Posts on lo0 — Tech Blog

Cooling the AI datacenter (1/4): the thermal challenge

Sat, 27 Jun 2026 20:00:00 +0200

The power series (in Spanish) followed the watt from the utility feed all the way to the chip. But there is an implacable physical law that closes the loop: almost every watt that goes into a server comes back out as heat. An inference factory does not only have to deliver megawatts to its GPUs; it has to remove those same megawatts as heat, and remove them fast enough that the silicon does not cook itself. That is the other half of physical infrastructure, and it is the half the AI era has turned upside down. This article opens the second series in the vertical: cooling.

The approach is the same one used in the power series: universal concepts, valid for any datacenter, read throughout against the question of why AI, with its 130 kW racks and its roadmap toward the megawatt, is forcing a historic transition from air to liquid.

The law that governs everything: power is heat

Start with the physical principle, because everything else follows from it. The electrical energy a server consumes does not disappear: by conservation of energy, practically all of it turns into heat that has to be evacuated. For practical purposes, the thermal power to dissipate equals the electrical power consumed. A rack that swallows 130 kW is a 130 kW radiator, the equivalent of several dozen domestic heaters switched on at once inside a single cabinet. Cooling means moving that heat from the chip to, eventually, the atmosphere, crossing a chain of media (air, water, refrigerant) as real as the electrical chain that fed it.

The ability of a fluid to carry heat obeys a simple but revealing relation: the heat a flow can remove is proportional to its mass flow rate, its heat capacity and the temperature rise it accepts,

$$\dot{Q} = \dot{m} \cdot c_p \cdot \Delta T$$

That formula alone explains why air is in trouble. Air has a low heat capacity and a minimal density, so removing a lot of heat means moving enormous quantities of it. Water, for the same volume, carries on the order of a thousand times more heat than air. When rack density spikes, sticking with air is like emptying a swimming pool with a hairdryer: technically conceivable, practically absurd.

Put numbers on it. Solving the formula above for flow rate, removing a given power with a fixed temperature rise requires $ \dot{m} = \dot{Q} / (c_p \Delta T) $. For a 50 kW rack cooled by air with a reasonable rise of about 11 °C (the usual 20 °F), the required airflow is around 7,800 cubic feet per minute: literally a gale blowing through the cabinet. Doubling density to 100 kW doubles the gale, and there is no way to contain that much moving air without the turbulence, noise and recirculation that ruin efficiency. Water does the same job with a ridiculously smaller flow and volume, because its $ \rho \cdot c_p $ product, the capacity to carry heat per unit of volume, is overwhelmingly higher. Physics leaves no room for nostalgia: above a certain density, air simply cannot.

AI density breaks the model

For decades, air was more than enough. A traditional rack dissipated between 5 and 15 kW, and a well-designed room cooled it without breaking a sweat. Air is comfortable: cheap, safe, no pipes to drip onto the electronics. The problem is that AI has multiplied density by ten.

The root of it is the chip. An H100 (Hopper) GPU dissipated about 700 W; a B200 (Blackwell) goes up to 1,000 W, and as far as 1,200 W in its full-specification version. That increase in power per chip is the first cause of the whole problem: it is not only that there are more chips, it is that each one dissipates almost twice what the previous generation did, on a silicon area that barely grows, so heat density per square centimetre spikes. An air heatsink has a physical limit to how much heat it can pull off a small surface before the surrounding air saturates; past that point, no matter how many fans are added, the chip throttles thermally and drops its clock so as not to burn, which means wasting very expensive silicon. Liquid, with direct contact and far greater capacity, pulls that heat away without the chip having to back off.

And those chips do not travel alone: the NVIDIA GB200 NVL72 packs 72 GPUs and 36 CPUs into a single rack rated at 120 kW, with real deployments measuring 130-132 kW at full load. The roadmap does not let up: the Vera Rubin generation points at 190-230 kW per rack, Rubin Ultra “Kyber” at around 600 kW in 2027, and there is already talk of the 1 MW rack toward 2028. The distant figures should be taken for what they are, vendor roadmap rather than deployed installations, but the slope is unmistakable.

Where is the limit of air? Industry figures place it with fair consensus: air is efficient up to about 10-20 kW per rack, room-level cooling reaches about 27 kW with effort, and row-level cooling about 30 kW. With advanced techniques such as rear-door heat exchangers it can be stretched to 40-60 kW. Above that, thermodynamics stops cooperating: cooling a 50 kW rack with air would take on the order of 7,800 cubic feet per minute, a gale impossible to contain inside a cabinet. Between 50 and 100 kW, liquid stops being an option and becomes the only viable physics. AI racks sit well above that boundary. That is why 2025-2026 is the moment of the great transition.

The spectrum of solutions

It helps to see cooling not as one technology but as a spectrum ordered by density. As heat per rack rises, you climb from one rung to the next:

Method	Typical density supported	Medium that touches the heat
Room air (CRAC/CRAH) + containment	up to ~30 kW/rack	Air
Rear-door heat exchanger (RDHx)	up to ~40-60 kW/rack	Air → water
Direct liquid to chip (DLC)	~50 kW to >130 kW/rack	Water/refrigerant in a cold plate
Immersion	100-250+ kW/rack	Dielectric fluid

Each rung brings the coolant closer to the hot spot. Air cools the whole room; the rear door intercepts hot air at the rack outlet; direct liquid to chip brings water to a cold plate pressed against the processor itself; and immersion submerges the entire electronics in a fluid that does not conduct electricity. The closer the liquid gets to the chip, the more heat is captured and the higher the density that can be operated. The next three articles in this series walk that spectrum in detail: air and its limits, direct liquid to chip, and immersion.

The reality of 2026, though, is not a clean replacement but coexistence. Even a direct-liquid rack still evacuates a fraction of its heat through air, between 2 % and 20 % depending on how close the liquid gets to each component, so the hybrid room, liquid for the GPUs and air for the rest, is the norm during the transition. Vendors have understood this: the most widely deployed solutions combine both worlds.

The journey of heat, end to end

Just as the power series followed the watt from the utility feed to the chip, it is worth following the heat in the opposite direction, from chip to atmosphere, because cooling is a chain with as many links as the electrical one. Heat is born in the transistor and conducts to the chip package; from there it passes to a heatsink or a cold plate; that first medium, air or liquid, carries it to a heat exchanger; a second loop, the facility water loop, takes it out of the room; and a final rejection stage, a cooling tower, a dry cooler or a chiller, hands it to the atmosphere. Every hop has a temperature difference and a loss, and the art of cooling consists of minimising the number of hops and maximising the temperature at which you work, so that less energy (and less water) is needed for the final rejection.

Here a concept appears that runs through the whole series: the coolant distribution unit (CDU), the thermal equivalent of a transformer. The CDU separates the loop that touches the equipment, clean and controlled at precise flow and temperature, from the facility loop, and exchanges heat between the two without mixing them. It is the boundary between the server vendor’s world and the building operator’s, and where a good part of the efficiency is decided. We will look at it in detail when we get to direct liquid to chip.

The metrics of heat

To govern cooling you have to measure it, and several metrics belong in an architect’s toolkit. The first is already familiar from the power series: PUE (Power Usage Effectiveness), the ratio between total facility energy and the energy that reaches IT equipment. What matters for cooling is that cooling and air movement consume between 30 % and 45 % of a traditional datacenter’s energy: by a wide margin, the largest energy overhead on top of compute. The industry average has been stuck around 1.54 for years, according to the Uptime Institute, precisely because air cooling is inefficient and hard to improve on.

The second metric, which AI has pushed to the foreground, is WUE (Water Usage Effectiveness), the litres of water consumed per kWh of IT:

$$\text{WUE} = \frac{\text{litres of water}}{\text{kWh of IT}}$$

The average sits around 1.8-1.9 litres per kWh, although the most efficient facilities aim to go below 0.2. Water is the hidden face of cooling, and we will come back to it.

The third is the ASHRAE TC9.9 framework, the reference that defines the temperature at which equipment may operate. For air, its classes run from A1 (15-32 °C inlet) to A4 (up to 45 °C); for liquid, a recent update defines water classes W17 to W45, where the number is the maximum supply temperature in degrees. The underlying trend is clear and counterintuitive: raise operating temperatures. The hotter the coolant can come in, the less it has to be chilled artificially, and the more ambient air or water can be used without compressors. Operating hot is operating efficiently.

That idea crystallises in the concept of warm-water cooling and the ambition to operate chillerless. The reasoning is elegant: if the chip is perfectly happy with 40 or 45 °C inlet water, there is no point spending energy producing 15 °C chilled water with compressors; a dry cooler rejecting heat to outside air, which is almost always below 45 °C, is enough. NVIDIA promotes exactly that, 45 °C inlet water (class W45) for its Vera Rubin generation, with a wide thermal rise (45 °C in, 65 °C return) that reduces pumping flow and, above all, eliminates chillers in most climates. The term deserves a caveat: “chillerless” usually means “almost chillerless”, because on the hottest days some mechanical cooling backup is still needed. But the direction is unmistakable, and it is the one that saves energy, saves water and, as we will see, allows heat reuse.

One measurement caveat to close on: PUE does not capture the consumption of the fans that live inside the server, which in an AI node is not negligible. Hence TUE (Total Usage Effectiveness), which combines PUE with the equipment’s own internal efficiency (ITUE) to give a more honest picture of how much energy actually reaches the transistors. As liquid removes server fans, the distinction gains importance: in an air-cooled AI node, internal fans can consume a far from negligible percentage of the server’s own power, energy that liquid recovers by making them unnecessary. It is one of the gains of liquid that PUE, on its own, cannot even see.

Water: the new constraint

If energy is the AI datacenter’s input constraint, water is becoming an equally hard output constraint. Many efficient cooling techniques, cooling towers and evaporative cooling among them, trade water for energy by evaporating water, and at AI scale that bill has become enormous. Estimates of AI datacenter water consumption already run into hundreds of billions of litres per year, with projections doubling toward 2030. The per-query figures in circulation, from a few millilitres to half a litre per request, should be handled with care, because they vary enormously with methodology, but the aggregate order of magnitude is real and concerning.

It is worth understanding the underlying technical dilemma, because it is a trade-off rather than a defect. A datacenter’s heat rejection equipment moves inside a triangle of energy, water and temperature. A cooling tower evaporates water to cool better: little energy, a lot of water (good PUE, bad WUE). A dry cooler evaporates nothing, so it barely uses water, but it can only cool to near ambient temperature, which demands operating with hotter water and, in warm climates, more backup energy. A mechanical chiller cools to any temperature but at a high energy cost. Raising operating temperatures, as noted above, is what allows the move from tower to dry cooler without an energy penalty: warm water and the waterless datacenter are two sides of the same coin. AI, by concentrating so much heat, forces a conscious choice inside that triangle where the choice used to be made by inertia.

And it is not only a technical problem but a social one. Competition for water with communities and agriculture has begun to stall projects: in the United States, “Project Blue” near Tucson, which expected to consume hundreds of millions of gallons a year, was cancelled after local opposition. The industry’s answer is twofold: on one side, waterless designs using dry coolers in exchange for operating with hotter water (Microsoft reports reaching a WUE of 0.30 and deploying zero-evaporation cooling); on the other, direct liquid to chip itself, which by capturing heat at a higher temperature makes it easier to reject to ambient without evaporating. The paradox of 2026 is that liquid, well designed, can use less water than the evaporative air it replaces.

Waste heat: from problem to resource

There is an optimistic twist. All that heat that costs so much to remove is, after all, energy, and reusing it rather than dumping it into the atmosphere is increasingly on the table. Waste heat reuse for district heating has gone from curiosity to regulatory obligation: the European energy efficiency directive pushes in that direction, and national laws already set quotas, with Germany requiring 10 % reuse in 2026 and up to 30 % in 2028.

Here liquid again has a decisive advantage over air. Modern district heating networks need water at 50-70 °C, a temperature that a liquid-cooled datacenter with its wide thermal rise (45 °C inlet and 65 °C return, for example) can deliver almost directly, whereas the hot air of a traditional room is too cool and too diffuse to be useful. It is no accident that the large reuse cases, heating tens of thousands of homes in northern Europe with Microsoft, Google and Meta datacenters wired into district heating networks, rest on liquid cooling. For an operator, waste heat can go from cost to revenue: heat delivered to a district network can be worth considerably less than the gas it displaces, with payback periods under a decade when the datacenter sits one or two kilometres from the network. At continental scale the potential is large: European datacenter waste heat is estimated to be able to cover around 10 % of EU space heating by 2030. What was vented to the atmosphere for decades is starting to be seen for what it is: wasted high-quality energy.

What is holding the transition back

If liquid is thermodynamically inevitable, why has it not swept air aside already? Because changing cooling paradigm touches everything an operator fears touching. The first brake is fear of leaks: bringing water within centimetres of the most expensive electronics on the planet makes anyone nervous, and although modern solutions use dripless quick connectors, leak detection and drip trays, the perception of risk weighs. The second is lack of standardisation: for years every vendor brought its own fluid, its own connectors and its own temperatures, which locked customers into a supplier and complicated maintenance; the sector is converging, but slowly. The third is cost and construction: liquid demands piping, CDUs, manifolds and often a rebuilt room, an investment that only high densities justify. And the fourth is operational and cultural: a team used to managing air conditioning has to learn to operate a hydraulic system, with its water chemistry, its pumps and its gaskets.

The result is that air does not disappear: it is reserved for loads that need nothing more, while liquid enters where density forces it. The question for an architect is rarely “air or liquid?” in the abstract, but “which part of my room needs liquid, when, and how does it coexist with air everywhere else?”. Retrofit, adapting existing air rooms to host liquid islands, is one of the big businesses of 2025-2026, precisely because almost nobody starts from scratch.

Where the market is in 2026

How far along is the transition? Uptime Institute data from 2025 gives a nuanced picture: direct liquid to chip is already present in around 22 % of operators, but perimeter air remains the majority at 75 %, and 61 % of those not yet using liquid are considering it. The barriers they cite are revealing, being lack of standards, cost and fear of a leak failing over the electronics, the same ones that hold back any paradigm change. The liquid cooling market, however, is growing at high double-digit rates, with estimates multiplying it by five or six over a decade, and the large players are repositioning fast: Vertiv leads with a good share of the first GB200 racks and has launched solutions combining direct liquid to chip and air in a single module; Schneider Electric bought Motivair to enter in earnest; and alongside them live cold-plate specialists (CoolIT, Boyd, JetCool, Asetek) and immersion specialists (Submer, GRC, LiquidStack). The dominant pattern is hybrid: rooms combining liquid for the GPUs and air for everything else, as a bridge to a mostly liquid future. That vendors package both worlds into a single product says a great deal about where the transition is heading: not to a clean cut, but to a managed coexistence lasting years.

For an inference factory

What should someone building an inference factory take from this? Three ideas that run through the whole series. First, that cooling is not an add-on but a co-design: the density of AI racks forces the cooling method to be decided at the same time as the power method, because a 130 kW rack cannot simply be dropped into a room designed for air. Second, that the move to liquid is no longer optional above a certain density; the question is not whether, but which, direct to chip or immersion, and how to manage coexistence with air for years. And third, that the metrics that matter have widened: watching PUE is no longer enough, water (WUE) has to be watched too and, increasingly, what is done with waste heat, because operating cost, the social licence to build and, in some geographies, legal compliance are all decided there.

It is also worth keeping in mind that cooling is not only efficiency: it is availability. As the power series showed, a thermal failure can bring a factory down as fast as a power cut, in fact faster, because a 130 kW rack without cooling reaches dangerous temperatures in seconds. That is why critical cooling is designed with the same philosophy as power: redundancy in pumps and CDUs, early detection, and testing. A cooling system nobody has exercised at full load is, once again, an unverified promise.

This article has drawn the map of the thermal challenge. The next three go down into the detail of each rung of the spectrum: starting with air cooling and its limits, the point everyone is coming from, moving on to the direct liquid to chip that is feeding GB200 racks, and finishing with immersion and the horizon of the megawatt rack. Feeding the GPUs was half the battle; the other half is keeping them from melting.

Sources

Network World, Why AI rack densities make liquid cooling nonnegotiable — https://www.networkworld.com/article/4149069/why-ai-rack-densities-make-liquid-cooling-nonnegotiable.html
Introl, Liquid Cooling vs Air: The 50kW GPU Rack Guide (2025) — https://introl.com/blog/liquid-cooling-gpu-data-centers-50kw-thermal-limits-guide
ToneCooling, NVIDIA GB200 NVL72 Cooling Requirements — https://tonecooling.com/nvidia-gb200-nvl72-cooling-requirements/
TweakTown, NVIDIA full-spec Blackwell B200 uses 1200W — https://www.tweaktown.com/news/97059/nvidias-full-spec-blackwell-b200-ai-gpu-uses-1200w-of-power-up-from-700w-on-hopper-h100/index.html
DCD, Nvidia’s Rubin Ultra NVL576 rack expected to be 600kW — https://www.datacenterdynamics.com/en/news/nvidias-rubin-ultra-nvl576-rack-expected-to-be-600kw-coming-second-half-of-2027/
Computer Weekly, Huge grid and heat challenges ahead as Nvidia set for 1MW rack — https://www.computerweekly.com/news/366639658/Huge-grid-and-heat-challenges-ahead-as-Nvidia-set-for-1MW-rack
mgrid / Uptime Institute, Global Data Center PUE Stalls at 1.54 (2025) — https://mgrid.org/2025/10/01/uptime-institute-data-center-pue-stagnation-2025-liquid-cooling/
ClearComfort, 2026 PUE and WUE Checklist — https://clearcomfort.com/pue-wue-ai-data-centers/
Introl, Water Usage Effectiveness (WUE) AI Data Center Cooling Guide 2025 — https://introl.com/blog/water-usage-efficiency-wue-ai-data-center-cooling-guide-2025
CKY, ASHRAE TC 9.9 Thermal Guidelines (5th Ed.) — https://www.cky.com.tw/en/insights/ashrae-tc9-datacenter-thermal-guidelines
DCD, Hot water, cold water: the right temperature for water cooling — https://www.datacenterdynamics.com/en/analysis/hot-water-cold-water/
EESI, Data Centers and Water Consumption — https://www.eesi.org/articles/view/data-centers-and-water-consumption
Microsoft Datacenters, Measuring energy and water efficiency — https://datacenters.microsoft.com/sustainability/efficiency/
Energy Solutions Intelligence, Data Center Waste Heat Recovery 2026 — https://energy-solutions.co/articles/sub/data-center-waste-heat-district-heating
NVIDIA Blog, Hotter Than a Hot Tub: 45°C liquid cooling for AI factories — https://blogs.nvidia.com/blog/liquid-cooling-ai-factories/
Uptime Institute, Global Data Center Survey 2025 — https://uptimeinstitute.com/resources/research-and-reports/uptime-institute-global-data-center-survey-results-2025

Cooling the AI datacenter (2/4): air and its limits

Sat, 27 Jun 2026 19:00:00 +0200

The first article drew the map of the thermal challenge and laid out the spectrum of solutions. Now for the bottom rung, the one almost everybody already has installed: air cooling. There is a temptation to treat it as legacy technology, a leftover of the pre-AI datacenter that liquid will sweep away in a couple of years. The data says otherwise. The Uptime Institute’s 2025 cooling systems survey puts perimeter air, the classic room CRAC or CRAH, in 75 % of installations, against 22 % reporting direct liquid cooling. Air is the base case, and it will remain so across most of the installed base for years.

This article has two goals. The first is to understand air well enough to squeeze it: where the energy goes, which measures pay back, and how much headroom is left in a poorly tuned room. The second is to understand exactly why it breaks, because the limit of air is not set by a liquid-cooling vendor’s catalogue: it is set by room geometry and by an exponent.

Inside the air-cooled room

The equipment that cools a traditional room comes in two variants that are often confused. A CRAC (Computer Room Air Conditioner) is a direct-expansion machine: it carries its own compressor and its own refrigerant circuit, and runs autonomously without depending on anything external. A CRAH (Computer Room Air Handler) has no compressor: it receives chilled water from a central plant and passes it through a coil while the fan moves room air. The difference matters for two reasons. The CRAH is usually more efficient, because compression work is concentrated in large, well-optimised chillers instead of being spread across dozens of small machines. And above all, the CRAH integrates naturally with an economiser: if the water can be cooled with outside air, the whole room runs without compressors. The CRAC needs its compressor whatever the weather.

Behind those units sits the final heat rejection equipment, and there the picture of the installed base is revealing. According to the same Uptime survey, rejection splits between air-cooled chillers (35 %), evaporative cooling towers (22 %), direct expansion (11 %), dry coolers and fluid coolers (6 %), free cooling with outside air (4 %) and adiabatic misting (3 %). It is a base dominated by the mechanical chiller, which is exactly the piece modern design tries to keep switched off for as many hours as possible.

Between the room and the rack there is an intermediate step, close-coupled cooling: in-row units placed between cabinets, shortening the air path and raising the density that can be handled. A typical row unit today, such as the Vertiv CoolPhase Row range, moves 30 to 40 kW per unit. Some 32 % of surveyed installations already have some form of this cooling, almost always coexisting with perimeter air.

The geometry of airflow

The formula that governs any fluid cooling system already appeared in the previous article:

$$\dot{Q} = \dot{m} \cdot c_p \cdot \Delta T$$

Solved for air with a realistic server temperature rise, it yields a rule of thumb every room designer knows: roughly 140 to 160 CFM per kW dissipated. Schneider works through an example with an 18 kW rack needing 2,500 CFM (1,180 l/s) at the inlet; an air-cooled DGX B200 asks for 1,550 CFM on its own. Those figures are not large in the abstract. They become a problem when crossed with the other half of the system, which is where that air comes in.

A perforated raised-floor tile delivers around 300 CFM (142 l/s). A grate-type tile reaches 700 CFM. Going beyond 500 CFM per tile takes special pieces and a very carefully designed plenum, because pressure under the floor is not infinite and tiles compete with each other. That is where the most quoted and least disputed limit of classic air comes from: a conventional distribution with one vented tile per rack does not sustain much more than 6 kW per rack. With containment and serious airflow management you reach around 10 kW sustained, and highly optimised designs stretch into the twenties.

The arithmetic of an AI rack breaks that beyond argument. At 140 CFM/kW, a 50 kW rack asks for about 7,000 CFM. With standard 300 CFM tiles that would take twenty-three of them under a cabinet whose footprint is worth two. With 700 CFM grates, ten. There is no floor area through which to push that flow, nor plenum pressure to distribute it. Air fails on geometry before it fails on thermodynamics: the room runs out of places to deliver the fluid long before the fluid runs out of capacity to absorb heat.

And there is a second, subtler problem that shows up when someone tries to force the machine.

The cube law, friend and executioner

The power a fan draws grows with the cube of its rotational speed. The US Department of Energy illustrates it with a direct case: a fan drawing 6.5 kW at full speed drops to 3.3 kW at 80 % of revolutions, because $ 0.8^3 \approx 0.51 $. The same relation seen from the air circuit says the required pressure grows with the square of the flow, $ P = Z \cdot Q^2 $, so doubling the flow quadruples the pressure and multiplies fan power by eight.

That exponent hands out very unequal prizes. At low density it is the greatest gift air has, and the measured numbers back it up. Variable-speed fan retrofits documented by the DOE at three facilities cut cooling system consumption by 22 % to 32 %, with payback under two years. The LBNL demonstration at a Digital Realty site in El Segundo, with 72 CRAHs and about 2.6 MW of IT load, is even more emphatic: swapping scroll fans for plug fans dropped CRAH power from 439 to 233 kW, a 47 % cut; adding intelligent speed control brought it to 147 kW. The combined effect was a 66 % reduction in CRAH-associated power, 2.9 million kWh a year, with payback again under two years.

At high density that same exponent presents the bill. Every attempt to compensate for missing airflow by raising revolutions is paid at the cube, and not only in room units. The same happens inside the server, and there Uptime has a clean data point from HPE: two single-socket servers of practically identical configuration, one in 1U and one in 2U, differ by 30 to 60 W of consumption, and at low utilisation the 1U burns around 30 % more power to do the same work. The cause is geometric: a 1U chassis forces small fans that have to spin fast. In an air-cooled AI node, where fans in high-performance systems can take 10 % to 20 % of system power, that penalty stops being anecdotal and becomes a cost line.

Containment: cold aisle or hot aisle

Before buying anything, almost every room has headroom in airflow management. The enemy goes by two names: recirculation, hot exhaust air finding its way back to server inlets, and bypass, cold air reaching the return without passing through any equipment. Without airflow management, the installed cooling capacity needed can reach two or three times what theory requires, because the designer compensates for poor mixing by lowering the setpoint of the whole room.

The metrics to diagnose it have existed for years and remain underused. The RCI (Rack Cooling Index) measures how effectively inlet temperatures stay within specification, with two separate indices for the high and low ends; a well-designed room should be at 90 % or above. The RTI (Return Temperature Index) compares the temperature rise of the cooling units with that of the IT equipment and separates the two ills: below 100 % there is bypass, above it there is recirculation. A couple of measurement sessions and a handful of blanking panels in the rack gaps recover more efficiency than any catalogue promises.

The next step is physical containment, and here the choice between containing the cold aisle (CACS) or the hot one (HACS) has a consequence that gets overlooked. Schneider’s model of a 700 kW, 100-rack facility in Chicago, comparing both under the realistic condition that the uncontained zone stays at 24 °C so staff can work, shows a large difference: hot-aisle containment consumes 43 % less cooling energy and drops annualised PUE from 1.98 to 1.69. The reason lies in the setpoint, not in the physics of the exchange. Contain the heat and the whole room lives at 24 °C, equipment draws air at 24 °C and the economiser can run 5,319 hours a year. Contain the cold and the room is the hot aisle, and keeping it habitable at 24 °C forces supply air at 10 °C, which leaves the economiser at zero hours.

Figures published by ENERGY STAR point the same way with similar orders of magnitude: hot and cold aisle layout without containment already saves 10 % to 35 %; cold-aisle containment reaches up to a 30 % reduction in cooling cost; hot-aisle containment adds around 40 % on top of that. Typical payback on containment is under two years.

The thermal envelope and class H1

All of the above rests on a prior question: at what temperature can the equipment work. ASHRAE TC 9.9 answers with its air classes, and the ranges are less demanding than many rooms actually practise.

Class	Recommended	Allowable	Maximum dew point
A1	18-27 °C	15-32 °C	17 °C
A2	18-27 °C	10-35 °C	21 °C
A3	18-27 °C	5-40 °C	24 °C
A4	18-27 °C	5-45 °C	24 °C
H1	18-22 °C	18-25 °C	17 °C

The recommended range is identical for every A class: 18 to 27 °C equipment inlet temperature. What changes between classes is how far you can stray from it without losing warranty. A decade of PUE improvement was built on that table, raising setpoints and widening the hours in which outside air is enough.

The fifth edition of the guidelines, published in 2021, added a class that deserves attention: H1, intended for high-density air-cooled equipment. Its recommended range is 18 to 22 °C and its allowable limit drops to 25 °C, against A1’s 32 °C. ASHRAE’s stated reason is physical and direct: in some dense systems there is no room for heatsinks or fans capable of keeping components within their limit if the incoming air is warm. Class H1 is the formal acknowledgement, by the body that wrote the thermal envelope, that high density on air no longer fits inside it.

The efficiency consequence is the part that interests an architect. Going from an allowable of 32 °C to one of 25 °C, and from a recommended of 27 °C to 22 °C, destroys much of the annual economiser hour budget. Facilities designed around free cooling and adiabatic cooling do not sit well in that envelope, and Uptime says so without hedging: the way out is to deploy liquid cooling for the dense part. Air does not run out of thermal capacity when density rises. It runs out of efficient thermal capacity, which is what made the model attractive.

Free cooling and a climate on the move

What the economiser is worth depends on where you are. Vertiv’s analysis of a 1 MW datacenter gives three figures per city that serve as a quick reference:

City	Air-cooled chiller	Free cooling	Adiabatic free cooling
London	pPUE 1.21	pPUE 1.09	pPUE 1.06
Madrid	pPUE 1.22	pPUE 1.12	pPUE 1.07
Dubai	pPUE 1.31	pPUE 1.24	pPUE 1.18

In annual operating modes, London spends 81 % of the year in free cooling or adiabatic free cooling, and Madrid 74 %, which is on the order of 7,100 and 6,480 hours without mechanical compression. Adiabatic savings against a standard air-cooled chiller at full load land between 60 % and 65 %.

That window is closing, and it has been measured. A paper published in Scientific Reports in June 2026 quantifies out-of-specification hours (inlet above 27 °C, or dew point above 15 °C, or relative humidity above 70 %) for the 1980-2024 period and projects to 2050. Northern Europe remains comfortable, with about two hours a day on average out of range. The Mediterranean and the Middle East add more than four hours a day per decade. Dallas-Fort Worth exceeds 20 % of the year; Singapore, 85 %. The fraction of sites in the optimal tier falls 6 % per decade while the most constrained tier grows 9.1 %. For a design with a fifteen-year life in southern Europe, the 2026 economiser hour budget is not the 2040 one.

What it costs in water

Air has a cheap route to better energy efficiency: evaporate water. And there lies one of the most uncomfortable trade-offs in physical infrastructure. Vertiv’s numbers settle it in a single sentence. Adding an adiabatic misting system on top of an air-cooled chiller improved pPUE by 0.01 points and raised WUE from 0 to 0.165 gallons per kWh (about 0.62 l/kWh). At a 3 MW facility that is 4.3 million gallons a year, on the order of 16,000 m³, the consumption of about forty homes. Indirect evaporative cooling sits around one litre per kWh. An air-cooled chiller without adiabatic assist, or a pumped-refrigerant direct-expansion system, has zero WUE and pPUE of 1.12 to 1.13.

One hundredth of a pPUE point in exchange for four million gallons a year was for years an obvious trade in water’s favour, because energy was paid for and water almost was not. In 2026, with moratoria and public opposition around several projects, that sum no longer always works out, and in some jurisdictions it is not even legal. The reference case for evaporative air done well is still Prineville, where Meta reported a WUE of 0.22 l/kWh with 100 % outside-air economisation, direct evaporative misting and evaporation efficiency around 85 %, with no chillers and no towers.

The ceiling of air, and who reaches it

Well-designed air goes a long way. Google reports a fleet PUE of 1.09, and its best site, in central Ohio, sustains 1.04. Yahoo’s Computing Coop in Lockport was designed for a PUE of 1.08 using ambient air 99 % of the time, with the roof acting as a chimney and cooling dropping from 25 % to 1 % of total power, although that figure is design rather than measurement. Kyoto-type thermal wheels, with 4.6 metre rotors turning at 1 to 6 rpm and around 300 kW per unit, have taken whole facilities in Chicago to pPUE 1.19 with up to ten months of free cooling a year.

Against that, the industry average measured by Uptime has been stuck around 1.54 for six years: 1.59 in 2020, 1.55 in 2022, 1.54 in 2025. The gap between 1.09 and 1.54 is not technological. It is about design, scale and the ability to build from scratch in the right place. Air has a very high efficiency ceiling that almost nobody reaches, and that is the best news for an operator with an existing room: there is nearly always headroom before touching infrastructure.

Rear doors: the last rung

Between room air and liquid on the chip there is a step that has become the big retrofit product of 2025 and 2026: the rear-door heat exchanger, or RDHx. The idea is simple. Instead of cooling the room, hot air is intercepted right at the rack outlet, its heat removed with water, and it is returned to the room at the temperature it came in at. From the rest of the facility’s point of view, that 60 kW rack does not exist thermally. This is what vendors call room neutrality, and it explains why a rear door forces neither a containment redesign nor a plenum recalculation.

There are two families. The passive kind consumes no electricity: it uses the push of the server’s own fans to cross the coil, and sits in the lower capacity band. The active kind adds variable-speed fans in the door, compensating the coil’s pressure drop and responding better to hot spots, at the cost of its own consumption. Capacities declared by serious vendors converge on a reasonable band: Vertiv announces up to 80 kW per rack in its CoolLoop, Motivair up to 75 kW in the ChilledDoor, nVent up to 78 kW with 14 °C water and a maximum draw of 1,500 W per door. Other catalogues publish figures of up to 200 kW per cabinet without specifying water temperature or flow, and those should be read for what they are. The defensible band today is 40 to 80 kW per rack.

The detail that makes the rear door attractive is water temperature. It works with warm water, on the order of 14 °C, well above the 7 °C of a classic chiller plant, which widens free cooling hours instead of reducing them. And it deploys rack by rack, with no general construction and no room shutdown. Adoption reflects this: the 2026 AFCOM report puts operators who have deployed or are evaluating rear doors at 37 %, and Schneider Electric’s purchase of Motivair confirms that the large vendors see them as a strategic product.

The real density of the installed base

Public conversation revolves around 130 kW racks. The real density distribution of the global installed base, measured by Uptime in 2025, tells another story:

Average rack density	Facilities
1-3 kW	16 %
4-5 kW	30 %
6-7 kW	18 %
8-9 kW	10 %
10-14 kW	14 %
15-19 kW	5 %
20-29 kW	4 %
30 kW or more	4 %

Some 74 % of facilities have an average density below 10 kW per rack, and 82 % do not exceed 30 kW even in their densest cabinet. Asked about the threshold beyond which air stops being enough, operators give a modal answer of 40-49 kW (25 % of responses), and 44 % put it at 40 kW or above, well past the “over 20 kW already needs liquid” repeated in presentations. Uptime marks the impracticable boundary at around 50 kW per rack.

From that comes the correct reading of the moment: air is not being replaced, it is being segmented. The training and dense inference frontier is moving to liquid, and the headline goes with it. The bulk of the installed base, still below 30 kW, is solved with optimised air, containment and, when a GPU island appears, a rear door. The useful question for an operator is not when air dies, but which density percentile their room sits in.

For an inference factory

Three concrete decisions follow from all of the above.

The first is to exhaust airflow management before buying hardware. Measuring RCI and RTI, closing rack gaps with blanking panels, matching tiles to the real load distribution and raising the setpoint to the top of ASHRAE’s recommended range costs little and frees capacity that is already installed. The ventilation retrofits measured by DOE and LBNL, with 22 % to 66 % reductions in cooling power and payback under two years, are the best return available in an existing room.

The second is to choose hot-aisle containment if the goal is the economiser, and to design the system around the highest water temperature the equipment tolerates. The decision chain runs in that order: the setpoint sets free cooling hours, free cooling hours set the PUE and, in many geographies, the absence of compression is what allows the evaporative tower and its water consumption to be dropped. Class H1 pushes the other way for dense equipment, and that tension is what decides when liquid becomes due.

The third is to size by percentiles rather than by the hottest rack in the catalogue. A room about to host two 40 kW GPU racks alongside forty 6 kW cabinets does not need rebuilding: it needs rear doors on those two and serious airflow management on the rest. Moving to liquid makes sense when sustained density forces it, and that moment arrives sooner through efficiency than through capacity.

One idea runs through the rest of the series. Even when a whole room moves to direct liquid to chip, 2 % to 20 % of the heat still leaves through air, and somebody has to remove it. Air does not disappear from the AI datacenter. It changes role: it stops being the cooling system and becomes the system that picks up what the liquid does not capture. The next article goes into that liquid, into the cold plate that carries it to the silicon, and into the CDU that governs the loop.

Sources

Uptime Intelligence, Data Center Cooling Systems Survey 2025 (UI Field Report 181) — https://intelligence.uptimeinstitute.com/sites/default/files/2025-07/UI%20Field%20181_Data%20center%20cooling.pdf
Uptime Institute, Global Data Center Survey 2025 — https://datacenter.uptimeinstitute.com/rs/711-RIA-145/images/2025.Annual.Survey.Report.pdf
Uptime Institute Journal, New ASHRAE guidelines challenge efficiency drive — https://journal.uptimeinstitute.com/new-ashrae-guidelines-challenge-efficiency-drive/
Uptime Institute Journal, Lower density brings server efficiency and cooling gains — https://journal.uptimeinstitute.com/lower-density-brings-server-efficiency-and-cooling-gains/
Upsite Technologies, Major Changes to ASHRAE’s Fifth Edition of Thermal Guidelines, Part 2: New Air-Cooled Class for High Density Compute Equipment — https://www.upsite.com/blog/major-changes-to-ashraes-fifth-edition-of-thermal-guidelines-part-2-new-air-cooled-class-for-high-density-compute-equipment/
LBNL / DOE Center of Expertise, Thermal Guidelines and Temperature Measurements in Data Centers — https://datacenters.lbl.gov/sites/default/files/FINAL%20Thermal%20Guidelines%20and%20Temp%20Measurements%209-15-2020.pdf
US DOE / FEMP, Variable-Speed Fan Retrofits for Computer-Room Air Conditioners — https://www.energy.gov/sites/prod/files/2013/10/f3/dc_fancasestudy.pdf
LBNL, Demonstration of Intelligent Control and Fan Improvements in Computer Room Air Handlers — https://eta-publications.lbl.gov/sites/default/files/control-and-fan-improvements-in-crahs.pdf
Schneider Electric / APC, Impact of Hot and Cold Aisle Containment on Data Center Temperature and Efficiency (WP 135) — https://docs.media.bitpipe.com/io_10x/io_105811/item_561497/DBOY-7EDLE8_R2_EN.pdf
Schneider Electric / APC, Cooling Strategies for Ultra-High Density Racks and Blade Servers — https://www.facilitiesnet.com/whitepapers/pdfs/20130227-SchneiderElectric.pdf
ENERGY STAR (US EPA), Utilize Containment / Enclosures — https://www.energystar.gov/products/data_center_equipment/16-more-ways-cut-energy-waste-data-center/containmentenclosures
Scientific Reports (Nature), Limitations to air free cooling in data centers under rising heat and humidity — https://www.nature.com/articles/s41598-026-56926-3
Vertiv, Freecooling, Evaporative and Adiabatic Cooling Technologies in Data Center Applications — https://www.vertiv.com/48ee5c/globalassets/products/thermal-management/free-cooling-chillers/freecooling-evaporative-and-adiabatic-cooling-technologies-in-data-center-applications.pdf
Vertiv, Calculating the Impact of Water Usage on Data Center Costs and Sustainability — https://www.vertiv.com/4ad73a/globalassets/documents/white-papers/vertiv-water-usage-and-sustainability-wp-en-na-web_317594_0.pdf
Vertiv, How rear door heat exchangers (RDHx) support high-density rack cooling — https://www.vertiv.com/en-asia/insights/articles/educational-articles/how-rear-door-heat-exchangers-rdhx-support-high-density-rack-cooling/
Motivair, ChilledDoor Rear Door Heat Exchanger — https://www.motivaircorp.com/products/chilleddoor/
DatacenterDynamics, New rear door coolers from nVent for high-density racks — https://www.datacenterdynamics.com/en/product-news/new-rear-door-coolers-from-nvent-to-offer-a-scalable-solution-for-high-density-racks/
Google, Power usage effectiveness, Google Data Centers — https://datacenters.google/efficiency/
Open Compute Project, Water Efficiency at Facebook’s Prineville Data Center — https://www.opencompute.org/blog/water-efficiency-at-facebooks-prineville-data-center
US DOE / EERE, Yahoo! Compute Coop: Next Generation Passive Cooling Design for Data Centers — https://www1.eere.energy.gov/manufacturing/datacenters/pdfs/yahoo_passive_cooling.pdf
Data Center Frontier, KyotoCooling Gains Traction With Service Providers — https://www.datacenterfrontier.com/cooling/article/11431141/kyotocooling-gains-traction-with-service-providers
Upsite Technologies, The State of Data Center Cooling in 2026 — https://www.upsite.com/blog/the-state-of-data-center-cooling-in-2026/
Data Center Knowledge, AI Rack Density’s Real Limits: Power, Cooling, Failure Risk — https://www.datacenterknowledge.com/ai-data-centers/ai-rack-density-s-real-limits-power-cooling-failure-risk

Cooling the AI datacenter (3/4): direct liquid to chip

Sat, 27 Jun 2026 18:00:00 +0200

The previous article ended with air exhausting its headroom on geometry: there is no floor through which to push seven thousand cubic feet per minute into a cabinet with a two-tile footprint. The way out is to bring the fluid to the hot spot until it touches the chip package itself. That is direct-to-chip liquid cooling, DLC, and it is the technology feeding the GB200 and GB300 racks shipping today.

The figure that justifies the change is textbook: water carries on the order of 4,000 times more heat per unit volume than air and has around 30 times the thermal conductivity. What follows is the engineering between that statement and a room in production, which turns out to be considerably richer than “add a pipe”.

Inside the cold plate

The part that touches the silicon is the cold plate: a copper block with an internal cavity crossed by microchannels, resting on the processor package through a thermal interface material. The fluid enters cold, crosses the channels absorbing heat by forced convection, and leaves hot. Typical geometry for a plate serving a processor of around 2,000 W works with fins and channels 0.15 to 0.25 mm wide, channel heights of 1 to 2 mm, and a copper base of about 2.5 mm.

Its figure of merit is thermal resistance, defined by the OCP specification as the ratio between the package temperature rise over the fluid inlet and the applied power:

$$R = \frac{T_c - T_L}{Q}$$

OCP insists on a nuance with practical consequences: resistance must be characterised as a curve against flow rate, not as a loose catalogue number. Published bench tests with commercial loops give values on the order of 0.02 °C/W at 550 W, with pressure drops of about 6.5 psi in the best of the three loops tested. That same bench produced a result anyone designing a rack should keep in mind: most of the circuit’s pressure drop came not from the cold plate but from the quick disconnects.

The cold plate is also where material constraints concentrate. Fluid velocity inside the microchannels must stay below 1.5 m/s to avoid erosion. The electrochemical potential difference between any pair of metals in contact inside the plate must not exceed 0.15 V, a far stricter limit than for the rest of the circuit, because slow galvanic corrosion in a 0.2 mm channel blocks it. The mechanical qualification OCP demands includes hydrostatic testing at three times maximum operating pressure, a pressure-drop test with less than 0.5 % loss, salt spray per ASTM B117, and internal X-ray inspection to find voids and manufacturing debris. Fin pitch on commercial plates runs around 50 microns and tends to narrow, which directly sets the loop’s filtration requirement.

How much heat the plate captures

Here is the nuance that separates the brochure from the engineering. A cold plate cools the component it covers, and a server has many more components than the GPU. The fraction of heat the liquid removes depends on how many of them are covered.

Experimental tests on real racks give capture ratios of 94 % on a 53 kW rack and 93 % on a 128 kW multi-rack assembly. Vendor deployment guides are considerably more conservative: Vertiv talks of 70 to 75 % in cold plates and 25 to 30 % left to air. Reference designs sit in the middle band: OCP’s model for 125 kW racks assumes 80/20, the Vertiv 360AI design for a 1.2 MW pod splits 76/24, and Chilldyne’s design for an NVL72 declares 72/28, which leaves 320 kW of residual air load in the room. At the high end, Supermicro announces up to 98 % capture per rack on its DLC-2 platform, and achieves it the only way possible: by putting cold plates on memory, PCIe switches and voltage regulators too, not only on CPU and GPU.

That residual fraction has two design consequences. The first is that the air room still exists, with its containment and its perimeter equipment, sized for 5 or 25 % of the heat depending on how complete the cold-plate solution is. The second is economic: internal fans in a high-performance server consume 10 % to 20 % of system power, and that line item only disappears entirely when the need to move air through the chassis disappears. A DLC capturing 72 % leaves the fans spinning, slower but spinning. The difference between 72 % and 95 % capture is not a nuance of thermal efficiency: it is the difference between paying that bill and not paying it.

The two loops

A direct-to-chip liquid system always has two separate hydraulic circuits whose fluids never mix.

The TCS (Technology Cooling System) is the secondary loop: it runs from the CDU to the cold plates, through manifolds and quick disconnects. It is a closed circuit of known volume with a controlled fluid, and it is the one that touches the IT equipment. The FWS (Facility Water System) is the building’s primary loop: it carries water from the heat rejection equipment to the CDU. Between them sits a plate heat exchanger inside the CDU, transferring heat without transferring fluid.

That separation is what makes the system governable. The TCS can have its own chemistry, its own pressure and its own filtration, far finer than will ever be reasonable in a building’s pipework. And because the exchange is never perfect, the TCS is always hotter than the FWS, by a difference that is one of the numbers most conditioning the design of the whole facility.

The CDU: the thermal transformer

The coolant distribution unit (CDU) is the boundary between the server vendor’s world and the building operator’s. It holds the heat exchanger, the pumps, the filtration, the buffer tank and the control. There are two architectures and three locations.

By architecture, a liquid-to-liquid (L2L) CDU rejects heat to building water, and is what gets installed where there is pipework. A liquid-to-air (L2A) CDU rejects it to the room’s own air, with no need to connect anything to the hydraulic installation, in exchange for returning all that heat to the existing air system. The usual sizing rule places L2A in deployments of one to ten racks and recommends L2L above 500 kW. By location, they come in-rack, in-row and perimeter.

Commercial capacities today span three orders of magnitude. Vertiv’s range runs from 70 kW in L2A format to 2,300 kW in L2L. The Liebert XDU1350 declares 1,368 kW nominal at 4 °C approach and up to 2,912 kW at 8 °C, with a nominal flow of 1,200 l/min in N+1 two-pump mode and triple-redundant 50 micron filtration cleanable in line. The CoolIT CHx2000 reaches 2,000 kW at 5 °C approach and serves twelve GB300 NVL72 racks from a single unit. At the hyperscale end, Google’s Project Deschutes CDU, contributed to OCP in 2025, is a 2 MW unit with a 3 °C approach, 80 psi of available pressure and 0.2 micron side-stream filtration. Motivair, now inside Schneider Electric, announces architectures scaling to 10 MW.

Of all those numbers, the one that governs the whole facility is the approach temperature: how many degrees separate the building’s inlet water from the coolant leaving toward the equipment. The state of the art is between 2 and 4 °C, and some vendors offer 6. The consequence is direct and often overlooked at purchase time: the lower the CDU’s approach, the higher the building water setpoint can be, and the higher that setpoint, the more hours a year the dry cooler runs without compressor help. Two degrees of approach translate into weeks of free cooling.

Redundancy follows the logic already seen in the electrical chain. Pumps are configured N+1 with automatic changeover, with isolation valves and quick disconnects so they can be replaced without draining the circuit. And there is a failure-domain rule that deserves a place in any specification: size the CDUs so that each one serves less than 10-20 % of the liquid-cooled servers, so a loop failure does not take out the whole room. The CDU also keeps secondary supply above the room’s dew point, which guarantees purely sensible cooling and avoids condensation on the electronics.

Flow, temperature rise and the price of warm water

The flow specification almost every design works with is 1.5 l/min per kW, with the server side sitting between 1.2 and 1.5. The design temperature rise of the TCS loop is around 10 °C for GPUs, although full-load tests with 128 kW racks have measured rises of nearly 15 °C. A 1 MW pod split across twenty 50 kW racks asks for on the order of 1,100 l/min in total, a figure that gives a sense of pipe sizing and pump size.

On that basis comes the most interesting argument of the moment. NVIDIA has set as a platform target that all MGX racks operate with 45 °C inlet water, and describes the full free-cooling scenario: the facility delivers 41 °C to the CDUs and these supply 45 °C to the racks. The argument is sound, because at that temperature almost any climate allows heat rejection with dry coolers, without mechanical compression, and the electrical budget freed up translates into more racks per contracted megawatt.

The price of that decision is in the hydraulics, and platform data for the GB300 quantifies it:

Supply temperature	Flow per rack	Pressure drop
25 °C	59 l/min	2.3 psi
30 °C	71 l/min	3.2 psi
35 °C	89 l/min	4.9 psi
40 °C	119 l/min	8.5 psi
45 °C	177 l/min	18.4 psi

Going from 25 to 45 °C triples the flow and multiplies pressure drop by eight, paid for in pumping, manifold sizing and CDU size. Which is why OCP defends a different position in its roadmap: 30 °C coolant as a durable minimum, arguing that at that temperature free cooling or mixed mode already covers almost the whole year in many climate zones, that fifteen years of PUE improvement are preserved, and that heat reuse options stay open. The same document warns that dropping to 20 °C raises annual chiller power by around 20 %, and that even at 30 °C a chillerless facility cannot be promised in every scenario, because capacity has to be sustained at full load and under fault conditions.

The two positions contradict each other less than it seems. NVIDIA specifies the upper limit the equipment tolerates; OCP recommends the operating point that optimises the whole. A reasonable design in southern Europe sits in the 30 to 35 °C supply band, with hydraulic capacity held in reserve to go higher if climate or tariff advise it.

Manifolds and connectors

Between the CDU and the cold plates lies a two-level distribution network. Row manifolds join the CDU to the racks and mount overhead or under the raised floor. Rack manifolds distribute inside the cabinet: vertical ones mount at the rear, horizontal ones can go at the front. NVIDIA already integrates tray manifolds, rack manifolds with UQD08 connectors and liquid-cooled busbars up to 5,000 A into the Vera Rubin platform, a sign that the boundary between electrical and thermal distribution is blurring inside the cabinet.

The part that makes all this maintainable is the dripless quick disconnect. OCP maintains two separate specifications: UQD for manual connection and UQDB for blind mate, the one that lets a compute tray connect hydraulically as it slides into the rack with nobody screwing anything. The UQDB test methodology validates interchangeability between suppliers with a full millimetre of radial misalignment, and vendors declare full flow with gaps of one to two millimetres. Spillage on disconnecting a reference flat-face connector is below 0.015 cc, on the order of one drop. Typical working ranges reach 150 psi with burst pressure four times higher and lives of more than 10,000 cycles.

That an open connector specification exists matters more than it seems. For years every vendor brought its own, which tied the customer to one supplier for the room’s whole life. The connector is the piece of plumbing that decides whether another vendor’s rack fits in your row.

The chemistry of the loop

The TCS fluid is not plain water. The de facto standard is PG25, a mixture of roughly 25 % by volume propylene glycol in water with corrosion inhibitors and antifoaming agents. Its properties at 20 °C give a specific heat of about 3.9 kJ/kg·K, thermal conductivity of 0.49 W/mK, viscosity of 2.4 cP and a freezing point of -10 °C. Pure deionised water has more than 4 % more heat capacity and pumps better; glycol is paid for in thermal performance and pumping energy, and in exchange gives two things: freeze protection and biological control.

That biological control has a surprising detail that explains a fair number of failures. Glycols are biostatic from 25 % by volume, but at concentrations on the order of 10 % biomass tends to feed on the propylene glycol rather than be inhibited by it. A badly dosed loop, or one topped up with water after a purge, is worse off than if it carried no glycol at all. The resulting biofilm blocks 0.2 mm channels and ruins heat transfer with no warning beyond a pressure drop that creeps up slowly.

The fluid quality specification is short and deserves a place in the maintenance contract:

Parameter	PG25 (TCS)	Plant water (FWS)
pH	8.0-10.5	7.0-9.0
Total hardness	< 50 ppm	< 200 ppm
Copper	< 2 ppm	—
Iron	< 2 ppm	—
Chlorides	< 25 ppm	< 50 ppm
Sulfates	< 25 ppm	< 100 ppm

Chlorides and sulfates accelerate pitting and galvanic attack, especially in aluminium and copper alloys. The design recommendation is to keep wetted materials as monometallic as possible, copper and brass or stainless steel, with a close anodic index between parts, and to avoid uninhibited aluminium and galvanised steel.

Filtration changes completely between loops. The TCS works below 25 microns, and commercial CDUs fit redundant 50 micron filters; the FWS accepts the 300 to 500 micron range. At the extreme, Google’s 2 MW CDU filters to 0.2 microns in a side stream to extend coolant life. OCP recommends on-site fluid testing quarterly and laboratory analysis annually, with standardised methods for metal ions, pH, reserve alkalinity and chlorides. Common operating practice is more intense at the start: daily analysis during commissioning, weekly after the first week and monthly in steady state. Documented coolant degradations are almost always attributed to the same causes: a badly executed initial flush, insufficient maintenance, or air ingress into the circuit.

Leaks, or the fear that was not the problem

The objection that always turns up in the first meeting is water over the electronics. The system defends itself in three layers.

The first is detection. Direct sensors, leak detection cables and point probes at the low points of trays and manifolds are combined with indirect detection through buffer tank level and bubble sensors: if air comes in or liquid is missing, something is open. The GB300 NVL72 platform already integrates leak detection at tray and rack level, with the signal available to the BMS.

The second is hydraulic design. The most elegant solution is to operate the server side below atmospheric pressure: with supply at about 0.13 bar of vacuum and return at 0.61 bar, a rupture draws air in rather than pushing coolant out. Incoming airflow is limited with check valves and venturis, and the measured thermal effect during a leak is a rise of one to three degrees in CPU temperature. A DOE cluster operated on this scheme has accumulated several years without a single recorded leak.

The third is factory qualification: hydrostatic testing at three times operating pressure, shock and vibration testing with subsequent leak-tightness verification, and at some vendors PG25 leak testing on 100 % of units plus an acceptance test simulating megawatts of load.

The result in production is good. Google reports CDU fleet availability of around 99.999 % since 2020, with liquid cooling in production since 2018, close to a gigawatt of liquid-cooled capacity and roughly half its global footprint with liquid deployed or enabled.

And here is the data point that dismantles the usual narrative. Asked about the barriers holding back DLC adoption, operators surveyed by Uptime in 2025 cite lack of standardisation (39 %), cost (38 %) and reliability (35 %), while safety concerns, the fear of water over electronics, appear in barely 7 %. What holds liquid back is not the leak. It is that there is still no comfortable standard, that it is expensive, and that it introduces failure modes operations teams do not know. Uptime adds an honest caveat: the sector is not applying enterprise IT redundancy standards to liquid today, because most installed capacity is in HPC and AI, where batch workloads tolerate interruptions a transactional system would not.

Getting liquid into an air-cooled room

Almost nobody starts from scratch, which is why the factor weighing most in the decision to adopt DLC, according to the same survey, is ease of retrofit onto existing infrastructure (46 % of responses), ahead even of operating savings.

The canonical retrofit route is the liquid-to-air CDU in sidecar format: a cabinet placed next to the rack, closing its own circuit with the cold plates and rejecting heat to room air. Commercial units run from 200 kW, at 1,145 kg and 19 kW of own consumption, to 500 kW at 2,300 kg and 55 kW. Building pipework does not have to be touched, and that is the whole appeal: it turns a construction problem into a purchasing problem. The cost is that the room has to be able to absorb those kilowatts in air, which in practice limits the approach to small islands.

The two physical limits of retrofit are weight and clear height. A GB200 NVL72 rack weighs on the order of 1,360 kg empty and reaches 1,500 kg with coolant, which demands floor positions rated for that load and, frequently, a concrete slab rather than a raised floor. Adding a one or two tonne sidecar alongside does not help. The conclusion deployment guides keep repeating is uncomfortable but realistic: most existing rooms cannot take a rack of this class without structural reinforcement, electrical expansion and thermal redesign, and the typical timeline for a full conversion, across manufacturing, delivery, integration and commissioning, is around a year.

For an inference factory

Four decisions concentrate almost all the value.

The first is to specify the capture fraction, not just the power. A tender asking for “liquid cooling for 130 kW racks” without saying which components carry a cold plate ends up with a room that has to remove 30 kW of air per cabinet and with internal fans consuming what liquid was supposed to save. The right question to the vendor is what percentage of node heat the fluid captures and on which components.

The second is to treat CDU approach temperature as a purchasing parameter. Two degrees of difference between a 4 °C CDU and a 2 °C one translate into two more degrees of FWS setpoint, and those two degrees are compressor-off hours in every year of the facility’s life. It is the thermal equivalent of choosing a UPS efficiency well.

The third is chemistry and filtration, which is where facilities are lost at the two-year mark. PG25 at a measured 25 %, not at the 12 % left after a top-up with water; a documented initial flush; filtration matched to the plates’ fin pitch; and a maintenance contract with quarterly analysis and written thresholds. A dirty loop gives no warning: performance drops slowly until a GPU starts throttling.

The fourth is the failure domain. Spread the load across CDUs so none serves more than 10 to 20 % of the servers, fit N+1 pumps with automatic changeover, and feed the CDUs from the backed-up system. The reason was given in the first article: a 130 kW rack without flow reaches dangerous temperatures in seconds, and critical cooling in an inference factory is designed with the same discipline as its electrical chain. The comparison is exact: the CDU is to cooling what the UPS is to power, and nobody would buy a UPS without module redundancy.

Direct liquid to chip is today the default answer between 50 and 150 kW per rack, and the whole ecosystem, from NVIDIA reference architectures to OCP specifications, is built around the cold plate. What remains to be seen is the next rung, the one that submerges the entire electronics, and what happens when the rack approaches a megawatt. That is the last article in the series.

Sources

Open Compute Project, Cold Plate Development and Qualification — https://www.opencompute.org/documents/ocp-cold-plate-development-and-qualification-with-integrated-comments-pdf
Open Compute Project, 30 °C Coolant: A Durable Roadmap for the Future, Rev 1.0 — https://ashrae.org.vn/wp-content/uploads/2024/12/30%C2%B0C-Coolant-A-Durable-Roadmap-for-the-Future-REV1_0.pdf
Open Compute Project, Universal Quick Disconnect (UQD) Specification Rev 1.0 — https://www.opencompute.org/documents/ocp-universal-quick-disconnect-uqd-specification-rev-1-0-2-pdf
Open Compute Project, Universal Quick Disconnect Blind-Mate (UQDB) Specification Rev 1.0 — https://www.opencompute.org/documents/uqdb-spec-1-0-pdf
Open Compute Project, Open Rack V3 Blind Mate Manifold Specification Rev 1.0 — https://www.opencompute.org/documents/open-rack-v3-blind-mate-manifold-specification-rev-1-0-review-april05-2024-pdf
ASHRAE TC 9.9, Emergence and Expansion of Liquid Cooling in Mainstream Data Centers — https://www.ashrae.org/file%20library/technical%20resources/bookstore/emergence-and-expansion-of-liquid-cooling-in-mainstream-data-centers_wp.pdf
Dell Technologies / OCP, Liquid Coolants Guidance for Technology Cooling System and Facility Water System — https://www.delltechnologies.com/asset/en-us/products/servers/industry-market/liquid-coolants-guidance-for-technology-cooling-system-and-facility-water-system-whitepaper.pdf
Schneider Electric, Direct Liquid Cooling System Challenges in Data Centers (White Paper 210) — https://media.datacenterdynamics.com/media/documents/WP210_V1_EN.pdf
Schneider Electric Blog, What coolant for data center liquid cooling: PG25, dielectric fluids, ASHRAE recommendations — https://blog.se.com/datacenter/2026/08/14/what-coolant-for-data-center-liquid-cooling-pg25-dielectric-fluids-ashrae-recommendations/
Schneider Electric Blog, Liquid cooling failures in AI data centers: causes, risks, prevention — https://blog.se.com/datacenter/2026/07/15/liquid-cooling-failures-ai-data-centers-causes-risks-prevention/
NSF Public Access Repository, Experimental evaluation of direct-to-chip cold plate liquid cooling for high heat density data centers — https://par.nsf.gov/servlets/purl/10529800
NVIDIA Technical Blog, NVIDIA Vera Rubin POD: Seven Chips, Five Rack-Scale Systems, One AI Supercomputer — https://developer.nvidia.com/blog/nvidia-vera-rubin-pod-seven-chips-five-rack-scale-systems-one-ai-supercomputer/
NVIDIA, NVL72 AI Factory Enterprise Reference Architecture: System Hardware and Components — https://docs.nvidia.com/enterprise-reference-architectures/nvl72-ai-factory/latest/components.html
NVIDIA Blog, Chill Factor: Blackwell Platform Boosts Water Efficiency by Over 300x — https://blogs.nvidia.com/blog/blackwell-platform-water-efficiency-liquid-cooling-data-centers-ai-factories/
Uptime Intelligence, Data Center Cooling Systems Survey 2025 (UI Field Report 181) — https://intelligence.uptimeinstitute.com/sites/default/files/2025-07/UI%20Field%20181_Data%20center%20cooling.pdf
Uptime Institute Journal, Performance expectations of liquid cooling need a reality check — https://journal.uptimeinstitute.com/performance-expectations-of-liquid-cooling-need-a-reality-check/
Uptime Institute Journal, AI embraces liquid cooling, but enterprise IT is slow to follow — https://journal.uptimeinstitute.com/ai-embraces-liquid-cooling-but-enterprise-it-is-slow-to-follow/
Vertiv, Deploying Liquid Cooling in the Data Center: A Guide to High-Density Cooling — https://www.vertiv.com/4926c8/globalassets/documents/white-papers/liquid-cooling/deploying-liquid-cooling-in-the-data-center-a-guide-to-high-density-cooling-white-paper.pdf
Vertiv, Liebert XDU1350 Coolant Distribution Unit, datasheet — https://www.vertiv.com/49971a/globalassets/shared/liebert-xdu1350-coolant-distribution-unitcdu-ds-en-na-sl-70799-web.pdf
Vertiv, 360AI High Density Reference Design RD0019 — https://www.vertiv.com/493c85/globalassets/documents/ai-hub-reference-design/vertiv-360ai-reference-design---rd0019.pdf
CoolIT Systems, CHx2000 CDU — https://www.coolitsystems.com/cdu-product/chx2000/
Chilldyne, FAQ Guide to Data Center Liquid Cooling — https://chilldyne.com/wp-content/uploads/2025/03/Chilldyne-FAQ-X.4.pdf
Chilldyne, Reference Design for AI: NVIDIA NVL72 — https://chilldyne.com/wp-content/uploads/2025/03/Chilldyne-Reference-Design-for-AI-NVIDIA-NVL72-X.3.pdf
Businesswire / Boyd, Google’s Two Megawatt Project Deschutes Coolant Distribution Unit at OCP Summit 2025 — https://www.businesswire.com/news/home/20251010498843/en/Boyd-to-Showcase-Googles-Recently-Released-Two-Megawatt-Project-Deschutes-Coolant-Distribution-Unit-at-the-2025-Open-Compute-Project-OCP-Summit
StorageReview, Inside Google’s plan to deliver 1MW racks and cool them too — https://www.storagereview.com/news/inside-googles-plan-to-deliver-1mw-racks-and-cool-them-too
Supermicro, DLC-2 Next Generation Direct Liquid Cooling Solutions — https://www.supermicro.com/en/pressreleases/supermicros-dlc-2-next-generation-direct-liquid-cooling-solutions-aims-reduce-data
Supermicro, Liquid-to-Air Sidecar CDU datasheet — https://www.supermicro.com/datasheet/datasheet_Sidecar_CDU.pdf
Dell’Oro Group, Data Center Liquid Cooling Market to Approach 7 Billion USD by 2029 — https://www.delloro.com/news/data-center-liquid-cooling-market-to-approach-7-billion-by-2029-as-ai-deployments-accelerate/

Cooling the AI datacenter (4/4): immersion and the megawatt horizon

Sat, 27 Jun 2026 17:00:00 +0200

We reach the last rung of the spectrum laid out in the first article. If the cold plate brings liquid up to the processor package, immersion removes the last intermediary: it submerges the whole board in a fluid that does not conduct electricity. On paper it is the definitive solution, with the best physics in the catalogue. In practice, 2026 has left it in a secondary role, and the reasons behind that distance between thermodynamics and deployment are the interesting part of this article.

This also closes the series, so the final section looks ahead: to the 600 kW rack announced for 2027, to the megawatt per cabinet, and to a technology that could make the whole argument obsolete.

The single-phase tank

In single-phase immersion the server lives inside a bath of dielectric fluid that never changes state. The fluid absorbs heat by convection, natural or pump-forced, and hands it to a heat exchanger that passes it to facility water. The fluid families admitted by OCP’s base specification are four: hydrocarbons, fluorochemicals, natural esters and synthetic esters.

That specification is useful because it puts hard numbers where marketing puts adjectives. An immersion fluid must maintain dielectric strength above 6 kV/mm throughout its service life, a flash point of at least 150 °C, a boiling point above 155 °C in single-phase, volume resistivity above $ 10^{11} $ ohm·cm and dynamic viscosity below 0.015 N·s/m². That last figure is the one governing operating cost: a fluid too viscous forces so much pumping that it eats the efficiency gain. And there is a requirement almost nobody expects to find on a cooling datasheet: dielectric constant below 2.3 and loss tangent below 0.05 between 20 MHz and 40 GHz, because the fluid is physically inside the path of high-speed signals.

Commercial capacities per tank have grown fast. Submer declares up to 361 kW per unit in its SmartPod EXO, with redundant pumps and CDU and operation on hot water up to 60 °C. GRC offers 368 kW CDUs with 13 °C water and a range going down to 13 kW for edge installations. Asperitas illustrates the design axis well: its natural convection model runs without pumps, with zero watts of non-IT consumption and around 32 kW per tank, while its forced convection version reaches 60 kW per tank with vertical cassettes. Academic reviews place commercialised single-phase around 100 kW per cabinet, with vendor-declared pPUE below 1.08 and in some cases close to 1.03.

Those pPUE figures need careful reading: they are partial, covering only the cooling loop and not total facility consumption. GRC itself, which sells single-phase, puts the efficiency difference between single-phase and two-phase at around 1 %, which already suggests the choice between them is not settled on thermal performance.

What has to change in the server

Here is the part commercial presentations tiptoe around, and the one that decides a deployment’s operational viability. A standard server is not submerged as it comes. OCP’s design guidelines for immersion equipment list the modifications, and the list is long.

Fans are not supported in most immersion strategies, so they have to be disconnected. The practical detail is that many firmwares refuse to boot without a tachometer signal, so a retrofit needs an emulator supplying the pulses the BMC expects to see. Conventional thermal paste can dissolve into the fluid, with a double effect: heat transfer at the chip is lost and the bath is contaminated; the recommendation is to replace it with a solid interface material such as indium foil, or to remove the heatsink if bath cooling is enough. Mechanical hard drives are excluded except helium-sealed ones, and in practice storage becomes SSD or NVMe.

Optics present a little-known and rather elegant problem: submerging a transceiver fills the ferrule’s air gap with fluid, changes the refractive index and introduces reflection losses. The ways out are copper cables with sealed connectors, silicon photonics, or extenders keeping the optics out of the bath. Labels are erased, because adhesives and inks dissolve, and with them serial numbers and MACs, which forces engraving or external documentation. Elastomers also suffer: EPDM swells, and in swelling can bend capacitor leads. Cable jackets with plasticisers, chlorine or sulfur dissolve and contaminate the fluid. And heat-shrink materials degrade above 60 °C.

None of this is insurmountable, and server vendors certified for immersion have it solved at the factory. But it explains why the ecosystem took so long to start: for years the real blocker was not thermal, it was contractual, because nobody warranted a submerged server. That changed in 2025, when Intel certified single-phase immersion fluids for the first time, Shell’s, and introduced a specific warranty extension for Xeon processors in immersion. Supermicro certified the first complete system under that scheme, and ExxonMobil added a second fluid family late in the year. Immersion has been viable in the lab for a decade; it has been purchasable with a warranty for eighteen months.

Boiling: the two-phase case

Two-phase immersion changes the mechanism. The fluid, a fluorochemical with a low boiling point, boils on contact with the hot chip surface; the vapour rises and condenses on a cooled coil mounted in the tank lid, and the condensate falls back into the bath. The cycle needs no pumps for the primary fluid, and phase-change heat transfer is an order of magnitude apart.

The numbers make it clear. Published reviews place the heat flux limit of forced air around 1.6 W/cm² for room air and up to a few tens with an optimised heatsink; water in a cold plate reaches the 16 to 200 W/cm² band depending on conditions; boiling reaches 500 W/cm² and remains effective above 1,000 W/cm². Critical heat fluxes measured with FC-72 on smooth copper are around 48 W/cm², and double with structured surfaces. Typical boiling points sit between 49 and 61 °C depending on fluid. Microsoft deployed the first two-phase system in production at a cloud provider in Quincy in 2021, with a fluid boiling at 50 °C, and measured per-server consumption reductions of 5 to 15 %. Commercially, LiquidStack declares up to 250 kW per 4U tank and 1.5 MW in a 48U format.

Two-phase’s problem is not thermal. It is the fluid, along two routes. The first is economic and operational: the vapour escapes. One system documented in the literature lost 102 litres a year on a total volume of 595, around 17 % annually, with a reference fluid at 75 dollars a litre. That is a recurring expense no business model digests well. The second is regulatory, and it is the one that has settled the contest.

The PFAS wall

In December 2022, 3M announced its complete exit from PFAS manufacturing by the end of 2025, including the fluorinated fluids that underpinned almost all two-phase immersion on the market. It was a business of some 1.3 billion dollars a year, abandoned under regulatory and litigation pressure.

At the same time, the universal PFAS restriction under REACH was advancing in Europe. Submitted in January 2023, it covers more than ten thousand substances and affects manufacture, placing on the market and use. ECHA’s risk assessment committee issued a final opinion favouring a broad restriction in March 2026, with the socio-economic committee’s opinion in public consultation and expected by year end. There are temporary derogations of up to twelve years for medical devices and semiconductors, and intermediate derogations for other sectors, but datacenter heat transfer fluids do not appear with specific treatment, and that is the risk for anyone who has built their room around a fluorochemical.

Alternatives exist and are nascent. Chemours is developing Opteon 2P50, a hydrofluoroolefin with a GWP of 10 that it trialled at full scale in Japan in March 2025 alongside NTT DATA and Hibiya, still in development and without a firm commercial date. Other suppliers cited as substitutes continue to offer fluorinated chemistry, and that is the nuance usually lost in the debate: a low GWP solves the fluid’s climate problem, not necessarily its fit within a regulatory definition of PFAS that rests on molecular structure. Until that uncertainty closes, an operator signing up to fifteen years of service life on a fluorinated fluid is accepting a risk they do not control.

Single-phase immersion with hydrocarbons or esters sits outside this problem, and it is one of the reasons it holds more than 70 % of the immersion segment today.

The third way: precision immersion

Between the open tank and the cold plate there is an intermediate approach worth mentioning. So-called chassis-level precision immersion encapsulates each server in its own sealed enclosure and floods only that volume, with flow directed at the components that need it. It captures practically 100 % of system heat, memory, storage and power supplies included, removes the fans and operates at noise levels below 50 dB, all inside a standard rack and without rebuilding the facility.

Its advocates’ commercial argument is direct and describes the tank’s problem well: classic immersion demands facility re-engineering, is heavy, hard to maintain and hard to deploy outside hyperscale environments. That the argument comes from inside the immersion field itself says a good deal.

Why direct liquid to chip is winning

With better physics, published standards and silicon vendor certifications, immersion ought to be eating the market. It is not, and the 2025 and 2026 figures show it from several angles.

Uptime’s 2025 cooling systems survey puts direct liquid cooling adoption at 22 % of installations. Immersion does not appear broken out in the report, which indicates its share is below the reporting threshold. Dell’Oro, which estimates the global liquid cooling market at some 3 billion dollars in 2025 and projects close to 7 billion by 2029, states that single-phase DLC accounts for the great majority of capacity entering service and describes immersion as a technology that “finds adoption selectively, where architectural trade-offs justify it”. Estimates of immersion market size range between 349 and 570 million dollars for 2025 depending on the analyst house, with annual growth of 24 to 25 %, healthy but on a small base and below the rate at which the total liquid market is doubling. An appreciable share of that figure also corresponds to cryptocurrency mining and edge installations, which is where immersion found its first mass market.

The reasons appear in that same Uptime survey, in the question about what makes liquid viable. The two most-voted factors are ease of retrofit onto existing infrastructure (46 %) and ease of maintenance (35 %), which are precisely the two axes where the tank loses by definition. A liquid-to-air sidecar is placed next to an existing rack in an afternoon. An immersion tank is construction work: it changes room layout, floor loading, the server handling procedure and staff training.

Add to that the gravity of the ecosystem. NVIDIA’s reference architectures for GB300 NVL72 and for Vera Rubin are cold plate, with manifolds, UQD connectors and liquid-cooled busbars. Google, with roughly half its global footprint on liquid deployed or enabled and around a gigawatt of capacity across two thousand AI chip pods, does it with cold plates and has contributed its 2 MW CDU to OCP. The capital of 2025 went the same way: Eaton’s purchase of Boyd Thermal, Daikin’s entry into Chilldyne, Motivair’s integration into Schneider. And the most eloquent indicator comes from two-phase’s own standard-bearer: LiquidStack launched a one-megawatt CDU for direct liquid to chip in 2024 and keeps extending that line. No company abandons its founding thesis, but all of them add the product that sells.

There is a final argument immersion has lost along the way: water. For years the tank was sold as the only route to a datacenter with no water consumption. Microsoft reached a fleet WUE of 0.30 litres per kWh in 2024, against 0.49 in 2021, and its next-generation design with closed-loop chip-level cooling avoids more than 125 million litres a year per facility with a WUE close to zero. Every design of theirs since August 2024 uses it, with the first sites operational in 2026. The water advantage no longer distinguishes immersion.

What the tank does keep, and it is real, is outlet temperature. In both single-phase and two-phase, the fluid returns heat above 60 °C stably, which is exactly what a district heating network needs. With heat reuse quotas entering European regulation, that is the niche where immersion has an argument the cold plate does not match as comfortably.

The horizon: from the 142 kW rack to the megawatt

The roadmap NVIDIA has published sets the pace for all physical infrastructure:

Generation	Power per rack	Year	Type of figure
Hopper	~40 kW	2023	deployed
GB200 NVL72	120-130 kW	2024-25	deployed
GB300 NVL72	up to 142 kW	2025	specification
Vera Rubin NVL144	190-230 kW	2026	roadmap
Rubin Ultra NVL576 “Kyber”	600 kW	H2 2027	roadmap
1 MW rack	1,000 kW	from 2027	roadmap

The Rubin Ultra announced for the second half of 2027 declares 576 GPUs, 2.5 million parts, 365 TB of fast memory and mandatory liquid cooling, with projected TDPs on the order of 3.6 kW per GPU. Vera Rubin reuses the Blackwell generation chassis with minor changes; Rubin Ultra demands a new rack and a new cooling system.

At that scale electrical distribution changes too, and here the series closes on itself. The 800 V DC architecture NVIDIA promotes for these racks carries 85 % more power through the same conductor than a 415 V AC distribution, with 45 % less copper and up to five points of improvement in end-to-end efficiency. In physical terms: it avoids having to run two-hundred-kilo copper bars to a single cabinet. It is the same trend anticipated in the power series (in Spanish), now with real facilities operating under that scheme and with the power vendor ecosystem lined up behind it.

None of those platforms contemplates immersion. All assume cold plate, warm water and CDU.

Microfluidics: attacking heat at the source

One line of work remains that could redraw the whole map, and it comes not from the facility side but from chip packaging.

The starting point is a paper published in Nature in 2020 by the EPFL group: microchannels fabricated monolithically inside the semiconductor substrate itself, co-designed with the electronics rather than bolted on top. The results were of another scale: heat fluxes above 1.7 kW/cm², with coefficients of performance above 10,000 and pumping powers of barely 0.57 W/cm², fifty times better than conventional straight microchannels.

Six years later that is leaving the lab along two paths. Microsoft announced silicon-etched microfluidics in September 2025, with channels the thickness of a hair and leaf-vein-inspired patterns designed with AI help to follow the chip’s thermal signatures; it declares up to three times better heat extraction than a cold plate and a 65 % reduction in maximum temperature rise inside a GPU, after four design iterations in a year. And TSMC presented direct-to-silicon liquid cooling integrated into its CoWoS packaging platform at ECTC 2025: a silicon microcooler fusion-bonded to the back of the chip, with no thermal interface layer, thermal resistance of 0.055 °C/W and demonstrated sustained TDP above 2.6 kW while keeping junction temperature rise below 63 °C.

If that line matures, the tank versus cold plate debate loses much of its point, because the bottleneck stops being how the board is bathed and moves inside the package. And one detail decides where the industry migrates: once heat leaves the chip in water at 45 or 60 °C, the infrastructure needed downstream is exactly the one direct liquid to chip already builds. Manifolds, CDUs, dry coolers. The tank has no natural path there.

Takeaways, and closing the series

Immersion has the best physics in the catalogue and a modest market position. That apparent contradiction sums up well what physical infrastructure teaches: the winning technology is not always the one that cools best, but the one that fits into rooms that already exist, with teams that already know how to operate them, with warranties somebody signs and with standards that allow changing supplier. Immersion still has its place, in new facilities built around it, in geographies with aggressive regulatory PUE caps, at the silent edge, and in deployments where high-temperature heat recovery is part of the business model. But anyone building an inference factory in Europe in 2026 is going to fit cold plates, and will be right to.

Three ideas run through all four articles.

The first is that cooling is a co-design, not an add-on. Rack density decides electrical distribution, floor loading, thermal method and water temperature at the same time, and those decisions are taken together or they are not taken well. A room designed for 40 kW per cabinet that then tries to host 130 ends up paying twice.

The second is that operating temperature is the most profitable efficiency lever there is. Raising the setpoint, reducing CDU approach and working with warm water is what allows the compressor to be switched off, the evaporative tower to be dropped and, along the way, waste heat to be delivered at a useful temperature. Energy, water and reuse move together, and all depend on the same parameter.

The third is that cooling is availability. A 130 kW rack without flow reaches dangerous temperatures in seconds, faster than it would fall over from a power cut. Pump and CDU redundancy, early detection, a bounded failure domain and real load testing belong to the same discipline as the UPS and the generator set of the power series (in Spanish). A thermal system nobody has exercised at full load is, once again, an unverified promise.

This closes the physical infrastructure vertical opened with the power chain. Between the two series the watt’s full journey is described: it enters through the utility feed, crosses switching and the UPS, reaches the silicon, converts entirely into heat and leaves through water toward the atmosphere or toward a heating network. What happens in the middle, the compute, is what usually takes the attention. Without the rest, it does not happen.

Sources

Open Compute Project, Base Specification for Immersion Fluids Rev 1.0 — https://www.opencompute.org/documents/ocp-base-specification-for-immersion-fluids-20221201-pdf
Open Compute Project, ACS Immersion Requirements Rev 2.10 — https://www.opencompute.org/documents/ocp-acs-immersion-requirements-rev-2-1-pdf
Open Compute Project, Design Guidelines for Immersion-Cooled IT Equipment Rev 1.01 — https://www.opencompute.org/documents/design-guidelines-for-immersion-cooled-it-equipment-revision-1-01-pdf
Science and Technology for Energy Transition, Immersion cooling technology development status of data center — https://www.stet-review.org/articles/stet/full_html/2024/01/stet20240005/stet20240005.html
PSEcommunity / LAPSE, Recent Advances in Two-Phase Immersion Cooling — https://psecommunity.org/wp-content/plugins/wpor/includes/file/2303/LAPSE-2023.16048-1v1.pdf
Uptime Intelligence, Data Center Cooling Systems Survey 2025 (UI Field Report 181) — https://intelligence.uptimeinstitute.com/sites/default/files/2025-07/UI%20Field%20181_Data%20center%20cooling.pdf
3M, 3M to Exit PFAS Manufacturing by the End of 2025 — https://news.3m.com/2022-12-20-3M-to-Exit-PFAS-Manufacturing-by-the-End-of-2025
Arnold & Porter, ECHA Committees Advance Broad PFAS Restriction Under REACH — https://www.arnoldporter.com/en/perspectives/advisories/2026/03/echa-committees-advance-broad-pfas-restriction-under-reach
DatacenterDynamics, Two-phase cooling will be hit by EPA rules and 3M’s exit from PFAS — https://www.datacenterdynamics.com/en/news/two-phase-cooling-will-be-hit-by-epa-rules-and-3ms-exit-from-pfas-forever-chemicals/
Chemours, Chemours joins forces with NTT DATA and Hibiya Engineering for full-scale Opteon 2P50 trial — https://www.chemours.com/en/news-media-center/all-news/press-releases/2025/chemours-joins-forces-with-ntt-data-and-hibiya-engineering-ltd-for-full-scale-opteon-2p50
Microsoft Source, To cool datacenter servers, Microsoft turns to boiling liquid — https://news.microsoft.com/source/features/innovation/datacenter-liquid-cooling/
Microsoft Source, AI chips are getting hotter: a microfluidics breakthrough goes straight to the silicon — https://news.microsoft.com/source/features/innovation/microfluidics-liquid-cooling-ai-chips/
Microsoft Cloud Blog, Sustainable by design: next-generation datacenters consume zero water for cooling — https://www.microsoft.com/en-us/microsoft-cloud/blog/2024/12/09/sustainable-by-design-next-generation-datacenters-consume-zero-water-for-cooling/
van Erp et al., Co-designing electronics with microfluidics for more sustainable cooling, Nature (2020) — https://www.nature.com/articles/s41586-020-2666-1
SemiWiki, Breaking the Thermal Wall: TSMC demonstrates direct-to-silicon liquid cooling on CoWoS (ECTC 2025) — https://semiwiki.com/semiconductor-manufacturers/362017-breaking-the-thermal-wall-tsmc-demonstrates-direct-to-silicon-liquid-cooling-on-cowos/
NVIDIA Technical Blog, NVIDIA 800 VDC Architecture Will Power the Next Generation of AI Factories — https://developer.nvidia.com/blog/nvidia-800-v-hvdc-architecture-will-power-the-next-generation-of-ai-factories/
NVIDIA, NVL72 AI Factory Enterprise Reference Architecture: System Hardware and Components — https://docs.nvidia.com/enterprise-reference-architectures/nvl72-ai-factory/latest/components.html
DatacenterDynamics, Nvidia’s Rubin Ultra NVL576 rack expected to be 600kW — https://www.datacenterdynamics.com/en/news/nvidias-rubin-ultra-nvl576-rack-expected-to-be-600kw-coming-second-half-of-2027/
DatacenterDynamics, Nvidia prepares data center industry for 1MW racks and 800-volt DC power architectures — https://www.datacenterdynamics.com/en/news/nvidia-prepares-data-center-industry-for-1mw-racks-and-800-volt-dc-power-architectures/
DatacenterDynamics, Chilling out in 2025: a year in data center cooling — https://www.datacenterdynamics.com/en/analysis/chilling-out-in-2025-a-year-in-data-center-cooling/
Dell’Oro Group, Data Center Liquid Cooling Market to Approach 7 Billion USD by 2029 — https://www.delloro.com/news/data-center-liquid-cooling-market-to-approach-7-billion-by-2029-as-ai-deployments-accelerate/
Network World, Shell’s immersion cooling liquids, first to receive official certification from Intel — https://www.networkworld.com/article/3984562/shells-immersive-cooling-liquids-the-first-to-receive-official-certification-from-intel.html
Supermicro, Industry First: Supermicro systems certified by Intel for an immersion cooling solution — https://ir.supermicro.com/news/news-details/2025/Industry-First----Supermicro-Systems-Certified-by-Intel-for-an-Immersion-Cooling-Solution/default.aspx
Submer, SmartPod EXO — https://submer.com/smartpod/exo/
GRC, ICEraQ product family — https://www.grcooling.com/iceraq/
Asperitas, Immersion Cooling Technology — https://www.asperitas.com/technology
LiquidStack, Two-Phase Immersion Cooling — https://liquidstack.com/two-phase-immersion
Iceotope, KUL chassis-level precision immersion cooling — https://www.iceotope.com/technologies/kul-chassis-level-hybrid-immersion-cooling/
Grand View Research, Immersion Cooling Market Report — https://www.grandviewresearch.com/industry-analysis/immersion-cooling-market
MarketsandMarkets, Immersion Cooling Market Report — https://www.marketsandmarkets.com/Market-Reports/immersion-cooling-market-107040948.html

Posts on lo0 — Tech Blog

Cooling the AI datacenter (1/4): the thermal challenge

The law that governs everything: power is heat

AI density breaks the model

The spectrum of solutions

The journey of heat, end to end

The metrics of heat

Water: the new constraint

Waste heat: from problem to resource

What is holding the transition back

Where the market is in 2026

For an inference factory

See also

Sources

Cooling the AI datacenter (2/4): air and its limits

Inside the air-cooled room

The geometry of airflow

The cube law, friend and executioner

Containment: cold aisle or hot aisle

The thermal envelope and class H1

Free cooling and a climate on the move

What it costs in water

The ceiling of air, and who reaches it

Rear doors: the last rung

The real density of the installed base

For an inference factory

See also

Sources

Cooling the AI datacenter (3/4): direct liquid to chip

Inside the cold plate

How much heat the plate captures

The two loops

The CDU: the thermal transformer

Flow, temperature rise and the price of warm water

Manifolds and connectors

The chemistry of the loop

Leaks, or the fear that was not the problem

Getting liquid into an air-cooled room

For an inference factory

See also

Sources

Cooling the AI datacenter (4/4): immersion and the megawatt horizon

The single-phase tank

What has to change in the server

Boiling: the two-phase case

The PFAS wall

The third way: precision immersion

Why direct liquid to chip is winning

The horizon: from the 142 kW rack to the megawatt

Microfluidics: attacking heat at the source

Takeaways, and closing the series

See also

Sources