The short answer first: for most grid-storage projects reaching financial close in 2026, LFP is still the default — cheaper per usable kWh, proven, and bankable — and sodium-ion earns the shortlist in the specific niches where LFP gives ground: deep cold, heavy cycling, and lithium-price exposure. The rest of this page is the why, and where the exceptions live.
Ten years ago, “which chemistry?” was a genuine debate for storage projects. Today the honest answer for most grid-scale projects is “LFP, next question” — but the reasons why matter, the exceptions matter, and for the first time in years there’s a credible challenger arriving: sodium-ion.
Here’s the practitioner’s view of the chemistry landscape as it actually stands in mid-2026.
To ground the discussion, it helps to see what’s actually inside a cell. This Lithium-Ion Cell visual walks through the anode, cathode, separator, and electrolyte — the parts where a chemistry choice like LFP versus NMC actually lives.
Why LFP won
Lithium iron phosphate took over stationary storage for three compounding reasons:
1. Cost. LFP uses no nickel and no cobalt. In BNEF’s 2025 Lithium-Ion Battery Price Survey, average LFP pack prices across all segments were $81/kWh versus $128/kWh for NMC. For stationary storage specifically, average pack prices collapsed to $70/kWh in 2025 — a 45% drop in a single year — with the lowest observed LFP cell and pack prices at $36/kWh and $50/kWh. Stationary storage became the cheapest battery segment for the first time. The driver is blunt: BNEF estimates China produced around 557 GWh of stationary-storage cells in 2025 — more than double global installations in the sector. Overcapacity of that scale sets prices.
2. Cycle life. LFP cells routinely carry 6,000–8,000+ cycle ratings, with newer storage-dedicated cells claiming 10,000 or more. For an asset that may cycle daily for 20 years, this is decisive.
3. Thermal stability. The phosphate cathode’s strong phosphorus–oxygen bond means LFP releases far less oxygen during decomposition than nickel-based cathodes, so cells are harder to ignite and fail less energetically. Important nuance: LFP is not fire-proof, and its vent gas is hydrogen-rich, which makes explosion prevention — not fire spread — the first-order design issue. More on the standards that govern this in the NFPA 855 fire-safety framework.
Where NMC still appears
Nickel manganese cobalt chemistry dominates where energy density per kilogram matters — long-range EVs, aviation, consumer devices. In stationary storage, weight barely matters, so NMC’s advantage evaporates while its cost and safety burdens remain.
You’ll still encounter NMC in three places:
- Legacy fleets. Many projects built through the early 2020s used NMC. The Moss Landing 300 facility that burned in January 2025 combined NMC chemistry with an indoor building architecture — a pairing modern design practice has moved decisively away from.
- Korean and Japanese supply. Manufacturers there historically specialized in nickel chemistries and continue serving some storage demand.
- Tariff-distorted markets. BNEF has noted that with steep US tariffs on Chinese LFP, non-Chinese NMC can remain viable in US utility-scale projects until at least 2027.
Sodium-ion: 2026 is the year it got real
I’ve watched sodium-ion be “two years away” for a decade. That finally changed, and the milestones stacked up fast:
- April 2025: CATL launched its Naxtra sodium-ion brand — up to 175 Wh/kg at the cell level, an operating range of −40 °C to +70 °C, and the first sodium-ion cells certified to China’s new GB 38031-2025 traction-battery safety standard (which takes effect in mid-2026).
- December 2025: at its supplier conference, CATL confirmed large-scale 2026 deployment across four fields including energy storage — its “dual-star” sodium-plus-lithium strategy.
- April 2026: CATL unveiled what it calls the first platform-based sodium-ion battery designed specifically for energy storage — deliberately compatible with its 587 Ah lithium storage cell platform — and signed a three-year, 60 GWh supply agreement with system integrator HyperStrong, by far the largest sodium-ion order ever placed. First ESS deliveries are expected in September 2026.
Where sodium wins: extreme cold (CATL claims roughly 90% usable power at −40 °C and about three times the discharge power of an equivalent LFP pack in extreme cold — vendor figures, but directionally credible), very high cycle-life claims, abundant raw materials, and zero lithium-price exposure.
Where it doesn’t (yet): on paper, 175 Wh/kg is close to gravimetric parity with mainstream LFP — the real gaps are volumetric energy density (less energy per container footprint), a supply chain outside China that barely exists, and bankability: lenders have no fleet performance data. And sodium’s toughest competitor isn’t physics — it’s LFP’s own collapsing price. Treat 2026–2028 sodium projects as early-adopter territory where vendor-backed guarantees do the heavy lifting.
For the technology behind these headlines — how sodium-ion actually works, why it behaves so differently from lithium, and the full family tree of sodium chemistries — see Sodium Batteries, Part 1.
Sodium-ion vs. LFP for grid storage
Strip the field down to what a stationary project actually shortlists and it’s a two-horse race: NMC has mostly bowed out of storage and flow is a long-duration specialist, so the live question is sodium-ion vs. LFP. For a mainstream grid-storage project today, LFP is still the default — cheaper per usable kWh, a proven supply chain and warranty book, and a decade of fleet data lenders can underwrite. Sodium-ion earns the look exactly where LFP is weakest: sites that run cold, duty cycles heavy enough to reward its higher cycle-life claims, and buyers who want out of the lithium-price cycle.
The tie-breaker in 2026 isn’t energy density — at ~175 Wh/kg sodium is close to LFP by weight. It’s bankability and price: sodium still has almost no field track record and barely a supply chain outside China, while LFP keeps getting cheaper. So for a project reaching financial close this year, sodium-ion belongs on the shortlist as a cold-climate or hedge play with vendor-backed guarantees — not yet as the default.
Side by side, the trade looks like this:
| Criterion | LFP | Sodium-ion |
|---|---|---|
| Price signal | Stationary pack average $70/kWh in 2025 (BNEF), still falling | No fleet price index yet; competing head-on with LFP’s collapsing price |
| Cycle life | 6,000–8,000+ rated; storage-dedicated cells claim 10,000+ | Very high claims, but vendor claims — no fleet data behind them yet |
| Energy density | Higher per container footprint | ~175 Wh/kg is near parity by weight; the real gap is volumetric |
| Deep cold | Must heat cells before charging | ~90% usable power claimed at −40 °C; −40 to +70 °C operating range (vendor figures) |
| Track record | A decade of fleet data lenders can underwrite | First large ESS deliveries expected late 2026 |
| Supply chain | Mature, but China-dominated | Barely exists outside China |
| Commodity exposure | Rides the lithium-price cycle | Abundant materials, zero lithium exposure |
| Bankability | The default — lenders are comfortable | Early-adopter territory; vendor-backed guarantees carry the deal |
Flow batteries: the long-duration specialist
Vanadium redox flow batteries decouple power (stack size) from energy (tank size), tolerate essentially unlimited cycling, and use a non-flammable aqueous electrolyte. Their weaknesses: low energy density, higher upfront cost, and round-trip efficiency typically in the 65–75% range.
The practical rule of thumb: below roughly 4 hours of duration, lithium wins on cost almost everywhere. Somewhere in the 6–12+ hour range, flow batteries and other long-duration technologies start to compete — especially with heavy daily cycling and cheap land. China operates the flagship projects, including the 100 MW / 400 MWh Dalian system.
The selection framework
When I evaluate chemistry for a project, the shortlist test looks like this:
| Criterion | LFP | NMC | Sodium-ion | Vanadium flow |
|---|---|---|---|---|
| Cost today | Lowest | Higher | Falling fast, still above LFP | Highest |
| Cycle life | 6,000–10,000+ | ~3,000–5,000 | 10,000+ (claimed) | 20,000+ |
| Energy density | Good | Best | Near LFP by weight; behind by volume | Lowest |
| Cold weather | Weak in deep cold — charging below 0 °C is the constraint | Moderate | Excellent | Poor (electrolyte freezing risk) |
| Failure profile | Harder to ignite; H₂-rich vent gas | Most energetic thermal runaway | Promising, little field data | Non-flammable |
| Bankability | Proven | Proven | Emerging | Niche |
One more name worth knowing: LMFP (lithium manganese iron phosphate) adds manganese to the LFP recipe for higher voltage and energy density. It’s an evolution of LFP rather than a rival, and you’ll see it appear in storage cells over the next few years.
FAQ
Is LFP completely safe? No. It is meaningfully harder to ignite than NMC and fails less energetically, but LFP cells in thermal runaway still produce flammable, hydrogen-rich gas. Explosion prevention is the core design problem.
Will sodium-ion replace LFP? Not soon. CATL itself frames a “dual-star” strategy: sodium for cold climates, heavy cycling, and cost-sensitive niches; lithium for everything else. Ever-cheaper LFP is sodium’s biggest obstacle.
Is sodium-ion better than LFP for grid storage? For most grid-storage projects today, no — LFP is cheaper, proven, and bankable, and lenders have a decade of fleet data to underwrite. Sodium-ion earns the shortlist in specific niches: cold climates where LFP has to heat cells before charging, heavy cycling, and buyers hedging lithium-price exposure. As of 2026, treat it as a cold-climate or hedge play with vendor-backed guarantees, not the default.
What about solid-state batteries? Watch the EV space first. Nothing bankable exists for grid-scale storage today.
My Grid-Scale BESS: Complete Guide goes deeper on cell chemistry, degradation mechanisms, and how chemistry choices ripple into warranties and financing.