Global Sodium-Ion Battery Materials Market Size, Share & Forecast 2026–2033

Forecast Snapshot 2026–2033

Market overview

The global sodium-ion battery materials market is valued at USD 162.6 million in 2025 and is projected to reach USD 5,376.5 million by 2033, a 54.86% baseline CAGR. That is a small number, and the size is the first finding: the chemistry that everyone describes as the answer to lithium is, in materials terms, 2.0% of the LFP cathode market it is meant to displace — USD 162.6 million against USD 8.30 billion.

Volume is 29,218 tonnes in 2025, rising to 1,789,500 tonnes by 2033 at a 67.25% CAGR. Tonnage therefore compounds 12.4 points faster than value. That gap is not a rounding artefact and it is not an assumption about competitive intensity: it is the arithmetic consequence of an industry whose entire commercial claim is that it will get cheaper. Every published price line in this chain is falling on purpose, and the fall is the product.

Two markets sit under this one number and they behave differently. Cathode active material was USD 89.9 million of the 2025 total, priced by three competing routes with three separate price levels. Hard carbon anode was USD 72.7 million — 44.70% of the market from 31.55% of the tonnage — because at RMB 56,390 a tonne it is the most expensive material in a sodium-ion cell and the only one whose price nobody can independently confirm.

The 2025 base: 8.1 GWh, and not 47.2

Published figures for 2025 sodium-ion output disagree by a factor of nearly six. Chinese industry shipment data puts it at 8.1 GWh. A separate 2026 industry compilation puts global installations at 47.2 GWh, up more than 150% year on year. One of these is wrong, and there is a way to settle it without choosing between institutions.

Cathode tonnage per gigawatt-hour is a property of the chemistry, not of the factory. Sodium-ion cathode powder delivers roughly 130 mAh/g at about 3.1 volts for the polyanion route and 150 mAh/g at 3.6 volts for layered oxide, which puts the requirement at 2,430–2,470 tonnes of cathode per GWh regardless of who is assembling the cell. Reported cathode shipments for 2025 were 20,000 tonnes. Twenty thousand divided by 2,450 is 8.16 GWh. The prior year offers a second, independent test: 9,000 tonnes of cathode against reported shipments of 3.7 GWh gives 2,432 tonnes per GWh. The 47.2 GWh figure would require 115,600 tonnes of cathode for 2025 — which is very nearly the 120,000 tonnes the same compilation reports for 2026.

That is the resolution. 47.2 GWh and 120,000 tonnes belong to the same year, and it is not 2025. The base year here is 8.1 GWh because two independent measurements of a chemical constant agree with it and the alternative is internally inconsistent with itself.

What a gigawatt-hour of sodium-ion buys

Sodium-ion is a materials-intensive chemistry, and that is its second commercial problem after price. At a mean cell energy density of roughly 144 Wh/kg — the central value across a survey of producers reporting 95 to 175 Wh/kg — a gigawatt-hour weighs about 6,900 tonnes. Of that, 3,579 tonnes is active material: 2,450 tonnes of cathode powder and 1,129 tonnes of hard carbon.

The anode is where the chemistry pays. Hard carbon stores roughly 300 mAh/g against graphite's 350–360, and sodium-ion cells carry a higher anode-to-cathode mass ratio than LFP because the cathode capacity is lower: at a cathode of 130 mAh/g and an N/P ratio of 1.05, the anode is 0.455 times the cathode mass. The observed 2025 ratio was 0.4609. Those two numbers were derived by different routes — one from electrochemistry, one from reported tonnage — and they agree to within 1.3%.

The price lines, and the one that has no price

Chinese domestic trade quotes for the sodium-ion chain in May 2026, per tonne of material:

The route that carries 70% of the tonnage is the one without a published price series. Polyanion, and NFPP within it, is quoted by participants and by cost studies, not by a weekly index. Layered oxide and Prussian blue are indexed. For a report that has to publish a blend, that asymmetry is the single largest source of error in the number above, and it is disclosed rather than smoothed.

Method and limits

Volume is bottom-up from reported cathode shipments by route — 14,000 tonnes polyanion, 5,000 layered oxide and 1,000 Prussian blue in 2025 — plus hard carbon derived at the observed anode-to-cathode mass ratio of 0.4609, giving 9,218 tonnes. The 2033 volume is cell output of 500 GWh multiplied by 3,579 tonnes of active material per GWh. The intensity construction alone would put 2025 at 28,990 tonnes against the reported 29,218; the 228-tonne gap, 0.8%, is the only place the two methods disagree and it is stated rather than netted out.

Value is tonnage multiplied by price. Cathode price is a weighted blend of the three route prices applied to the reported route tonnage, at RMB 28,000 for polyanion, RMB 45,877.50 for layered oxide and RMB 21,667.50 for Prussian blue, giving RMB 32,152.75 per tonne in 2025. The 2033 blend assumes polyanion at 88% of tonnage, NFPP at RMB 18,000, layered oxide at RMB 32,000 and Prussian blue at RMB 18,000 — a 39.7% decline in the blended cathode price over eight years. Hard carbon is RMB 56,390 in 2025 and RMB 26,000 in 2033, a 53.9% decline, which is the low end of what producers themselves say they expect. All values converted at a constant RMB 7.15 per USD.

Excluded: sodium carbonate, iron phosphate, sodium dihydrogen phosphate and glucose as merchant products, which are precursors inside the cathode price rather than separate markets here; electrolyte and sodium hexafluorophosphate, which are a separate market and sized elsewhere; cells, modules, packs and systems; current collectors, separators and casings; and LFP and LMFP cathode chemistry, which appear as the incumbent and the substitution risk rather than as volume. Cell output of 8.1 GWh in 2025 carries an estimated ±15% band, because the two independent measurements that corroborate it are both survey data rather than audited shipment records.

Related research

This report is the material layer beneath Global EV Battery Market (USD 91.7 billion in 2025 to USD 369.5 billion in 2033, 19.03%) and it is the substitution named in Global LFP Cathode Active Material Market (USD 8.30 billion to USD 26.20 billion at 15.45%), which sizes the chemistry this one displaces. It draws its anode-side raw material from the iron phosphate chain covered in Global Phosphorous Acid & Phosphonate Derivatives Market, and its largest end market is Global Stationary Energy Storage Systems (ESS) Market (USD 64.54 billion to USD 247.68 billion at 18.31%), which already names sodium-ion in its own perimeter.

Competitive Landscape

A 20,000-tonne cathode market with a top-ten concentration of 48.9% looks fragmented and is. The leader shipped 4,000 tonnes, which is 28.57% of polyanion tonnage and 20.0% of all sodium-ion cathode. Four thousand tonnes is a plant, not a company. The second through tenth positions in a pool that size are separated by a few hundred tonnes each, and the annual ranking is not stable enough to be a ranking.

That is about to stop being true, and the mechanism is entry from the lithium side. Ronbay Technology has 6,000 tonnes of polyanion capacity at Xiantao in Hubei, is expanding to 28,000 tonnes in 2026, and has published a plan for 300,000 tonnes of sodium-specific line by 2027. The 2027 figure is fifteen times the entire 2025 global sodium-ion cathode market. It is not a forecast; it is a statement about which balance sheet intends to own the route. Ronbay also claims a 30% to 50% reduction in polyanion processing cost, achieved by building process and equipment specifically for sodium rather than converting an LFP line. A cathode business that keeps 56.79% of its selling price as conversion spread is an invitation, and it has been accepted.

On the cell side the structure is the opposite and that is the problem for everyone upstream. The 2025 shipment leaders were Veiking Technology, Highpower, CATL, FinDreams and Pylontech. CATL has signed a three-year, 60 GWh sodium-ion storage order with HyperStrong. BYD’s Xuzhou plant is specified at 30 GWh. Five customers, each of them larger than the entire merchant material base that supplies them, negotiating with a cathode industry whose top ten hold less than half of 20,000 tonnes. The conversion spread described in the value-chain chapter is not a moat. It is a subsidy being paid by five buyers who can end it with a sourcing decision.

Capacity is not the constraint and the gap between announcement and commissioning is the competitive fact. At the end of 2025 China had announced roughly 275.8 GWh of sodium-ion cell capacity and had commissioned roughly 36.5 GWh — a realisation rate of 13.23%. Through 2026 total capacity is above 80 GWh against output of 47.2 GWh, or 59.0% utilisation. The industry is not short of plans. It is short of qualified lines and it is running the ones it has at well under nameplate.

Hard carbon is where the overbuild is concentrated. Planned capacity stands at 287,000 tonnes a year against effective capacity of 63,000 tonnes, up from 18,000 at the end of 2025 — a 250% increase in twelve months. Industry utilisation is 25.9%. Expected 2026 shipments of 32,000 tonnes against 63,000 tonnes of effective capacity is 50.79%, which is the number that matters: the dedicated hard-carbon fleet could serve twice the demand it has. That is not a bottleneck. That is a price collapse waiting for a delivery schedule, and it is the reason the 2033 hard carbon price in this report is RMB 26,000 rather than the RMB 35,000 to 40,000 the industry is promising for 2026.

Outside China the industry is an order of magnitude smaller and it is not catching up. Peak Energy in the United States is building NFPP cells for storage and targets more than 20 GWh of annual capacity by 2030, with first production in 2026 or 2027. Tiamat in France is targeting about 0.7 GWh in its first phase in 2026 and 5 GWh for the full gigafactory around 2030. Ningbo Jinglan has signed for 5 GWh, 2 GWh of it due in December 2025. Novasis is producing Prussian blue cells at 130 Wh/kg. Against that, the Chinese base commissioned 36.5 GWh by the end of 2025. The European and American entrants hold patents and qualification positions and almost no tonnage, which is a defensible position in a market that is supply-constrained and an indefensible one in a market that is not.

The competitive question for the next three years is not who leads sodium-ion. It is whether the material layer of a Chinese-dominated industry can hold a conversion spread while the cell layer consolidates and the lithium cathode incumbents arrive with capacity that is fifteen times the current market. On the evidence of the 2024-to-2025 route war, the answer is that it will not hold it for long.

Value Chain

One tonne of NFPP is a formula before it is a product. Sodium iron pyrophosphate runs Na4Fe3(PO4)2P2O7, molar mass 623.39 grams. Three of its four phosphorus atoms arrive through iron phosphate and one through sodium dihydrogen phosphate, so the input bill per tonne of finished powder is 0.726 tonnes of iron phosphate, 0.340 tonnes of sodium carbonate and 0.192 tonnes of sodium dihydrogen phosphate. That is 1.258 tonnes of purchased chemical per tonne of product, and the arithmetic is molar mass rather than assumption.

The cost line proves the arithmetic. At RMB 11,650 for iron phosphate, RMB 4,492.50 for battery-grade sodium carbonate and RMB 5,930 for sodium dihydrogen phosphate, the three inputs cost RMB 8,458, RMB 1,527 and RMB 1,142 — RMB 11,127 a tonne, or 91.96% of the RMB 12,100 raw-material cost that a Chinese brokerage estimate puts on NFPP. The reported structure says iron phosphate is about 72% of that cost; the stoichiometry says 69.90%. The difference is within the measurement, and the residual RMB 973 a tonne, 8.04%, pays for the carbon source and the processing aids.

The alkali saving is real, and it is smaller than the industry implies. At 2,450 tonnes of cathode per GWh, one kilowatt-hour of sodium-ion cell carries 0.833 kilograms of sodium carbonate. At RMB 4,492.50 a tonne that is RMB 3.74 a kWh, or RMB 0.0037 a Wh. An LFP cell carries 0.5975 kilograms of lithium carbonate at 0.2342 tonnes per tonne of cathode, which at RMB 55,800 is RMB 33.34 a kWh, or RMB 0.0333 a Wh. Substituting sodium for lithium therefore saves RMB 0.0296 a Wh on the alkali line — an 88.78% reduction, and the single largest structural advantage the chemistry has.

Set it against the number the cell makers themselves quote. The sodium-to-LFP cost gap is put at RMB 0.10 to 0.15 a Wh. The entire sodium-for-lithium substitution is worth 19.7% to 29.6% of that gap. The other 70% to 80% has to come from somewhere else, and where it has to come from is the anode.

Hard carbon is the expensive half. At 1,129 tonnes per GWh and RMB 56,390 a tonne, hard carbon costs RMB 63.66 million per GWh, or RMB 0.0637 a Wh. Cathode at 2,450 tonnes and RMB 28,000 costs RMB 68.6 million per GWh, or RMB 0.0686 a Wh. Active material is therefore RMB 0.1323 a Wh, or 20.35% of a quoted storage cell at RMB 0.65. Hard carbon alone is 48.15% of the active-material cost and 9.80% of the cell price — from 31.55% of the tonnage. That is the inversion this market runs on: the cheap chemistry has an expensive anode.

**The cathode spread is fatter than LFP’s, and that is the competitive opportunity and the competitive warning.** NFPP at RMB 28,000 against RMB 12,100 of raw material leaves 56.79% of the selling price for carbon coating, firing, milling, yield and margin. An LFP cathode maker at RMB 18,000 a tonne with RMB 11,137 of precursor inside it keeps 38.13% of the fee. The sodium cathode business is a materially better spread than the lithium one, because it is earlier on the cost curve and nobody has yet competed it away. Nothing about the chemistry defends it.

The iron phosphate is shared with LFP, and that is the supply-chain linkage that matters. NFPP needs 0.726 tonnes of iron phosphate per tonne against LFP’s 0.956 — 24.06% less, because three quarters of the phosphorus load travels through the precursor and the fourth arrives separately. But it is the same precursor, bought from the same plants, qualified to the same tap density and magnetic-foreign-matter limits. Sodium-ion cathode plants are built on conversion of existing LFP lines: same solid-state route, spray drying and high-temperature sintering, short retooling cycles. The industry does not need a new supply chain to make sodium cathodes. It needs the one it already has, which is why the capital arrives faster than the demand and why the spread above will not last.

The perimeter, stated here rather than in a footnote: this chapter prices merchant cathode and merchant hard carbon. Sodium carbonate, iron phosphate, sodium dihydrogen phosphate and glucose are inputs inside the cathode price, not separate markets here. Electrolyte and sodium hexafluorophosphate are excluded and sized elsewhere. The export price of a sodium-ion cathode is not published anywhere, which means every number above is a domestic Chinese transaction and the international market for these materials does not yet have a price.

Investment Activity

The capacity being announced is not the capacity being built, and the gap is the investment case.

The unit economics, from the one project with a disclosed ticket. HiNa Battery signed a sodium-ion base in Wuhan in February 2026 worth RMB 5 billion for 10 GWh. That is RMB 500 million per gigawatt-hour, or USD 69.9 million. At that unit cost, 500 GWh of cell capacity by 2033 implies roughly RMB 250 billion, or USD 35.0 billion, of cell capital — before a single tonne of cathode or hard carbon capacity is financed. This is a capital-intensive industry being entered by companies whose disclosed projects are one to two orders of magnitude smaller than the incumbents’ balance sheets.

What has actually been built. At the end of 2025, announced Chinese sodium-ion cell capacity was about 275.8 GWh against roughly 36.5 GWh commissioned — 13.23% realised. The 2026 installed base exceeds 80 GWh. The pattern is consistent across the chain: HiNa’s Fuyang line is 1 GWh, completed in July 2024; BYD’s Xuzhou plant is 30 GWh announced in 2024 with no delivery date; CATL has built a sodium-specific line and says it will deliver its first storage systems in September 2026 with GWh-level shipments for the year. One of the three largest battery companies in the world is guiding to GWh-level output in the first year of commercial delivery. That is the scale this market is actually at.

The material-side capital is where the leverage sits. Hard carbon effective capacity went from 18,000 tonnes at the end of 2025 to 63,000 tonnes in 2026 against 287,000 tonnes planned and utilisation at 25.9%. Cathode capacity is being added even faster: 6,000 tonnes to 28,000 tonnes at Ronbay in one step, with a 300,000-tonne plan behind it. To serve 500 GWh in 2033 the chain needs 1,225,000 tonnes of cathode and 564,500 tonnes of hard carbon a year — 43.75 times the cathode capacity being commissioned in 2026 and 8.96 times the effective hard carbon base. Those multiples look enormous and they are the honest description of a market that has to grow 60-fold.

What is worth underwriting and what is not. The precursor is worth underwriting: iron phosphate is shared with the LFP industry, qualified to published tap-density and magnetic-foreign-matter limits, and the same plants serve both chemistries. Whoever owns low-cost, high-purity iron phosphate owns the input to both branches of the phosphate cathode business, and sodium-ion adds demand to a base that already exists rather than requiring a new one. Carbonisation capacity is worth underwriting once the price settles, because the coal-based and resin-based routes escape the biomass constraint and the dedicated-line fleet is running at half its effective capacity.

The cathode blending step is not worth underwriting at a 56.79% conversion spread against an entrant planning fifteen times the current market. That spread is the return on being early, and being early is not a durable asset.

What consolidation will look like. The precedent is in the adjacent chemistry: ICL paid USD 285 million for an LFP cathode position, which is the price of a qualification, not of a plant. In sodium-ion the same logic will apply earlier, because the qualified-supplier list is shorter, the cell customers are fewer, and the technology is not protected by an export control. The assets that get bought will be the ones holding a customer qualification at a tier-one cell maker, and the price will be set by what it costs a competitor to obtain that qualification independently rather than by replacement cost.

The one number that should govern any commitment here. 86.77% of announced Chinese cell capacity had not been commissioned at the end of 2025. Capital is not scarce in this market. Commissioned, qualified, yielding capacity is scarce, and the difference between those two things is where every investment in this chain will be won or lost.

Technology & Innovation

The route war is over and it was decided in twelve months. Polyanion cathode went from 3,000 tonnes and 33.3% of sodium-ion cathode tonnage in 2024 to 14,000 tonnes and 70.0% in 2025, up 366.6%. Layered oxide fell 16.6% to 5,000 tonnes and 25.0%. Prussian blue held 1,000 tonnes and 5.0%. Within polyanion, NFPP alone is 68% of global shipments. A technology transition that normally takes a decade happened between two annual counts.

The price explains it and the price is not the whole story. NFPP is quoted at RMB 28,000 a tonne against RMB 45,877.50 for layered oxide, a 38.97% discount on the cathode line. On a storage project buying hundreds of megawatt-hours, that is the largest single controllable line in the bill of materials. What the buyer gives up is energy density: 130 Wh/kg for NFPP cells against 160 to 175 for layered oxide. Across a survey of producers the reported range is 95 to 175 Wh/kg with a mean of 144 and a median of 150.

That trade is only rational for a customer who does not care about kilograms. More than 60% of 2025 sodium-ion cell demand was stationary storage, where volume and weight are site costs rather than product costs, cycle life is revenue, and cold-weather capacity retention decides whether the project is financeable at all. CATL’s sodium storage system is specified at 15,000 cycles, a 20-year system life and 92% capacity retention at minus 20 degrees Celsius. NFPP is quoted above 15,000 cycles. Layered oxide is not.

Where the technology frontier actually sits is the anode, and almost nobody says so. Hard carbon stores sodium at roughly 300 mAh/g against graphite’s 350 to 360 for lithium, and the disordered, porous structure that makes that possible is also what makes it hard to manufacture: moisture control, first-cycle irreversible capacity and gas generation in the first formation cycles. CATL, describing its own mass-production problems, named hard carbon electrode moisture control and gas generation specifically rather than any cathode issue. First-cycle efficiency is the number that matters and the industry is trying to hold it above 90%, because every point below that is sodium bought and thrown away.

The precursor question is a feedstock question. Most hard carbon is biomass-derived — coconut shell is quoted at RMB 5,415 a tonne — and biomass carries two problems the chemists cannot solve: seasonal supply and batch-to-batch performance variation. The industrial answer is to stop using biomass. Wanhua Chemical has put two engineered routes into production, one coal-based and one resin-based, explicitly to escape the supply and consistency limits of natural feedstock. The coal route has the cost advantage; the resin route has the consistency and the performance. That is a petrochemical company treating an anode as a formulated product, and it is the direction the industry has to go if hard carbon is to reach the RMB 35,000 the cell makers need.

Prussian blue is the cheapest cathode per tonne and it is losing. At RMB 21,667.50 it undercuts NFPP by 22.6% and layered oxide by more than half. It has held 1,000 tonnes and 5% because the economics of the cathode line are not the constraint: cyanide-bearing process handling, low tap density and a sodium-ion cell built around it that tops out near 130 Wh/kg make it a niche for stationary backup rather than a route to scale.

The number that governs this whole report is a chemistry constant, and that is the technology risk nobody prices. Cathode requirement per gigawatt-hour is set by specific capacity and operating voltage, not by plant design. At 130 mAh/g and 3.1 volts, a kilowatt-hour needs 2.48 kilograms of NFPP; at 150 mAh/g and 3.6 volts, a layered oxide cell needs 2.14 kilograms. The forecast here uses 2,450 tonnes per GWh. Every improvement in specific capacity reduces the tonnage this report sizes while leaving the cell market untouched. A 10% gain in delivered cathode capacity takes 245 tonnes per GWh out of the market at unchanged energy output. There is no other materials market in this estate where the technology roadmap is a volume headwind of that size, and it is the reason the value line compounds at 54.86% while the cell line compounds at 67.42%.

The other half of the roadmap points the other way. CATL’s second-generation sodium product targets above 200 Wh/kg in 2026, against an LFP range of 160 to 180. If a sodium cell reaches LFP energy density while carrying a cheaper alkali and an aluminium anode current collector, the addressable market stops being storage and becomes the bottom of the vehicle market. That is the upside case, and it is worth being precise about what it does to this report: it raises the cell volume and it lowers the tonnes per cell, and the two effects do not cancel.

Market Risk

The largest risk in this report is that the cost claim is a projection and the price is a fact, and they currently disagree.

Risk one: the spread between claimed cost and quoted price is implausible. Sodium-ion cell manufacturing cost is put at RMB 0.35 to 0.40 a Wh in the first quarter of 2026, with RMB 0.40 the figure for a 1 GWh line and a path below RMB 0.35 by the end of the year. The quoted price for the product storage actually buys, NFPP prismatic cells at 160 to 180 Ah, is RMB 0.65 a Wh. The spread between the quote and the claimed cost is 62.5% to 85.7%. No lithium cell maker in the world earns that gross margin, and the most plausible explanations are both uncomfortable: either the cost figure is a target rather than a cost, or the quote is a small-batch trade price for a product with a handful of suppliers. Either way, the industry is asking the market to accept a parity claim whose two halves have never been observed in the same transaction.

Risk two: the sodium-ion price is not one price, and the product you buy decides whether the chemistry is cheap. The quoted ladder in May 2026 runs from RMB 0.53 a Wh for layered-oxide cylindrical small-power cells to RMB 1.21 a Wh for layered-oxide start-stop cells — a factor of 2.28 across products that all describe themselves as sodium-ion. NFPP storage cells are RMB 0.65 and NFPP start-stop cells are RMB 0.91. Against a quoted LFP storage cell in the RMB 0.50 to 0.55 range, implied by the RMB 0.10 to 0.15 gap the cell makers quote, the NFPP storage cell that carries the volume thesis is 18.2% to 30.0% more expensive than the technology it is supposed to undercut. The forecast here does not require sodium to be cheaper than LFP until 2027, and it does require it thereafter.

Risk three: hard carbon has to deflate on schedule and the capacity base argues against it. Three published claims describe the same product in the same window: an average of RMB 86,000 a tonne for 2025, a traded high-grade quote of RMB 56,390 in May 2026, and an expected 2026 average of RMB 35,000 to 40,000. The first and third differ by a factor of 2.46. The forecast in this report assumes RMB 26,000 by 2033, which is the low end of what producers promise, and the reason to believe it is 25.9% utilisation. The reason to doubt it is that an industry running at a quarter of capacity and still quoting RMB 56,390 is not an industry whose price is falling on a schedule.

Risk four: the tax exemption expires two thirds of the way through the forecast. Sodium-ion pays no consumption tax until 31 December 2028 while lithium pays 4%. At 84.4% of the 2033 tonnage arriving after that date, the central years of this forecast depend on a chemistry that must be cheaper than LFP on its own cost base rather than on a 4% advantage it will no longer have. If the exemption is extended, the forecast is conservative. If it is not, the 2029 to 2031 price line in this report is optimistic by the value of the tax.

Risk five: the tonnage is the technology risk. Cathode requirement per gigawatt-hour is a function of specific capacity and voltage. A 10% gain in delivered cathode capacity removes 245 tonnes per GWh at unchanged energy output, which is 10% of the volume in this report with no change to the cell market. A chemistry whose energy density roadmap is the industry’s principal selling point is a chemistry whose materials market shrinks every time the roadmap delivers. This is the structural reason the value line compounds at 54.86% against the cell line’s 67.42%, and it is not a modelling assumption that can be relaxed.

Risk six: the demand forecast has been wildly unstable and the numbers behind it disagree. Published 2030 forecasts for sodium-ion cell shipments span 435 GWh to 1,000 GWh — a factor of 2.3 — and the lower of the two was itself restored from a figure that had been cut by three quarters a year earlier. The forecast in this report reaches 500 GWh in 2033, below the lower of those 2030 numbers, deliberately. A market whose long-range outlook is revised by 75% down and then 300% up within eighteen months is not a market with a demand curve. It is a market with a capacity announcement, and the 13.23% realisation rate is what announcements are worth.

The substitution risk runs in both directions and the incumbent is not standing still. At 500 GWh, sodium-ion displaces about 1,275,500 tonnes of LFP cathode demand — consistent with the arithmetic in the LFP report that ten points of storage migrating to sodium removes roughly 1.4 million tonnes of LFP demand. But the incumbent can also move toward sodium, and the production of the leading LFP cathode maker can be retooled to polyanion on the same solid-state line in a short cycle. The most likely outcome is not substitution. It is that the same Chinese cathode industry runs both chemistries on the same assets, and that neither the sodium premium nor the LFP incumbency survives it.

Regulatory Landscape

The most consequential regulation in this market is not an industrial policy, a subsidy or an export control. It is a consumption tax asymmetry, and it has an expiry date inside the forecast window.

The instrument. From 1 September 2026 to the end of 2028, sodium-ion batteries are exempt from China’s battery consumption tax while lithium-ion batteries are restored to a 4% rate. On a lithium cell selling at RMB 0.55 a Wh, 4% is RMB 0.022 a Wh of tax that sodium does not pay. Set against the RMB 0.10 to 0.15 a Wh cost gap the cell makers themselves quote, the exemption is worth 14.7% to 22.0% of the entire difference between the two chemistries. It is the largest single policy support in the chain, it costs the state revenue rather than appropriating any, and it ends on 31 December 2028 — which is the fiscal cliff this report has to price.

The demand-side policy is a volume target, not a technology mandate. The August 2025 special action plan for scaled new-type energy storage covering 2025 to 2027 sets a national target above 180 gigawatts of new-type storage by 2027 and roughly RMB 250 billion of direct project investment, and names sodium-ion among the routes that must reach commercial deployment. That is a target for storage, not for a chemistry. Sodium-ion qualifies for it by being cheap enough, which returns the argument to the cost line.

The export-control position is the sharpest asymmetry in the report and it favours sodium-ion. China’s Announcement No. 28 of 2025, issued by the Ministry of Commerce and the Ministry of Science and Technology on 15 July 2025, added item 252604X — battery cathode material preparation technology — to the restricted-export catalogue. The thresholds are written for lithium iron phosphate at 2.58 g/cm3 compaction, 160 mAh/g and 97% first-cycle efficiency, for LMFP at 2.38 g/cm3, and for precursors above 1.2 g/cc tap density and below 10 parts per billion magnetic foreign matter. No threshold is published for polyanion sodium cathodes. On the face of the catalogue as published, the route carrying 70% of sodium-ion cathode tonnage sits outside the control. That is an inference from the thresholds rather than a statement of intent, and it matters: a state that restricts what it leads in has chosen not to restrict this, either because it does not yet consider the technology a chokepoint or because it wants the world to adopt it.

**The American instrument works against the licence, and it lands exactly where this industry’s cost advantage comes from.** The One Big Beautiful Bill Act of 2025 replaced the Inflation Reduction Act’s sourcing tests with a cost-weighted material assistance cost ratio under Section 70512, applied to the advanced manufacturing credit. The ratio counts technology licensing, equipment and processing relationships as assistance, and tightens through 2030. A sodium-ion cell line built on Chinese process licences and Chinese equipment is precisely what the ratio is designed to catch. The consequence is a two-tier market: a Chinese cost base that competes on price everywhere, and a Western capacity that must be qualified without the licences that make the price, which is why Peak Energy’s 20 GWh and Tiamat’s 5 GWh are strategically important and commercially fragile at the same time.

Europe regulates the artefact. Regulation 2023/1542 requires a battery passport from February 2027 and recycled content from 2031. Sodium-ion has no established recycling stream, no black-mass economics and no declared recovery route for hard carbon or for sodium salts, because nobody has yet retired a fleet. A documentation regime applied to a chemistry with no end-of-life data is a fixed cost that lands before the first revenue.

Three regimes, three different objects: China taxes the substitute’s competitor and lets the technology travel; America costs the licence; Europe documents the object. None of them raises supply, and only one of them is aimed at the price.