Rear-Door Cooling vs In-Row Units: The Air Path That Wins at 100kW per Rack

A rack that draws 100kW is no longer a box with a fan; it is a local furnace that must be cooled where the air leaves it. in-row precision cooling units places a cooled unit in the row, pulling warm air and pushing cold air back into the aisle. rear-door liquid-to-air heat exchangers bolts a water-coil door onto the back of the cabinet, capturing heat at the source before it ever reaches the room. The comparison looks like two ways to move the same heat, but it is really about how far the hot air travels and who pays for moving it that far.

Above roughly 100kW per rack the cooling question stops being which unit moves heat and becomes how far the hot air travels before it is caught.

Part One: What is actually being compared

Before comparing rear-door liquid-to-air heat exchangers with in-row precision cooling units, it helps to name the unit of account. In high-density rack thermal design the tempting unit is the headline specification, because it is printed on a datasheet and can be ranked. The unit that actually appears in a budget is the sum of three things: the cost paid before the first unit of value, the cost paid on each unit, and the cost of the occasions when the system behaves badly. Once those three are written on one line, most of the specification argument stops being interesting.

The cost structure diverges in a way that a per-unit comparison hides. With rear-door liquid-to-air heat exchangers, a meaningful part of the spend is committed before the first useful result; with in-row precision cooling units, the spend arrives as a variable charge attached to each operation. A committed cost rewards volume and punishes a bad quarter, because it stays even when utilisation collapses. A variable cost is gentler operationally but removes the incentive to build spare capacity before it is needed.

Part Two: The arithmetic that decides it

The energy difference is mostly a fan problem. Moving air is expensive; moving water is cheap. in-row precision cooling units needs large airflow to overcome mixing and distance, and those fans run continuously at a fraction of total IT power. rear-door liquid-to-air heat exchangers moves the heat primarily through water, with far smaller air volumes, so the fan energy drops even though a pump and a chiller loop are added upstream. At 30kW a rack the difference is noise; at 100kW it becomes a measurable slice of the power bill.

The mechanism behind the trade is about information rather than physics. rear-door liquid-to-air heat exchangers makes one part of the problem observable and therefore cheaper to plan around, while in-row precision cooling units leaves it implicit and pays for it in variance. Whenever a design makes something measurable, somebody still has to build the instrument, keep it calibrated and act on its readings, and those three obligations are where a theoretical advantage is usually spent.

The second-order consequence is organisational rather than technical. Choosing rear-door liquid-to-air heat exchangers creates a dependency on a capability that must be staffed, monitored and kept current. Choosing in-row precision cooling units keeps that capability outside the team and converts it into a vendor relationship. Both are real costs, and the option that looks cheaper in a spreadsheet is frequently the one that adds a standing meeting to the calendar.

Part Three: Where the simple model breaks

The catch is water where you did not want it. rear-door liquid-to-air heat exchangers brings a water loop into the row and onto the cabinet, which triggers leak-detection, isolation valves and a facilities team that now owns something inside the IT space. in-row precision cooling units keeps water in a unit that is already treated like facilities equipment, so the IT side stays dry and the failure surface stays where electricians are comfortable. The safer-feeling option is not always the cheaper one, but it is often the one the insurance policy prefers.

Rear-door exchangers move heat mostly through water with far smaller air volumes, so fan energy drops even though a pump and chiller loop are added upstream.

Part Four: The cost that arrives later

There is a retrofit asymmetry that decides adoption. in-row precision cooling units slides into an existing row with little change to the cabinet, so a legacy data hall can be densified without rewiring every rack. rear-door liquid-to-air heat exchangers wants every cabinet to accept a door and a water connection, which is natural in a new build and awkward in a hall full of mismatched gear. The same technology that wins in a greenfield loses on a brownfield simply because the plumbing is not everywhere yet.

Expect a split by scale rather than by industry. Large operators can amortise the fixed side of rear-door liquid-to-air heat exchangers and will move first; smaller ones keep in-row precision cooling units because their load is spiky and their tolerance for operational complexity is low. The genuinely contested ground is the middle, where a managed service can make either option look like the other.

Part Five: What this changes outside the engineering team

For the office park down the road, the consequence is whether the new computing facility hums quietly or trips the local substation. A cooling choice that shaves a few percent off a 50MW campus is a visible reduction in trucks of diesel and in the neighbour's complaint about the heat haze. The daily-life version is an air conditioner that vents into the same room it cools versus one that dumps heat straight outside.

Outside the engineering team the consequence is about who gets access. A system that needs less of the scarce resource can be replicated by organisations that could not have afforded the original, which changes procurement, staffing and the kinds of work that get scheduled. That is a quieter but more durable effect than any single performance number.

A fair comparison holds the rack power fixed at the real density and measures cooling energy, water footprint and the retrofit cost to the existing hall. A vendor quoting a perfect new build hides the brownfield number that most operators actually face.

A market sizing sanity check

If rack densities keep climbing and a large share of new AI capacity crosses 80kW by 2029, the market for rear-door and in-row cooling, plus the water-loop and leak-detection kit around them, is plausibly $3B to $9B annually. The low end assumes densities plateau; the high end assumes most new halls are built for 100kW-plus racks. The driver is average rack power, not total square footage.

All figures here are an author's synthesis of published ranges and a stated model, not a measurement of a shipped system. The point of the arithmetic is to locate the crossover and expose which assumption drives it. A vendor quote will differ, and the direction of that difference is usually more informative than the final number.

The comparison that matters

The comparison that matters is between capturing heat at the source and catching it after it has mixed into the room. Rear-door exchangers win at high density where fan energy and air travel dominate; in-row units win on retrofit simplicity and keeping water out of the IT row. Greenfield high-density halls should plan around the door; brownfield densification often cannot.

The honest caveat is that this ranking is sensitive to one assumption: that the workload keeps its present shape. Shift the mix toward the case where in-row precision cooling units is strong and the result changes without any technology changing at all. That is why the durable output of an evaluation is not a verdict but the threshold at which the verdict would flip.

Plan the new high-density hall around the door, but densify a brownfield with in-row units, because water is not yet plumbed to every cabinet.

SOL: HCqPmPJM3NVtS6YrmVaTZYnuP8XZ2PFgQZKT94VK1LkS
Creating this takes real effort — thank you for the support.