Home battery backup in 2026 comes down to three chemistries: lithium iron phosphate (LiFePO4), lead-acid in its three flavors (flooded, AGM, gel), and saltwater — the chemistry that promised to change everything and mostly teaches a lesson about buying proven technology. I've banked my reputation on all three over the years: flooded banks in off-grid cabins, AGM strings in telecom shelters, and LiFePO4 racks in just about everything built lately. This guide is the comparison we wish every customer read before the sales call — real cycle-life numbers, cost-per-kWh-throughput math, maintenance reality, and the sizing arithmetic, with no chemistry sold as magic.
1. The Master Comparison Table
| Attribute | LiFePO4 (LFP) | Flooded Lead-Acid (FLA) | AGM Lead-Acid | Gel Lead-Acid | Saltwater (aqueous hybrid ion) |
|---|---|---|---|---|---|
| Usable depth of discharge | 90–100% | 50% max recommended | 50–60% | 50–60% | 100% claimed |
| Cycle life @ typical DoD | 4,000–6,000+ @ 80% DoD | 500–1,200 @ 50% DoD | 400–800 @ 50% DoD | 500–1,000 @ 50% DoD | ~3,000 claimed |
| Round-trip efficiency | 95–98% | 80–85% | 85–90% | 85–90% | ~90% claimed |
| Energy density (Wh/kg) | 90–160 | 30–40 | 35–45 | 35–45 | ~20–30 |
| Maintenance | None | Watering, equalization, ventilation | None (sealed) | None (sealed) | None |
| Charge temperature floor | 0 °C (32 °F) without heating | Below freezing OK at reduced rates | Below freezing OK at reduced rates | Sensitive to over-voltage | Wide range claimed |
| Upfront cost per nominal kWh | $300–$700 | $150–$300 | $250–$450 | $300–$500 | N/A (see Section 5) |
| Typical warranty | 10 years / 6,000 cycles class | 1–3 years | 2–5 years | 2–5 years | Varies; company history matters |
| Fire risk profile | Very low (LFP thermal stability) | Hydrogen off-gassing requires ventilation | Low; still hydrogen under fault | Low; still hydrogen under fault | Very low (aqueous electrolyte) |
Read the DoD row twice, because it inverts the sticker comparison. A 10 kWh lead-acid bank is really a 5 kWh bank if you want it to survive; a 10 kWh LiFePO4 bank is a 9–10 kWh bank. Every per-kWh comparison that ignores usable capacity is comparing apples to half-apples.
2. Cost Per Usable kWh Throughput — The Only Fair Metric
The honest way to compare chemistries is dollars per lifetime kilowatt-hour delivered: (bank cost) ÷ (nominal kWh × usable DoD × cycle life × round-trip efficiency). Here's the math at representative 2026 prices — your quotes will vary, the structure won't:
| Chemistry | 10 kWh Bank Cost | Usable kWh/Cycle | Cycles | Lifetime kWh Delivered (× efficiency) | Cost per Lifetime kWh |
|---|---|---|---|---|---|
| LiFePO4 ($4,500) | $4,500 | 9.0 (90% DoD) | 5,000 | 45,000 × 0.96 = 43,200 | ~$0.10/kWh |
| Flooded lead-acid ($2,200) | $2,200 | 5.0 (50% DoD) | 900 | 4,500 × 0.82 = 3,690 | ~$0.60/kWh |
| AGM ($3,200) | $3,200 | 5.5 (55% DoD) | 650 | 3,575 × 0.87 = 3,110 | ~$1.03/kWh |
| Gel ($3,800) | $3,800 | 5.5 (55% DoD) | 800 | 4,400 × 0.87 = 3,828 | ~$0.99/kWh |
That's the whole story of the last decade in one table: lead-acid costs half as much up front and three to ten times as much per kilowatt-hour actually delivered. When a customer tells me "lithium is too expensive," what they mean is the invoice is bigger this month. On every bank that cycles daily, lithium is the cheap option — it's just expensive the way insulation is expensive: once, up front, and then never again. For shallow-duty applications — a cabin used twelve weekends a year, a pure standby float application — the math flips and a quality AGM or even flooded bank can be the rational buy. Duty cycle decides; we run this exact arithmetic with customers on the whiteboard before anyone talks brands.
3. Lead-Acid: Three Flavors, One Family Trait
Flooded lead-acid is the century-old workhorse: cheapest upfront, tolerant of abuse in some directions (cold charging, mild overcharge) and utterly unforgiving in others (chronic undercharge sulfation, deep discharge). It demands monthly watering with distilled water, quarterly specific-gravity checks, periodic equalization charges, and ventilated space — the hydrogen off-gassing under charge is a real NEC 480/706 ventilation design item, not a footnote. I've walked into battery rooms where the watering schedule died with the previous owner's enthusiasm; sulfated FLA banks lose 30% of capacity before anyone notices the lights dimming sooner.
AGM seals the electrolyte in glass mat: no watering, no spills, better charge acceptance, safe for occupied spaces, and the standard for standby/UPS duty. The catch: AGM hates heat and hates being floated hot; every 10 °C above 25 °C roughly halves calendar life, a rule of thumb that has accurately predicted every premature AGM failure we've autopsied. Gel is AGM's niche cousin — immobilized electrolyte, excellent deep-discharge tolerance, very charge-voltage-sensitive; overcharge a gel bank and the voids that form are permanent. For renewable cycling, gel is mostly a legacy choice now.
| Lead-Acid Type | Best Duty | Worst Enemy | Charge Voltage (12 V, 25 °C) | Field Verdict |
|---|---|---|---|---|
| Flooded (FLA) | Off-grid cycling on a budget, owner does maintenance | Chronic undercharge (sulfation) | Absorb 14.4–14.8 V, EQ 15.3–15.8 V | Cheapest per upfront kWh; demands an owner |
| AGM | Standby/UPS, occasional cycling, indoor installs | Heat; hot float charging | Absorb 14.2–14.6 V, float 13.5–13.8 V | The default where lithium doesn't fit the budget |
| Gel | Deep-cycle niche, mobility, high-vibration | Over-voltage (permanent voids) | Absorb 13.9–14.2 V — strictly | Legacy choice; verify charger gel profile |
Quality matters more inside lead-acid than anywhere else — the difference between a true deep-cycle plate and a re-purposed starting battery is the difference between five years and one. That's why our lead-acid shelf is short and deliberate: Trojan and Rolls/Surrette flooded deep-cycle, proven AGM and gel lines, plus lead-carbon for the sulfation-resistant middle ground. Full maintenance protocols live in our battery life extension guide.
4. LiFePO4: Why It Ate the Market
Lithium iron phosphate won for stacked reasons, and none of them are marketing. Cycle life in the 4,000–6,000 range at 80% DoD turns a battery from a consumable into an appliance. Round-trip efficiency of 95–98% means your solar harvest mostly reaches your loads. Zero maintenance ends the watering calendar. Usable capacity at 90–100% DoD halves the nameplate you need to buy. And the LFP chemistry specifically — versus the NMC lithium in phones and EVs — trades some energy density for thermal stability that makes it the right risk profile for a box bolted to your house. The honest weaknesses, because every chemistry has them: charging below freezing requires either heated batteries or BMS charge-inhibit (nearly universal now, but confirm it — a LiFePO4 cell charged hard below 0 °C plates lithium and loses capacity permanently); the upfront invoice is real; and you're trusting a battery management system, which makes brand quality and warranty support part of the product. A BMS from a brand that answers the phone is worth money; we've watched "identical" no-name cells diverge within two years while name-brand racks from EG4, SimpliPhi, and Fortress Power (the eFlex Max 5.4 kWh is the value benchmark we quote against) hold balance for a decade. Browse the current LiFePO4 catalog and 48 V rack batteries for the build-your-own-bank route, or lithium batteries generally for drop-in 12 V replacements.
5. Saltwater Batteries: An Honest Post-Mortem
Saltwater deserves a section because customers still ask, and the honest answer requires history. The technology — aqueous hybrid ion chemistry, sodium ions shuttling between manganese oxide and carbon electrodes in a salt-water electrolyte — was real: non-toxic, non-flammable, 100% depth of discharge, wide temperature tolerance. Aquion Energy commercialized it, shipped the Aspen and S-Line stacks, and won legitimate praise for the greenest battery ever UL-listed. Then the economics hit: energy density a third of lead-acid, power density that made inverter-matching awkward, cost per kWh that never reached the promised curve — and the company filed Chapter 11 in 2017. The assets changed hands; production sputtered; warranty support for installed units became a scavenger hunt. The lesson isn't that saltwater chemistry is bad — it's that a battery is a 10–15-year commitment to a company, not just a chemistry. We tell customers to weight warranty bankability alongside the spec sheet: who answers the phone in year seven? That's why this store stocks chemistries with decade-deep supply chains. If a new aqueous or sodium-ion contender reaches UL-listed volume production with a solvent manufacturer behind it, we'll test one in the shop and report back. Until then, saltwater remains the proof that "non-toxic and fireproof" doesn't pay the inverter.
6. Sizing Math — A Worked Backup Example
Quick, verifiable arithmetic for a critical-loads backup: loads total 6.5 kWh/day (fridge 1.5, furnace blower 1.2, lights and outlets 2.3, well pump 0.8, misc 0.7). Target autonomy: 2 days with no sun. Required usable storage: 13 kWh. Now the chemistry multiplier: LiFePO4 at 90% DoD needs 13 ÷ 0.9 ≈ 14.5 kWh nameplate — call it three 5.12 kWh rack modules (15.4 kWh). Flooded at 50% DoD needs 26 kWh nameplate — twenty-plus golf-cart batteries, plus a ventilated room. AGM at 55% needs ~24 kWh. Then the aging factor: plan for 20% capacity loss over the design life and the lithium case lands at four modules or accepts slightly reduced autonomy in year ten; the lead cases grow by the same proportion from an already doubled base. Inverter surge is the other gate: the well pump's locked-rotor current, typically 3–5× running watts for a fraction of a second, must sit under the inverter's surge rating or carry a soft starter. Chemistry doesn't fix a stalled inverter. Run your own numbers in the battery sizing calculator, cross-check against the bank sizing guide, estimate runtime with the runtime calculator, and internalize the 20/80 rule — keeping daily swings inside 20–80% state of charge is the single cheapest lifespan extender on any chemistry that allows it.
7. Maintenance Calendar — What Each Chemistry Actually Demands
| Task | LiFePO4 | Flooded | AGM/Gel |
|---|---|---|---|
| Watering (distilled only) | Never | Monthly; more in heat | Never (sealed — do not open) |
| Equalization charge | Never (harmful) | Monthly–quarterly per SG readings | AGM: rarely, per maker. Gel: never |
| Terminal cleaning / torque check | Annual | Quarterly (corrosion is constant) | Annual |
| Specific gravity logging | N/A | Monthly — the real health metric | N/A (voltage only) |
| Ventilation check | Standard room air fine | Active venting; hydrogen is lighter than air | Normal ventilation |
| Temperature compensation verification | BMS handles; confirm cold charge-inhibit | Confirm −3 to −5 mV/°C/cell on charger | Same; gel narrower window |
Two field notes. First: most "dead" lead-acid banks we inspect were murdered by calendar — skipped watering and skipped equalization — not by cycles. If your household won't do monthly maintenance, buy sealed or lithium and consider it a maintenance contract paid in advance. Second: whatever the chemistry, log something. Specific gravity on flooded, resting voltage and mid-string variance on AGM, cell-balance spread on lithium. Trends catch problems; snapshots don't.
8. Safety and Code Placement in the Home
Garage, basement, or dedicated enclosure — the current NEC 706 and IRC rules push residential ESS toward listed equipment with maximum individual unit sizes (commonly 20 kWh per unit and aggregate limits per location under the residential provisions), spacing from doors and windows, and vehicle-impact protection in garages. Flooded banks add NEC 480 ventilation and eyewash-adjacent common sense. LiFePO4's thermal stability is why it passes these rules cheaply — but "LFP is safe" is not permission to skip clearances, working space, or the disconnect requirements. Your AHJ's adoption year matters; ask before you order, or let us pull the current requirements into the quote. See the full battery storage collection for listed residential options, and our disconnect and OCPD guide for the protection side.
Frequently Asked Questions
Is lithium really cheaper than lead-acid over time? On anything that cycles regularly, yes — by 3–10× per kilowatt-hour delivered. Lead-acid's 50% usable DoD and 500–1,200 cycle life against lithium's 90% DoD and 4,000–6,000 cycles swamps the upfront price difference. The exception is shallow-duty standby use, where AGM's lower invoice wins.
Can LiFePO4 batteries charge in a freezing garage? Not below 0 °C without protection — charging cold plates lithium metal inside the cell and permanently destroys capacity. Buy batteries with low-temperature charge cutoff or built-in heaters (both common in 2026), or keep the bank inside the thermal envelope. Discharging in the cold is fine at reduced capacity.
What happened to saltwater batteries? The chemistry worked — non-toxic, non-flammable, full depth of discharge — but the energy density and cost never competed, and the flagship manufacturer went through bankruptcy in 2017. It stands as the industry's reminder that a 10-year warranty is only as good as the company behind it.
How many kWh of battery do I need for home backup? Total your critical loads in kWh/day, multiply by days of autonomy, divide by usable DoD (0.9 lithium, 0.5 lead-acid), then add ~20% for aging. A typical critical panel of 6–8 kWh/day with 2-day autonomy lands at 15–18 kWh of lithium or roughly double that nameplate in lead-acid.
Can I mix old and new batteries in one bank? Don't. Mixed ages or chemistries in a shared bank charge and discharge unevenly — the weak units drag the strong ones to an early grave. Replace banks as matched sets; if expanding lithium racks, same model, same firmware, and per the manufacturer's expansion rules.
Do lead-acid batteries belong inside the house? Sealed AGM/gel, yes, with normal clearances — they're the standard for indoor standby. Flooded belongs in ventilated, dedicated space: hydrogen off-gassing under charge is a documented ignition hazard and a code-governed ventilation design, not a preference.
9. Warranty Reading — The Fine Print That Pays
Battery warranties look uniform on a landing page and diverge in the documents. What we teach customers to check, in order: throughput caps (a "10-year" warranty with a 15 MWh throughput ceiling is a 4-year warranty on a house that cycles daily — do the division before signing); end-of-warranty capacity floor (70% retained is typical; 60% is a red flag); cycle counting method (equivalent full cycles versus calendar events — the first is fairer to partial-cycling homes); transferability if you sell the house; and the labor/shipping terms for a warranted replacement, because a free module you pay $800 to freight and reinstall is a discount, not a warranty. Lead-acid pro-rata schedules deserve special suspicion: the "36-month warranty" that pro-rates from month 13 is worth about fourteen months. Keep commissioning photos, serials, and the install date in one folder; the claims that resolve fast are the ones that arrive documented.
10. The Verdict by Use Case
| Application | Our Pick | Runner-Up | Why |
|---|---|---|---|
| Daily-cycling solar self-consumption | LiFePO4 rack | Lead-carbon | Throughput cost decides; lithium wins by 3–10× |
| Whole-home backup, grid-tied | LiFePO4 (integrated or rack) | — | Cycle life + efficiency + zero maintenance |
| Weekend cabin, 15 cycles/year | Quality AGM | Flooded (if owner maintains) | Calendar aging kills both; buy the cheaper invoice |
| Off-grid full-time, tight budget, hands-on owner | Flooded Trojan/Rolls | Lead-carbon | Lowest entry cost for an owner who'll do the maintenance |
| Unheated space, cold climate | Heated LiFePO4 or AGM | — | Cold-charge protection is non-negotiable for lithium |
| Mobile / marine / RV | LiFePO4 drop-in 12 V | Gel | Weight, DoD, and vibration tolerance all favor LFP |
| Non-toxic mandate (schools, sensitive sites) | LiFePO4 with listed ESS | Watch sodium-ion | Saltwater's role is currently unfilled at scale |
Final word from the warehouse floor: we still sell pallet quantities of flooded deep-cycle every month, and they're the right answer for a shrinking but real slice of jobs. But the direction is unambiguous — lithium's cost curve crossed lead's throughput cost years ago, and the gap widens every year. Buy for your duty cycle, buy from brands with a decade of phones-answered history, and maintain whatever you buy like it cost real money. Because it did.
11. What's Coming: Sodium-Ion and the Next Decade
One chemistry deserves a forward look because it's closer than most people realize: sodium-ion. Same working principle as lithium-ion but with sodium — abundant, cheap, no lithium or cobalt supply chain — trading energy density for cost and cold-weather performance (sodium-ion cells charge happily at temperatures that freeze lithium out). Utility-scale sodium storage is already shipping; residential-format products are in pilot. If the cost curve holds, sodium-ion occupies the role flooded lead-acid plays today: the budget workhorse for stationary storage where weight doesn't matter — except with lithium-class cycle life and no maintenance. We're not stocking it yet, and we'd advise the same skepticism this article applied to saltwater: let someone else's project be the pilot, verify the UL listings and the warranty bankability, and check back in two years. The pattern of this industry is that the boring, bankable chemistry wins — LFP was once the boring option too.
12. End of Life: Recycling and Second Lives
The question arrives more often every year: what happens to the bank in fifteen years? Lead-acid has the best answer in the industry — a closed recycling loop that recovers 95%+ of the material and pays you a core credit; it's the most recycled consumer product in America, full stop. Lithium recycling is real but younger: hydrometallurgical recovery of lithium, copper and aluminum is scaling, and most major LiFePO4 brands now run take-back programs — ask for the end-of-life terms at purchase, because the brand that plans the battery's funeral probably engineered its life. Second-life applications (retired EV and solar packs serving stationary duty at reduced capacity) are a genuine market, but buy them only with documented capacity testing and a seller who warrants the result. A "cheap" second-life bank without test data is a lottery ticket wearing a battery costume.
That closes the loop on the chemistry question: buy for the duty cycle, from companies that will answer the phone in year ten, maintain what you buy, and plan the end at the beginning. The full energy storage catalog covers every chemistry discussed here, and the counter is open for the arithmetic on yours.
13. The 60-Second Self-Check Before You Buy
Five questions, answerable from your kitchen table, that resolve most chemistry debates before the sales call: How many days per year will this bank actually cycle? (Under 50, consider AGM.) Does anyone in the household do scheduled maintenance happily? (No rules out flooded.) Is the install location heated through winter? (No requires heated lithium or sealed lead.) What's the real budget — invoice, not monthly payment? (Be honest; the math only works with real numbers.) And what does a failure cost you — inconvenience or consequences? (Consequences buys warranty depth and brand support.) Write the five answers down and bring them in; the right chemistry is usually obvious from that page alone, and the backup kits page maps them to matched hardware.
