Grid operators now pay large loads to reduce demand for a few tens of minutes during a peak event. A data centre can accept that payment only if it can keep serving while the cooling plant draws less power, and the standard answer is a tank of cold water. The interesting question is not whether storage works, but how many minutes of ride-through one cubic metre buys you.
A chilled-water tank stores cold in sensible heat: the temperature difference between supply and return, multiplied by the heat capacity of water. It is simple, cheap and easy to permit, but the energy density is modest, so a 30-minute event on a large hall needs a tank measured in hundreds of cubic metres. That footprint competes directly with the white space the operator is trying to sell.
The useful metric is minutes of ride-through per cubic metre at the required discharge rate, not stored energy per cubic metre.
Part One: What is actually being compared
To compare phase-change thermal buffer with chilled-water buffer tank honestly you first have to name the quantity being optimised. In data-centre peak shaving it is tempting to use the headline specification, but the binding quantity is the one that appears in the bill after a full year of operation: the sum of the committed cost, the per-unit cost and the cost of the occasions when the system behaves badly. Once that is written down, most of the specification-sheet argument simply disappears.
The cost structure behind phase-change thermal buffer is dominated by a fixed commitment that must be paid before the first unit of value appears, while chilled-water buffer tank converts that commitment into a variable charge on every unit. That difference is invisible in a per-unit comparison and decisive in a budget, because a fixed cost rewards scale and a variable cost punishes it. Whichever side your volume forecast lands on determines the answer long before the specification sheet does.
Part Two: The arithmetic that decides it
Phase-change materials store cold as latent heat while freezing and release it while melting, at a nearly constant temperature. Per cubic metre the capacity is several times higher, and the flat discharge temperature is genuinely useful because it keeps the supply setpoint stable for the whole event. The cost is power at the ends: the material must be frozen back before the next event, and heat transfer slows as a solid layer builds on the exchanger surface.
The second-order consequence is organisational. Choosing phase-change thermal buffer creates a dependency on a capability that must be staffed, monitored and kept current; choosing chilled-water buffer tank keeps the capability outside the team and turns it into a vendor relationship. Neither is free, and the one that looks cheaper in a spreadsheet is frequently the one that adds a standing meeting to the calendar.
Part Three: Where the simple model breaks
That recharge behaviour is the part procurement comparisons usually omit. A water tank can be recharged quickly whenever the grid is cheap, because the whole volume convects. A phase-change store has a moving solid front, so the last fraction of capacity takes disproportionately long to recover. If peak events arrive twice in one afternoon, the second one may find a store that is nominally half empty but effectively unavailable at the required discharge rate.
An 18-minute event repeated twice in one afternoon can exhaust a phase-change store that still reports 30 percent capacity, because the solid front limits discharge power.
Part Four: The cost that arrives later
There is also a control advantage on the water side. Chilled-water systems are well understood by facilities teams, their failure modes are visible, and a bypass valve gives a mechanical fallback. Phase-change stores need instrumentation on the melt fraction, which is inferred rather than measured directly, and a controller that decides whether to serve the load or preserve capacity for a later event. That controller is where most of the engineering risk sits.
Expect a split by scale rather than by industry. Large operators can amortise the fixed side of phase-change thermal buffer and will move first; smaller ones keep chilled-water buffer tank because their utilisation is spiky and their tolerance for operational complexity is low. The interesting competition happens in the middle, where a managed service can make either option look like the other.
Part Five: What this changes outside the engineering team
The economics therefore hinge on event frequency and duration, not on headline energy density. For a site that sees a handful of 18-minute events per year, the extra capital is hard to justify against a bigger tank. For a site in a market with frequent 4-minute dispatches and a high capacity price, the footprint saving alone can pay for the material. Two identical halls in different grids should buy different systems.
A home refrigerator behaves the same way. A full freezer full of frozen food holds temperature through a power cut far longer than an empty one, but it also takes hours to freeze again afterwards. Capacity and recharge time are separate properties, and only one of them appears on the label.
A market sizing sanity check
Data-centre cooling equipment is roughly a 20 to 30 billion dollar annual market, and grid-interactive load programmes are growing from a small base into a multi-billion-dollar revenue stream for flexible operators. If 10 to 20 percent of new large halls install dedicated thermal storage by 2029, the buffer segment alone is plausibly a 2 to 5 billion dollar annual market, with most of the value captured in controls and integration rather than in the storage medium.
Treat the numbers as a scale model. They are built to show where the curves cross and which term dominates, and they deliberately round the inputs that vary most between deployments. The ranking is robust to those rounds; the exact threshold is not.
The comparison that matters
Chilled water buys simplicity, fast recharge and a mechanical fallback at the price of footprint. Phase-change storage buys density and a flat discharge temperature at the price of a slow recovery and a harder control problem. Choose from the histogram of peak-event durations in your own grid, not from a capacity table.
The honest caveat is that this comparison is sensitive to one assumption: that the workload keeps its current shape. Shift the mix toward the case where chilled-water buffer tank is strong and the result changes without any technology changing at all. That is why the durable output of an evaluation is not a verdict but the threshold at which the verdict flips.
Two identical halls in different electricity markets should not buy the same cooling plant; the event histogram decides the payback.
SOL: HCqPmPJM3NVtS6YrmVaTZYnuP8XZ2PFgQZKT94VK1LkS
Creating this takes real effort — thank you for the support.
Top comments (0)