The question we hear most these days isn't "should I be looking at lithium?" It's "is lithium always the right answer now?" Our honest answer: no. Most of the time, yes. But not always — and the exceptions matter, because getting it wrong costs you either capital you didn't need to spend, or a lifecycle that doesn't hold up the way the slide deck promised. This piece is how we actually think through the decision with customers.
A quick note on chemistry, because "lithium" covers a lot of ground
When people say "lithium" in the UPS world, they almost always mean LFP — lithium iron phosphate, or LiFePO₄. It's not the lithium in your laptop or EV (those are NMC or NCA chemistries, which pack more energy per pound but run hotter and handle abuse much less gracefully). For stationary UPS duty, [1] LFP is the chemistry of choice precisely because it trades a little energy density for a lot more thermal stability, which is exactly the right trade for a battery that lives in your building.
On the other side is VRLA — valve-regulated lead-acid. It's been the UPS standard for decades, in two main flavors: AGM (absorbent glass mat) and gel. Sealed, maintenance-reduced, well-understood, and supported by almost every major UPS platform. The technology isn't obsolete — it's evolving — and pure-lead front-terminal designs have pushed VRLA performance further than the old-school reputation would suggest. [1]
So the real comparison for most UPS work is LFP vs. VRLA. Here's where each one actually shines.
Where LFP clearly wins
On the spec sheet, it's not close. LFP wins almost every technical metric:
Figures drawn from manufacturer data and industry comparisons [2][3][4][5]; your actual numbers will vary by product, duty cycle, and room conditions — and any serious deployment should be sized against current manufacturer runtime charts.
What this means in the room: for the same backup minutes, an LFP-based UPS takes somewhere between half and a third of the footprint, weighs significantly less (floor-loading actually becomes interesting in old buildings), tolerates a warmer closet, and doesn't want to be swapped out at the 4-year mark. Most customers looking at this list conclude LFP is the obvious answer. And usually it is.
Where VRLA still makes sense
Sometimes it isn't. We've installed enough of both to be honest about this. Here are the scenarios where we still quote VRLA first:
1. The capital budget isn't flexible and the timeline isn't long
LFP runs roughly 2–3× the upfront cost of VRLA for equivalent kWh. [5] If the building is getting sold in three years, or the project lives on a capital budget that can't flex, or you're refreshing equipment that's going to be replaced wholesale by a different tenant, the lifecycle math that favors LFP never actually gets a chance to play out. VRLA is the right call.
2. The existing infrastructure and workflows are VRLA
If your facilities team has VRLA procedures, VRLA disposal contracts, VRLA spare stock, and VRLA-trained techs — and your system is working — there's a legitimate argument for staying on that path through the next refresh. [6] A full LFP transition isn't just new batteries; it's new BMS integration, different maintenance rhythms, different training, and in some cases different monitoring contracts. Our recommendation in this scenario is almost always to align the LFP transition with your next scheduled UPS refresh — not to rip and replace ahead of time just to chase the spec sheet.
3. Small systems that stay below the NFPA 855 threshold
This is the one that surprises people. NFPA 855 — more on it in a moment — sets energy thresholds below which most of the stringent fire-code requirements don't apply. For lithium-ion chemistries that threshold is 20 kWh. For VRLA, it's 70 kWh. [7] If your whole backup system is a 6 kVA single-phase UPS with 30 minutes of runtime in a telecom closet, you're well below both thresholds either way — but as systems scale up, VRLA gets you more runtime before the code conversation gets complicated. For small and mid-size closets right at the margin, that's genuinely a factor.
4. Float-service UPS duty, where the cycle advantage isn't used
This one is important and under-discussed. Most enterprise UPS batteries don't cycle — they sit on float, holding charge for the occasional outage. LFP's 3,000–6,000-cycle advantage is enormous in applications that actually cycle (solar, EVs, microgrids). In a UPS that sees maybe a dozen meaningful discharges a year, most of that cycle capability goes unused — and the TCO math gets closer than the marketing suggests. [1]
The fire-code conversation nobody wants to have
We're not going to pretend this part is simple. NFPA 855 — Standard for the Installation of Stationary Energy Storage Systems — is the document that governs how lithium (and, to a lesser extent, VRLA) UPS installations have to be designed and sited. [8] The 2023 edition is referenced by the 2024 International Fire Code; the 2026 edition was recently published and is starting to be adopted. [7]
The short version of what 855 does: once your stored energy exceeds the threshold (20 kWh for Li-ion, 70 kWh for VRLA), your installation becomes subject to a cascade of requirements — UL 9540 listing of the ESS as a whole, UL 9540A thermal runaway fire-propagation testing for lithium systems, specific detection and ventilation requirements, siting and separation rules, and a Hazard Mitigation Analysis depending on the situation. [9]
For a typical enterprise deployment above these thresholds, this often means: the lithium system needs to carry the right UL listings, your room may need a dedicated detection scheme, your local Authority Having Jurisdiction (AHJ) is going to have opinions, and your insurance carrier may have separate questions. None of this is a reason not to go lithium — tens of thousands of compliant lithium UPS systems are installed safely every year — but it is a reason to have the conversation early, before the slab is poured and the cabinet is on the dock.
A practical note: the 20 kWh threshold is nameplate stored energy, not usable energy. A 10 kVA LFP UPS with external battery cabinets can cross the line before you realize it. This is one of those places where an early conversation saves a lot of pain later.
The TCO math — and an honest nuance about it
Over a 10–15 year horizon, LFP-based UPS systems typically come out 30–45% lower in total cost of ownership than VRLA once you include battery replacements, maintenance labor, cooling load, and footprint-driven real-estate costs. [2] For a data center, a healthcare environment, or an edge AI deployment, that margin is real and worth acting on.
The honest nuance: that TCO advantage is driven mostly by fewer battery replacement events and less floor space consumed. If your VRLA system is in a building where replacement labor is cheap and floor space is already sunk cost, and the system sits on float and rarely discharges, the TCO gap closes. Published analyses of 15-year float-service UPS scenarios at utility scale have actually shown VRLA ahead on TCO once fire suppression, HVAC, and insurance costs specific to the lithium installation were factored in. [1]
So the right question isn't "does lithium have lower TCO?" — it's "does lithium have lower TCO in your building, in your duty cycle, with your fire code reality?" That's the conversation we actually want to have with customers, and it's why we don't default to either answer.
How we actually help customers pick
Here's the decision pattern we walk through with most clients. It's not a formula — every building is different — but it's the conversation we tend to have:
- Edge / small site / under ~10 kVA, heat-exposed closet. LFP most of the time. The heat tolerance alone is usually worth the premium, and the 15-year battery life means you're not sending a tech to the site every four years.
- Mid-size office UPS, VRLA system currently in place and working. Stay on VRLA through this refresh, transition to LFP at the next one. We'll help you plan the transition so the existing plant gets its full useful life before you invest in new chemistry.
- Enterprise / mission-critical, stored energy above 20 kWh. Fire-code conversation first — what does your AHJ want to see, what's your insurance posture, what does your building's fire suppression actually support? Then TCO. In most cases, the answer still comes out LFP with the right documentation in place, but we want to know the constraints before we quote.
- Scheduled UPS refresh, legacy VRLA is end-of-life. This is where LFP most often makes sense. You're buying new hardware anyway; the BMS integration and training cost is bundled into the new platform; and you get the 10–15 year forward runway before you touch batteries again.
- New build, AI / high-density, fresh slab. LFP almost always. Densification is the whole point of the project, heat is part of the profile, and designing the fire code in from day one is much easier than retrofitting it. [2]
- Budget-constrained, short-horizon project. VRLA without apology. The lifecycle math needs time to work, and if you don't have the time, you shouldn't be paying for capability you won't use.
- Critical Power Battery Solutions (2026). Lead Acid vs Lithium Battery Comparison 2026: VRLA vs Li-ion for UPS, Data Centers. criticalpowerbatterysolutions.com/lead-acid-vs-lithium-battery-comparison-2026
- Gottog Power (2026). Are Lithium Batteries The Right Choice For AI Data Centers And Modular Infrastructure? gottogpower.com
- Xtreme Power Conversion. Lithium UPS vs Lead-Acid UPS Battery Technology. xpcc.com/lithium-ups-vs-lead-acid-ups
- Enconnex. Benefits of Lithium-Ion UPS vs. Lead-Acid (VRLA). blog.enconnex.com/lithium-ion-ups-vs-lead-acid-vrla-ups
- Vertiv / Battcon (2016). Discussion on Lithium Iron Phosphate Batteries Used for IDCs Compared with VRLA Batteries. vertiv.com (PDF)
- Leoch Lithium America. VRLA vs Lithium Battery for UPS: A Practical Guide for Data Center Backup Power. leochlithium.us
- Telgian (2025). NFPA 855 Changes in the 2026 Edition. telgian.com/nfpa-855-changes-in-2026
- NFPA. NFPA 855, Standard for the Installation of Stationary Energy Storage Systems. nfpa.org
- Mitsubishi Electric (2026). Understanding The UL 9540 Listing. mitsubishicritical.com