Resource note

Four Kinds of Battery Storage Buyers, and What Each One Should Actually Care About

2026-11-10 · Elias Bergstrom

The question behind the question

When someone emails me asking which battery storage system to buy, I've learned not to answer right away. The question usually hides a second one: what kind of buyer are you?

A 200 MWh utility project, a 400 kWh factory backup, and a rural Victorian homeowner pricing out a solar battery in Wangaratta are all shopping for lithium storage. Their questions sound identical. Their answers shouldn't be.

Context on where I'm coming from: I've handled procurement and vendor management for a mid-size commercial solar and storage installer since 2021. That's roughly $2.1 million a year across 30-odd vendors—maybe $1.9M, I'd have to pull the actual number. I report to both the operations director and the CFO, which means every recommendation I make has to survive two very different sets of questions.

Here's the framework I use. Three things sort buyers into four groups: project scale, whether you're grid-tied or islanded, and who absorbs maintenance risk after commissioning. Get those straight and most of the buying confusion resolves itself.

Fair warning before we start: my numbers here were accurate as of early 2025. Battery pricing and certification standards move fast, so verify current figures before you build a budget on them.

Scenario A: Utility-scale developers (10 MWh and up)

You're building for a grid operator or a large IPP. You care about availability guarantees, cycle warranty terms, and who's liable when a module goes into thermal runaway.

Two buying routes exist. Route one: buy an integrated system. The Tesla Megapack energy storage system is the reference point here—each Megapack 2 XL unit ships around 3.9 MWh, pre-integrated with thermal management and controls. You're buying a warranty and a service network as much as you're buying cells.

Route two: source cells from a manufacturer and integrate in-house or through a third party. This only works if you have real integration engineering capability. Most developers don't, and the ones who think they do usually learn otherwise on their first project.

What should you verify on the cell side? Provenance and consistency, mainly. This is where supply relationships carry weight. EVE Energy's lithium battery production has fed into Tesla's supply chain, which tells you something about the qualification bar those cells cleared. EVE Energy's Indonesia battery factory, with first phases targeted for 2025–2026, matters for a less obvious reason: regional supply diversity. If you're building across Southeast Asia or the Pacific, a second manufacturing base shifts your logistics and tariff exposure in ways a spreadsheet won't show you.

Certifications to demand: UL 9540 for the system, UL 9540A test reports for thermal runaway fire propagation, and NFPA 855 compliance for installation. If a vendor hedges on the 9540A report, walk. That document is the difference between an insurable project and an uninsurable one.

A counterintuitive note for this group: don't optimize for the lowest cell price per kWh. Optimize for warranty language and the failure liability split. A 4% cell cost advantage vanishes the first time a module fails and nobody's sure who pays for the crane.

Scenario B: Commercial and industrial buyers (100 kWh to 10 MWh)

This is the messy middle. Factories, cold storage, hospitals, data centers. You want peak shaving or demand charge reduction, and you don't have a 40-person energy engineering team.

My strong opinion: in this range, buy a containerized system from an integrator. Do not source bare cells and build your own pack.

I learned this the hard way. In 2022 we took on a cold storage project—600 kWh nominal, 590 usable, I'd have to check the commissioning report—and tried to shave cost by sourcing LiFePO4 cells directly and pairing them with a third-party BMS. The BMS and the cell management firmware disagreed about balancing thresholds. Commissioning slipped eleven weeks. We burned more on engineering hours than we saved on cells, and I had to explain the delay to a client whose product was literally spoiling.

That experience shaped how I think about supplier boundaries. A manufacturer that makes excellent LiFePO4 cells isn't automatically the right party to specify your container-level fire suppression. An integrator that's brilliant at controls may have no deep visibility into cell aging behavior. The vendors who say "that part isn't our strength—here's who does it better" have earned more of my trust than the ones promising a full stack.

For this segment, ask for IEC 62619 cell-level certification, UL 9540 on the assembled unit, and a written maintenance schedule. NFPA 855 spacing requirements catch people off guard—clearance distances can eat 20% of your available floor area if you find out late.

Scenario C: Residential and remote installs

Wangaratta, rural Victoria. A homeowner wants a solar battery because the grid drops three or four times a year and the last outage ran 31 hours.

Here's what surprises people: in this scenario, the battery's cycle life spec sheet matters far less than three other factors. Local installer accreditation. Backup switchover time. And how fast someone can physically reach your house when something fails.

In Australia, you want a Clean Energy Council accredited installer working to AS/NZS 5139. That standard governs battery system safety in installations and it isn't optional. I've seen quotes from installers who waved off the accreditation question. Those quotes were 15–20% cheaper. They were also useless, because the system wouldn't pass inspection.

Switchover time is the spec nobody asks about until the first outage. A system that takes 8 seconds to transfer to backup will reboot your router and your fridge. A system that transfers in under 20 milliseconds won't. If you work from home, that gap matters far more than whether the battery is rated for 4,000 or 6,000 cycles.

One genuinely satisfying result from a project two years back: a remote property we equipped with a modest 13.5 kWh system rode out a three-day outage without the family noticing anything beyond the neighbor's generator noise. The battery wasn't the cheapest quote. It was the one with a local service tech 40 minutes away.

Scenario D: You already own the battery, and it's dead

Different problem entirely. Someone has a LiFePO4 bank reading 0 volts and wants to know how to revive a dead LiFePO4 battery.

First, define "dead." Three distinct failures can look identical on a multimeter:

1. BMS protection lockout. The cells are fine but the BMS tripped on over-discharge and won't reset. Most common case, most recoverable.

2. One cell group is deeply discharged. Pack voltage reads low because one series group is dragging the rest down. Recoverable, but it needs a proper balance charge, not a jump start.

3. Actual cell degradation or damage. Swollen cells, electrolyte leakage, or internal resistance that's climbed out of spec. Not recoverable. Replace it.

For case one, check whether the BMS has a wake-up procedure—many do, and it's often a specific charge voltage applied to the terminals. For case two, you need a bench charger with per-group monitoring. Here's where I'll be honest about my limits: I can walk you through a 12V or 24V pack revival. I will not walk you through a 48V or higher system. Above roughly 60V DC you're in territory that requires a licensed technician, and the arc flash risk is real. Any advice that skips that warning is bad advice.

Before you attempt anything: check for swelling. Press gently on the cell casing. If there's any give, or any bulge, stop. A swollen LiFePO4 cell is a fire risk and no revival procedure is worth it.

For case three, run a capacity test. If the pack delivers under 80% of rated capacity after a full balance cycle, it's done for serious duty. It might still serve as a low-draw backup, but don't put it back on a critical load.

So which one are you?

Three questions will sort you:

  1. What's your project size? Above 10 MWh, you're a utility buyer and warranty terms are your primary negotiation. Below that, you're almost certainly better served by an integrated system.
  2. Do you have in-house integration engineers? If the honest answer is no—or "one guy who's enthusiastic"—buy integrated. Every time.
  3. Who's doing maintenance in year three? If it's you, and you're not a battery engineer, weight local service presence over spec sheets.

If your system is already installed and not working, you're in scenario D regardless of size. In that case the first question isn't "how do I revive it" but "is this even safe to touch." Answer that one before anything else.

I'm not 100% sure this framework covers every edge case—there's a gap between 10 MWh C&I and small utility that doesn't fit cleanly. But in five years of buying this stuff, it's sorted roughly 90% of the inquiries I get. The other 10% usually turn out to be a scenario D problem wearing a scenario A costume.