement cycles, redundancy requirements, and integration expectations that differ sharply from marine or food processing. Alfa Laval will need to build a sales and engineering organization that speaks the language of PUE, WUE, Tier III, and N+1 redundancy. Fast development cycle or not, that organizational muscle takes time.
FreeWaterLoop is a signal. The data center cooling problem has gotten large enough and desperate enough that century-old industrial conglomerates are redirecting R&D budgets to chase it. Alfa Laval did not wander into this market on a whim. They saw a $50-billion-plus opportunity growing at double digits, recognized that their core competency in heat exchange maps almost perfectly onto the problem, and moved. When the heat exchange equivalent of the 1883 cream separator company decides data centers are the future, pay attention. The industrial giants are coming. They will not all succeed. But the ones carrying 88 years of plate heat exchanger R&D in their back pocket have a real shot.