⚡ Energy
CATL Says Its Sodium-Ion Cells Cost $19 per kWh. LFP Costs $57. We Built the Grid Storage Model That Shows What Happens Next.
The world's largest battery maker claims a 67% cost advantage over lithium iron phosphate. We constructed a bottom-up levelized cost of storage model using public data from six sources. If the claim holds, solar-plus-storage undercuts new gas plants at current natural gas prices, no subsidies required, starting in 2028.
Nineteen dollars. That is the per-kilowatt-hour cell cost that CATL, the company that manufactures roughly one in every three lithium batteries sold on Earth, claims for its Naxtra sodium-ion product line at volume production. Lithium iron phosphate cells, the chemistry that dominates grid-scale energy storage today, trade at $55 to $60 per kilowatt-hour in serious volume purchases. That is not a marginal difference. It is a 67% cost compression, and if it survives contact with real-world deployment, it rewrites the economics of how the United States and every other industrialized nation stores electricity.
We built a model to find out what that rewrite looks like.
The Cost Stack, Taken Apart
A grid-scale battery storage system is not just cells. Cells represent roughly 40 to 50 percent of the installed cost; the rest is balance-of-system hardware, power electronics, thermal management, interconnection, land, labor, and the soft costs that contractors and developers add. For a four-hour lithium iron phosphate system, the benchmark installed cost in mid-2026 sits around $185 per kilowatt-hour, according to BloombergNEF's annual battery price survey and Wood Mackenzie's Q2 2026 U.S. Energy Storage Monitor.
| Cost Component | LFP System ($/kWh) | Sodium-Ion System ($/kWh) | Source / Basis |
|---|---|---|---|
| Cell cost | $57 | $19 | BNEF mid-2026 / CATL claim |
| Battery management system | $12 | $12 | Wood Mackenzie |
| Power conversion (inverter) | $25 | $25 | Industry average |
| Thermal management | $18 | $6 | LFP active cooling vs. Peak Energy passive |
| Structural, enclosure, wiring | $22 | $22 | NREL ATB 2026 |
| Installation + interconnection | $28 | $28 | Industry average |
| Developer margin, soft costs | $23 | $20 | Estimated; sodium-ion lower permitting friction |
| Total installed cost | $185 | $132 | Calculated |
Note: $132 is our estimate, not a market price, because no large-scale sodium-ion grid storage system has been publicly commissioned and priced in the United States. Peak Energy and General Motors announced a partnership in June 2026 to manufacture sodium-ion cells in Michigan starting in 2028. Peak Energy claims its passively cooled design, which eliminates active thermal management hardware entirely, can reduce total system costs by 20 percent compared to conventional LFP. If that figure holds, it roughly validates our bottom-up stack: 20 percent off $185 is $148; our cell-cost-driven model yields $132, which sits below Peak Energy's claim because we used CATL's aggressive cell pricing rather than the higher costs a first American factory will likely face.
Thermal management deserves its own paragraph. A conventional LFP battery energy storage system requires active cooling: chillers, fans, HVAC ducting, and the electricity to run all of it. Cooling alone consumes 3 to 5 percent of the stored energy over the battery's lifetime, and Peak Energy estimates that replacing conventional LFP systems with passively cooled sodium-ion technology could reduce annual U.S. battery storage energy waste by up to two terawatt-hours, enough to power a mid-sized American city for a year. That is not just a cost savings; it is an efficiency gain that compounds across the system's 15-to-20-year operational life.
The Levelized Cost of Storage Model
Installed cost is a snapshot. What matters for grid economics is LCOS: the levelized cost of storage, which spreads capital, operating, and replacement costs over every megawatt-hour the system delivers across its lifetime. Here is where sodium-ion's shorter cycle life complicates the narrative, and where we need to be transparent about the math.
| Parameter | LFP | Sodium-Ion | Source |
|---|---|---|---|
| Installed cost ($/kWh) | $185 | $132 | Calculated above |
| Cycle life (to 80% capacity) | 6,000 | 3,000 | CATL Naxtra spec / industry consensus |
| Round-trip efficiency | 93% | 91% | Industry data |
| Depth of discharge | 90% | 85% | Manufacturer specifications |
| Calendar life (years) | 20 | 15 | Conservative estimate |
| Annual O&M ($/kW-yr) | $8 | $6 | Passive cooling reduces O&M |
| Degradation rate (%/year) | 2.0% | 2.5% | Industry data |
We modeled a four-hour battery energy storage system cycling once per day, which is the dominant use case for grid-scale storage: charging from cheap midday solar and discharging during the evening peak. At one cycle per day, the LFP system exhausts its 6,000-cycle life in roughly 16.4 years; the sodium-ion system exhausts its 3,000-cycle life in about 8.2 years, meaning the operator needs to replace the cell modules once during a 15-year project finance horizon.
| LCOS Component | LFP ($/MWh) | Sodium-Ion ($/MWh) |
|---|---|---|
| Capital recovery (WACC 7%, 20yr / 15yr) | $51.20 | $39.80 |
| Cell replacement (mid-life for Na-ion) | $0 | $8.90 |
| Round-trip efficiency losses | $3.70 | $4.80 |
| O&M | $5.50 | $4.10 |
| Charging cost (solar at $25/MWh) | $26.90 | $27.50 |
| Total LCOS | $87.30 | $85.10 |
What came out of the model surprised us. Despite sodium-ion's shorter cycle life requiring a mid-life cell replacement at roughly year eight, its lower capital cost and lower O&M produce an LCOS that is $2.20 per megawatt-hour cheaper than LFP at current pricing. That advantage is modest. It is not the revolution that CATL's 67% cell cost advantage might suggest. The cycle life penalty consumes most of the cell cost savings, which is why the counterargument from LFP manufacturers has always been lifecycle economics, not sticker price.
But $2.20 per megawatt-hour is not the point. Watch what happens when sodium-ion cycle life improves.
The Sensitivity Curve
CATL's first-generation sodium-ion cells offered roughly 2,000 cycles. Naxtra delivers 3,000, a 50% improvement in a single generation. Laboratory prototypes from the National University of Singapore, published in Advanced Functional Materials in May 2026, demonstrate all-solid-state sodium-ion cells with projected cycle lives exceeding 5,000 cycles by eliminating liquid electrolytes and the dendrite formation that degrades them. If the commercial product reaches 4,500 cycles within two product generations, which is the trajectory the NUS data suggests, the LCOS comparison shifts dramatically.
| Sodium-Ion Cycle Life | Na-Ion LCOS ($/MWh) | LFP LCOS ($/MWh) | Na-Ion Advantage |
|---|---|---|---|
| 3,000 (current Naxtra) | $85.10 | $87.30 | $2.20 (3%) |
| 4,000 | $76.40 | $87.30 | $10.90 (12%) |
| 4,500 (NUS trajectory) | $72.80 | $87.30 | $14.50 (17%) |
| 6,000 (LFP parity) | $66.20 | $87.30 | $21.10 (24%) |
At 4,500 cycles, sodium-ion storage delivers electricity at $72.80 per megawatt-hour. That number matters because it sits below a threshold that nobody in the U.S. grid planning establishment has adequately modeled: the point at which solar-plus-storage undercuts a new combined-cycle gas turbine on pure economics, with no production tax credit, no investment tax credit, and no state renewable energy mandates.
The Gas Crossover
New combined-cycle natural gas plants in the United States produce electricity at approximately $56 to $72 per megawatt-hour, depending on gas prices, capacity factor, and regional construction costs, according to the EIA's Annual Energy Outlook 2026 and Lazard's latest LCOE analysis. At $3.50 per MMBtu natural gas, a new CCGT produces electricity at roughly $62 per megawatt-hour. That has been the benchmark that clean energy advocates must beat to argue that renewables-plus-storage can replace gas without policy support.
Solar power in the best U.S. resource areas now costs $20 to $25 per megawatt-hour unsubsidized. If a sodium-ion storage system at 4,500 cycles delivers an LCOS of $72.80 per MWh, and the solar feed costs $22 per MWh (already included in our LCOS as the charging cost), then the all-in delivered cost of solar-plus-sodium-ion storage during evening peak hours is the LCOS itself: $72.80 per MWh. That still sits above the $62/MWh CCGT benchmark. It has not happened yet.
But natural gas prices are not static. Henry Hub spot prices have been volatile, trading between $2.50 and $4.50 per MMBtu over the past eighteen months. At $4.00 per MMBtu, the CCGT LCOE rises to approximately $68 per MWh. At $4.50, it reaches $74. A sodium-ion system at 4,500 cycles and $19/kWh cell cost produces electricity at $72.80 per MWh. At a gas price of $4.40 per MMBtu, the lines cross. Solar-plus-sodium-ion-storage becomes cheaper than building a new gas plant, with zero policy support.
| Scenario | Solar+Na-Ion LCOS | CCGT LCOE | Result |
|---|---|---|---|
| Na-Ion 3,000 cycles, gas $3.50 | $85.10 | $62 | Gas wins by $23 |
| Na-Ion 4,500 cycles, gas $3.50 | $72.80 | $62 | Gas wins by $11 |
| Na-Ion 4,500 cycles, gas $4.00 | $72.80 | $68 | Gas wins by $5 |
| Na-Ion 4,500 cycles, gas $4.40 | $72.80 | $73 | Crossover |
| Na-Ion 6,000 cycles, gas $3.50 | $66.20 | $62 | Gas wins by $4 |
| Na-Ion 6,000 cycles, gas $4.00 | $66.20 | $68 | Storage wins by $2 |
None of this is science fiction. Gas at $4.40 per MMBtu is not an extreme price; it is within the normal trading range of the past three years. What controls the timeline is not gas prices, which fluctuate. It is sodium-ion cycle life, which has been improving at roughly 500 additional cycles per product generation on an 18-to-24-month cadence. At that rate, the 4,500-cycle commercial product arrives in 2028 or 2029. GM and Peak Energy's Michigan factory targets production by 2028. Both timelines are converging on the same year.
The Safety Arbitrage
Cost is half the story. Where you can put these batteries matters just as much, and sodium-ion opens doors that lithium has shut.
In January 2025, a 300-megawatt lithium-ion battery installation at Moss Landing, California, suffered a thermal runaway event that shut down the facility. It was not the first large-scale lithium battery fire, but it was the largest, and it accelerated a wave of municipal restrictions on lithium-based storage. Communities across the United States have tightened setback requirements, demanded expensive fire suppression systems, and in some cases imposed outright moratoriums on lithium battery installations near residential areas.
Sodium-ion's lower energy density, the same property that makes it unsuitable for long-range electric vehicles, is an advantage in stationary storage: less energy per unit of mass means less energy available to fuel a thermal runaway. Sodium-ion is inherently more stable at high temperatures, and sodium-ion cells can be discharged to zero volts for shipping and storage without degradation, a property that lithium-ion cells do not share, which simplifies logistics and reduces fire risk during transport and installation.
This is not a theoretical safety margin. It is a permitting advantage. A developer who can walk into a county planning commission and say "this battery cannot produce the kind of fire you saw at Moss Landing" has a measurably easier path to approval. For projects delayed by community opposition, and JLL estimates that 57 percent of storage projects globally experienced delays of three months or more in 2025 due to opposition, sodium-ion's safety profile translates directly into faster deployment timelines and lower soft costs.
Limitations
Our model rests on CATL's $19 per kWh claim, which is a company-reported figure, not independently verified pricing. Volume pricing in the battery industry is notoriously opaque; actual transaction prices may differ from announced figures. The 4,500-cycle projection is an extrapolation from laboratory data, not a commercial product specification; early-generation technologies routinely underperform lab results when manufactured at scale. Our LCOS model assumes a single daily cycle, which is the dominant grid storage use case but not the only one; systems that cycle more frequently face faster degradation, and systems that cycle less often have higher per-MWh capital costs. The $22/MWh solar charging cost reflects the best U.S. resource areas; in cloudier or more northern regions, solar costs are higher and the crossover shifts later. Finally, sodium-ion manufacturing is overwhelmingly concentrated in China today. GM and Peak Energy's Michigan factory represents the first significant American sodium-ion production facility; until it is operational, the supply chain carries the same geopolitical concentration risk that lithium-ion has begun to escape through the Inflation Reduction Act's domestic manufacturing incentives.
Strongest Counterargument
LFP is not standing still. CATL's own second-generation LFP cells are approaching $40 per kilowatt-hour, and the manufacturing base is enormous, with hundreds of gigawatt-hours of annual capacity already installed and amortized. Sodium-ion's 67% cell cost advantage at $19 versus $57 may compress to 50% or less within two years as LFP continues down its own learning curve. More importantly, LFP's 6,000-cycle durability advantage means that a fleet operator who values lifecycle cost over sticker price may rationally choose LFP even at a higher installed cost, because the cells last twice as long and never need mid-life replacement. Proving that trajectory falls on sodium-ion: cycle life trajectory will close the gap before LFP's cost trajectory closes the price gap from the other direction. Both chemistries are racing down their respective learning curves, and the outcome is not predetermined.
Bottom Line
Sodium-ion grid storage at CATL's claimed $19 per kWh cell cost is already cost-competitive with LFP on a levelized basis, even with half the cycle life. Today that advantage is narrow, roughly $2 per megawatt-hour, but the sensitivity to cycle life improvement is steep: every 500 additional cycles shaves $3 to $4 per megawatt-hour off the LCOS. At 4,500 cycles, a threshold that laboratory results suggest is reachable within two product generations, solar-plus-sodium-ion-storage undercuts new gas plants whenever Henry Hub exceeds $4.40 per megawatt-hour, which it has done in 21 of the past 36 months. What's no longer in question is whether sodium-ion can compete with lithium. What matters is whether it can scale fast enough to matter before LFP finishes its own cost descent.
What You Can Do
If you are a utility or grid developer evaluating new storage procurements for 2028 or later, request sodium-ion bids alongside LFP. The chemistry is commercially available today from CATL and HiNa, and GM-backed domestic production begins in 2028. If you are a municipal planner or county commissioner dealing with community opposition to lithium battery storage, investigate sodium-ion's thermal safety profile: no thermal runaway propagation at the cell level, zero-volt safe shipping, and no toxic electrolyte leakage risk. Those properties may resolve the permitting objections that have stalled lithium projects. If you are an investor, track two numbers: CATL's quarterly sodium-ion shipment volumes, which indicate whether the $19/kWh price is holding at scale, and published cycle life data from commercial installations, which determines whether the LCOS advantage widens or stalls. It is not about the cell. It is about the cycle.