We build both air-cooled and liquid-cooled BESS. That means we've sat through enough commissioning calls, warranty discussions, and thermal modeling reviews to have a clear opinion on when each approach makes sense - and when it doesn't. This article lays out what we've learned, what the published data supports, and where the cooling decision usually gets made wrong.
The cooling method you choose for a battery energy storage system affects how long the batteries last, how hard you can cycle them, and whether the system holds its rated capacity in hot weather. Air cooling works for smaller, gently cycled systems. Liquid cooling is where most commercial and utility-scale projects land. The gap between the two isn't small.
Why Cooling Matters More Than Most Buyers Realize
Lithium-ion batteries don't like heat. That's not controversial - every cell manufacturer publishes a recommended operating range, typically somewhere between 15°C and 35°C, sometimes up to 40°C depending on the chemistry and cycling profile. NREL's Storage Futures Study and Annual Technology Baseline both emphasize that keeping cells within a moderate, stable temperature band is one of the most important factors in achieving the cycle life printed on the spec sheet.
What's less obvious is how steeply the penalties stack up once you leave that range. Pfannenberg's widely cited NREL-referenced analysis puts rough numbers on it: sustained operation at 30°C may shorten lifetime by around 20% compared to 20°C. At 40°C, losses approach 40%. At 45°C, usable life can drop by half. Those percentages shift depending on cell chemistry, pack design, and how aggressively the system cycles - but the direction doesn't change. Heat ages batteries. More heat ages them faster.
Now picture a 20-foot steel container sitting on a concrete pad in Phoenix or Riyadh. No shade, no climate control. Interior air temperature on a summer afternoon can blow past 50°C. That's not a hypothetical - it's the default condition for any outdoor BESS without active thermal management. And it's why the question isn't whether your system needs cooling, but which kind.
Cold weather brings a different problem that fewer buyers think about. Below 0°C, lithium-ion cells resist charging. Pushing current into a cold cell causes lithium plating - metal deposits that form on the anode, permanently reduce capacity, and increase internal short-circuit risk. NREL has flagged low-temperature charging as a specific degradation mechanism. If your site sees harsh winters, your thermal management system needs a heating function too, not just cooling.
One more thing that often gets overlooked: temperature uniformity within the battery pack matters almost as much as absolute temperature. When the hottest and coolest cells in a rack differ by 5°C or more, those cells age at different rates, charge at different speeds, and hit voltage limits at different times. The weakest cell sets the ceiling for the entire string. In a multi-MWh containerized system with thousands of cells, uneven thermal distribution is how you end up with capacity you paid for but can't safely access.
Sources referenced above: NREL Storage Futures Study and Annual Technology Baseline (temperature guidance, degradation modeling); UL 9540 (ESS equipment safety standard); UL 9540A (thermal runaway fire propagation test method, referenced by NFPA 855); published aging studies across LFP and NMC chemistries.
Air Cooling - Where It Works, Where It Doesn't
Air cooling uses fans to move ambient or conditioned air across battery modules. Simple, cheap, fewer things to break. We use it in our outdoor cabinet BESS for exactly those reasons - in a 60–120 kWh commercial cabinet that cycles once a day at moderate rates, air cooling keeps the thermal load in check without the plumbing complexity of a liquid loop.
The honest limitation: air doesn't transfer heat well. In high-density containerized formats, you need wide air channels between battery racks to maintain airflow, which eats into energy density. And even with good airflow design, cell-to-cell temperature spreads of 5–8°C are common. That spread drives uneven aging, and it gets worse in hot climates or during aggressive cycling - precisely the conditions where you need the cooling to work hardest.
We've had customers spec air cooling for cost reasons, then run into thermal throttling during summer peak-shaving. The BMS detects hot cells, pulls back discharge power to protect them, and the system delivers less than its rated output on the hottest days of the year. That's not a defect - it's the BMS doing its job. But if your business case depends on peak-day performance, air cooling in a hot outdoor installation is a mismatch.
For residential systems, small commercial installations under roughly 500 kWh, and anything sitting in a climate-controlled environment with gentle cycling, air cooling is the right call. Beyond that, we steer customers toward liquid.
Liquid Cooling - Why Most Commercial Projects End Up Here
Liquid cooling circulates a water-glycol coolant through metal plates pressed against battery cells. The coolant absorbs heat, carries it to an external chiller, and comes back cold. It's more expensive - the cost premium over air cooling runs in the range of 15–25% depending on system size and thermal architecture - and it adds plumbing, pumps, and a chiller that need maintenance.
So why do most C&I and utility-scale projects choose it anyway?
Because the physics gap is large. Water-glycol has dramatically higher heat capacity and thermal conductivity than air, which is why liquid-cooled systems can hold cell-to-cell temperature variation within 2–3°C. That uniformity translates directly into more even cell aging, more consistent usable capacity over the system's warranty period, and fewer surprises in year 5 when cells start diverging.
Density is the other factor. Without wide air channels between racks, you can pack more storage into the same container. Some liquid-cooled 20-foot containers now exceed 5 MWh - substantially more than typical air-cooled configurations in the same footprint. For projects where land cost or permitting constraints limit physical size, that density advantage matters.
There's also a revenue argument. Systems that can cycle aggressively without overheating are eligible for higher-paying grid services - frequency regulation, demand response, arbitrage strategies that require multiple cycles per day. The additional cycling headroom that liquid cooling provides can meaningfully improve annual returns, though the exact uplift depends on your market, dispatch strategy, and rate structure.
One project that shows the difference clearly: a 2 MWh containerized ESS we deployed in Australia. The system uses liquid cooling to manage thermal load across LFP cells in a hot outdoor environment - exactly the kind of site where air cooling would have forced the BMS into regular summer throttling. With the liquid loop maintaining tight cell-to-cell uniformity, the system cycles daily for peak shaving and renewable integration without the capacity derating that plagues underspec'd thermal designs in similar climates. That's the kind of result that's hard to put in a brochure but easy to see in the performance data twelve months in.
For any system above 500 kWh, cycling more than once daily, or sitting outdoors in a hot climate, we recommend liquid cooling as the starting configuration. The upfront premium is real, but it's small relative to the cost of premature battery replacement or lost revenue from thermal throttling.
Immersion Cooling - Worth Watching, Not Yet Standard
Immersion cooling submerges cells entirely in a non-conductive dielectric fluid. Every surface contacts the coolant directly - no plates, no thermal interface material, no air gaps. Cell-to-cell temperature variation drops to near zero, and the fluid itself acts as a fire barrier.
Some vendor testing suggests immersion-cooled batteries may last meaningfully longer than plate-cooled equivalents, though independent field data at grid scale is still thin. The technology is getting attention for data center backup power and extreme-heat deployments. Costs are trending down, but as of early 2026, immersion cooling is still a niche option for stationary storage - something we're watching, not yet something we'd recommend as a default.
The Budget Question, Answered Honestly
We get asked about cooling cost-benefit on nearly every commercial project. Here's how we frame it.
Take a 1 MWh LFP system cycling daily. With liquid cooling holding cells near 25°C, that system might deliver 6,000–8,000 cycles over its warranty period - the exact number depends on depth of discharge and cycling profile. If that same system runs consistently at 35°C because the cooling was underspec'd, cycle life could fall to 4,000 or less before hitting warranty-triggering degradation. At current LFP cell costs, the replacement gap between those two outcomes easily exceeds the cost of specifying liquid cooling at the start.
Financing is part of it too. When lenders and insurers evaluate a project, they look hard at safety documentation. UL 9540 - the ESS equipment safety standard - and UL 9540A - the test method for evaluating thermal runaway fire propagation, explicitly referenced by NFPA 855 - both probe how the system handles thermal stress. A system with a well-designed thermal management backbone that supports full UL certification tends to get better insurance terms and faster permitting. That's not a soft benefit - it's project timeline and cost of capital.
How We Help Customers Decide
When a customer comes to us early in project design, we walk through five variables before recommending a thermal configuration:
- System size: Under 500 kWh, air cooling usually handles the load. Above 1 MWh, liquid cooling is the practical default.
- Cycling profile: One gentle cycle per day at 0.25C? Air is fine. Multiple daily cycles or fast discharge for grid services? Liquid.
- Site climate: Indoor or temperate outdoor? Air can work. Desert, tropical, or extreme-cold deployment? Liquid with an integrated heating loop.
- Revenue model: Simple peak shaving? Air may suffice. Revenue stacking with frequency regulation and arbitrage? The system needs the cycling headroom that liquid cooling provides.
- Footprint constraints: Tight site? Liquid cooling's density advantage means fewer containers for the same capacity.
If you're comparing BESS configurations and thermal management is part of the decision, our article on real-world BESS performance factors covers the broader picture - including BMS quality, integration testing, and how thermal management interacts with warranty terms.
Air vs. Liquid vs. Immersion - Quick Reference
| Air Cooling | Liquid Cooling | Immersion Cooling | |
|---|---|---|---|
| System size | 5 kWh – 500 kWh | 500 kWh – multi-MWh | Specialty / pilot-scale |
| Cycling intensity | 1x/day, moderate C-rate | Multiple cycles/day, high C-rate | High C-rate, continuous duty |
| Cell-to-cell uniformity | 5–8°C (design-dependent) | 2–3°C typical | Near-zero |
| Climate suitability | Temperate, indoor, mild outdoor | All climates (with heating loop) | Extreme heat, high-density sites |
| Relative cost | Lowest | Moderate premium | Highest (declining) |
| Best for | Residential, small C&I, backup | C&I, utility-scale, grid services | Data centers, extreme environments |
What's Changing in Thermal Management
A few things we're paying attention to on the product development side.
Some BESS suppliers are integrating AI-driven thermal optimization into their energy management software - using weather forecasts and dispatch schedules to pre-cool batteries before heavy cycling rather than reacting after temperatures spike. Where it's deployed well, operators report tighter thermal control with lower auxiliary power consumption. We're seeing this mostly from the larger, software-forward integrators; it hasn't filtered down to mid-market systems yet.
Phase change materials are being explored as a passive thermal buffer in hybrid cooling architectures. IRENA's Innovation Outlook on thermal energy storage has identified improved PCMs as a potential pathway to better efficiency, though commercial use in stationary BESS is still limited. The idea - using a material that absorbs heat as it melts to smooth out transient spikes - is sound. Scaling it reliably in a containerized format is the remaining engineering challenge.
On the cell hardware side, the shift toward larger-format cells (from the 280 Ah cells that dominated 2022–2024, through 314 Ah, into 700+ Ah formats) has thermal management implications. Fewer cells per system means fewer cell-to-cell junctions where temperature gradients form. Whether that simplifies cooling enough to change the air-vs-liquid calculus depends on pack architecture - but it's moving in the right direction.
If the chemistry angle interests you, our piece on high voltage battery chemistry performance goes deeper into how LFP and NMC behave differently under thermal stress - and what that means for system design.
Common Questions We Get From Buyers
Does my facility actually need liquid cooling, or is that overselling?
It depends on how hard the system works. If you're installing a 200 kWh backup system in an air-conditioned utility room and cycling it a few times a month, liquid cooling is overkill - air cooling handles that fine. If you're putting a 1 MWh system outdoors for daily peak shaving plus demand response, liquid cooling is not overselling. It's protecting a six-figure investment from avoidable degradation. The cost of getting this wrong usually shows up in year 3–5, when air-cooled systems in hot climates start losing capacity faster than the financial model projected.
What about LFP vs. NMC - does the chemistry change the cooling requirement?
LFP has a wider thermal safety margin. Its thermal decomposition point is around 270°C versus 210°C for NMC, which makes LFP more forgiving of brief temperature excursions. But both chemistries degrade faster outside their optimal operating range. LFP's safety advantage means the consequences of a cooling failure are less catastrophic - not that you can skip cooling. The chemistry choice affects sizing and safety margins, not the fundamental need for thermal management.
Can I start with air cooling and upgrade later?
Technically yes, practically difficult. Retrofitting liquid cooling into an air-cooled container means redesigning the rack layout, adding plumbing runs, installing a chiller, and recalibrating the BMS. In most cases the cost and downtime exceed what you'd have spent specifying liquid cooling from the start. If there's any chance your cycling profile or revenue strategy will intensify over the system's life, spec the thermal system for the endgame, not the starting condition. Our BESS cost breakdown article covers how to budget for this correctly upfront.