Eaton closed its acquisition of Boyd Corporation's thermal division on March 12, paying $9.5 billion for a business that generated $1.7 billion in projected 2026 sales. Boyd Thermal was carved out specifically for this deal — Eaton was not buying a conglomerate. It was buying the liquid cooling hardware side of the data center market, as directly and deliberately as any acquisition in the sector's history.
The multiple is 22.5x estimated 2026 adjusted EBITDA. That number is a statement. The cooling industry is not priced like a components business anymore. It is priced like infrastructure software: recurring, mission-critical, and very hard to replace mid-deployment.
Of Boyd Thermal's projected $1.7 billion in 2026 sales, $1.5 billion is liquid cooling. Cold plates. CDUs. Thermal interface materials. Manifold assemblies. The components that sit between the chip and the rest of the cooling loop in direct-to-chip systems. Boyd's manufacturing footprint spans Asia, North America, and Europe, giving Eaton a global production base at a moment when supply chain geography is a competitive variable, not a back-office consideration.
Boyd was already integrated into several hyperscaler supply chains. The acquisition does not require Eaton to build a liquid cooling business from scratch. It requires Eaton to scale one that was already winning contracts.
Eaton's positioning for this deal is "grid-to-chip." The phrase is precise. Eaton's existing portfolio covers the power management side: UPS systems, switchgear, power distribution units, and busway. What it lacked was the thermal management hardware that handles the heat once the power has been converted and delivered to the server. Boyd fills that gap.
The operating logic is that a hyperscaler or large operator buying into liquid cooling at scale wants to reduce the number of vendors across their critical infrastructure chain. Power and thermal management are both mandatory. They share commissioning timelines, warranty periods, and service relationships. A single vendor who covers both has a structural advantage in large-scale procurement conversations, all else being equal.
Eaton's Boyd Thermal revenue now sits within its Electrical Global segment. That placement matters: Eaton is treating liquid cooling hardware as part of electrical infrastructure, not as an HVAC or mechanical add-on. That framing will show up in how it sells, how it services, and how it responds to the next Nvidia architecture change that moves the thermal requirements.
Transaction value: $9.5 billion. Multiple: 22.5x estimated 2026 adjusted EBITDA. Boyd Thermal projected 2026 sales: $1.7 billion, of which $1.5 billion is liquid cooling. Closed March 12, 2026. Boyd Thermal is now reported within Eaton's Electrical Global segment.
The Eaton deal closed nine days after Trane Technologies completed its acquisition of LiquidStack on March 3. Seven days after Eaton closed, Ecolab announced its $4.75 billion agreement to acquire CoolIT Systems from KKR. Three transactions totaling roughly $19 billion, all targeting the same liquid cooling market, across eleven days.
The pace tells you something. These are not opportunistic deals. The HVAC conglomerates and industrial companies who did not have a direct-to-chip liquid cooling position a year ago have concluded, independently and almost simultaneously, that waiting is no longer viable. Trane's LiquidStack acquisition was framed as a hedge against Nvidia's Vera Rubin architecture, which can run at 45°C inlet water temperature with no chiller. Eaton's Boyd deal was framed as a portfolio completion play. Ecolab's CoolIT deal was framed as a market expansion. The frames are different. The urgency is the same.
The market that existed two years ago, where liquid cooling was a specialty product built by specialty vendors, is gone. The buyers are now industrial conglomerates with balance sheets in the tens of billions, global manufacturing capacity, and existing relationships with the hyperscalers who are building out at 100 MW and above. Specialty vendors who did not get acquired are now competing against divisions of companies like Eaton and Trane.
Boyd's manufacturing geography creates a tariff exposure question that Eaton will be managing through the second half of 2026. Boyd operates in Asia — where Trump's reciprocal tariffs range from 24% (Japan, Malaysia) to 145% (China) depending on the production location. North American and European manufacturing within Boyd's footprint gets favorable treatment. Asian production does not.
Eaton has not disclosed the specific geographic breakdown of Boyd's manufacturing capacity or what percentage of its cold plate and CDU production sits in high-tariff jurisdictions. That figure will matter. Schneider Electric disclosed that 83% of its North American COGS are not imported, giving it a cost structure that can survive the current tariff environment without major restructuring. Eaton's Boyd division will face the same question from procurement teams within months of closing.
Eaton is now a credible end-to-end option for a hyperscaler or large colocation operator who wants power and thermal management from a single vendor with a global service network. That was not true six months ago. The question for the market is whether integrated vendors convert that positioning into contract wins at the scale of their ambition, or whether operators keep splitting the stack between best-in-class specialists.
The history of infrastructure procurement suggests both patterns coexist. Mission-critical buyers often prefer specialists for the highest-value components. They accept integrated vendor solutions for everything else. Cold plates and CDUs for the next generation of AI hardware, the components where a thermal failure at 130kW per rack is a catastrophic and expensive problem, may be exactly the category where best-in-class still wins. Eaton paid $9.5 billion on the assumption that Boyd is best-in-class, not just bundled-in.