Global Sodium-Ion Battery Electrolyte & NaPF6 Market Size, Share & Forecast 2026–2033

Sodium hexafluorophosphate is the cheapest electrolyte salt in the world, and the electrolyte made from it is more expensive per kilowatt-hour than the lithium electrolyte it is supposed to replace. Both statements are arithmetically true in the same month.

NaPF6 quoted RMB 79,500 a tonne in China in January 2026. LiPF6 quoted RMB 168,500. That is 52.82 per cent cheaper by weight, and 61.87 per cent cheaper for a mole of charge carrier, because NaPF6 weighs 187.95 against LiPF6's 151.90 and both carry one monovalent ion. On the salt line the substitution is not marginal. It is decisive, and it is the only line on which sodium-ion's cost claim is unambiguously true.

Then the cell drinks it. A sodium-ion cell absorbs roughly 1,300 tonnes of electrolyte per gigawatt-hour. The lithium chain runs at about 1,000, and that is not an estimate — it falls out of CATL's own 2026 procurement contracts, 770,000 tonnes of electrolyte against 770 gigawatt-hours of cells. Multiply through and sodium electrolyte at RMB 28,000 a tonne costs RMB 0.0364 per watt-hour, against RMB 0.03575 for LFP electrolyte at RMB 35,750. The cheap chemistry is 1.82 per cent more expensive on the line it was built to win. The dosage has cancelled the salt discount almost exactly.

Companion analysis. The dosage arithmetic in this report is written up for a non-specialist reader in The Cheaper Salt That Costs More, which shows why a salt 61.87% cheaper per mole of charge carrier produces an electrolyte 1.82% dearer per kilowatt-hour.

Market forecast snapshot 2026 – 2033

Key takeaways

• The value line grows at 45.40% and the volume line at 60.84%, and the 15.4-point gap is the product. Every tonne added in 2033 sells for 55.4 per cent less than the tonne added in 2025. A market whose proposition is that it gets cheaper cannot also compound its value at its tonnage rate.
• The salt's deflation stopped in 2025 and reversed. Sodium-ion electrolyte fell from RMB 180,000 a tonne in 2021 to RMB 28,000 in the first half of 2025, an 84.4 per cent collapse. SMM then expected NaPF6 to approach RMB 50,000–60,000 in the second half of 2025. It quoted RMB 79,500 in January 2026 — above the top of that range — because LiPF6 went from a RMB 49,000 low in July 2025 to about RMB 180,000 in November. The sodium salt is priced off the market it is trying to displace.
• The value accrued to the formulator, not the cell maker. On a January 2026 cost stack a tonne of sodium electrolyte carries RMB 18,450 of raw material — 65.89 per cent — and leaves a RMB 9,550 gross spread, 34.11 per cent. When the salt fell to RMB 70,000 by September 2026 that spread widened; the cell maker's cost did not move.
• Capacity is not the barrier, and that is the barrier. Entry costs about RMB 1,084 per tonne of nameplate — Tinci's Jiujiang approval is RMB 542 million for 500,000 tonnes. There are no refractory linings, no furnaces, no mines. The barrier is a formulation file that differs by more than 30 per cent between customers.
• The most-quoted cost structure in this chemistry is stale. The published split puts cathode and electrolyte at 26 per cent each of cell material cost. At 2025 prices the cathode reproduces at 25.75 per cent. The electrolyte reproduces at 13.66 per cent. The parity was real when electrolyte sold at six figures a tonne; it is not real now.
• The rebate ends on 1 January 2027. Announcement No. 2 of 2026 cut China's VAT export rebate on battery products from 9% to 6% from April 2026 and abolishes it from January 2027. That is a step change in realised export value inside the forecast window, and it is why 2026 volume growth front-loads.

Overview

This forecast covers the global market for formulated sodium-ion battery electrolyte sold as product: the sodium salt, whether NaPF6 alone or a blended NaPF6 / NaFSI system; the carbonate solvents — ethylene carbonate, propylene carbonate, dimethyl and ethyl methyl carbonate; the functional additives; and the blending, quality control and hazardous logistics that turn those inputs into a specification a cell maker will accept. It excludes merchant solvent sold to other industries, merchant salt sold outside electrolyte, lithium-ion electrolyte, solid-state electrolytes, and cell and pack manufacture. The contained NaPF6 pool is published as a memo line on the same basis and is never added to the electrolyte value, because the salt is inside the electrolyte.

The volume denominator is a construction and is labelled as one. 8.16 GWh of 2025 sodium-ion cell output is derived from 20,000 tonnes of reported cathode shipments at 2,450 tonnes of cathode per gigawatt-hour, a ratio set by specific capacity and voltage rather than by survey. That GWh figure is then corroborated from the electrolyte side: reported shipments of 5,060 tonnes in 2024 against 3.6 GWh of cells, and 4,818 tonnes in the first half of 2025 against 3.7 GWh, imply dosages of 1,406 and 1,302 tonnes per GWh respectively. Two materials, two reporting channels, one answer.

Value is that volume at a realised price. Sodium electrolyte sold at RMB 28,019 a tonne in the first half of 2025 on a shipment value of RMB 140 million for 4,818 tonnes; 10,608 tonnes at that price is RMB 297.2 million, or USD 42.58 million at RMB 6.98 to the dollar — the rate implied by the sodium hexafluorophosphate assessment that anchors the cost work. Growth is then deterministic: one volume path, one price path, and three scenarios that differ only in the rate of the volume path.

What this report argues

First, that the electrolyte is the wrong place to look for sodium-ion's cost advantage, and the right place to look for its cost problem. The salt is genuinely cheap. The dosage is not. Until the cell stores more energy per kilogram, the electrolyte line will not beat LFP's on a per-kilowatt-hour basis no matter what happens to NaPF6, because the advantage in price and the disadvantage in volume move through the same multiplication.

Second, that the deflation everybody is underwriting has already been interrupted once, by a market sodium is not in. NaPF6, PCl5 and anhydrous hydrogen fluoride are shared with the lithium salt chain. When LiPF6 tripled in the second half of 2025, the fluorine and the phosphorus went where the price was. A sodium salt producer is a residual claimant on a shared fluorine balance, exactly as a thermal phosphoric acid furnace is a residual claimant on cheap electricity — and the report treats both as the same structural phenomenon.

Third, that concentration in this market is a formulation fact, not a capital fact. Three suppliers hold more than 60 per cent of shipments, and roughly 90 per cent of the capacity they run is retrofitted lithium-ion plant. The moat is a specification that varies by more than 30 per cent between cell makers, which is why a 10,000-tonne world market can carry a 34 per cent gross spread and why a 50,000-tonne dedicated line being built now is 4.72 times the entire 2025 market.

Fourth, that the honest risk is not sodium losing to lithium — it is sodium winning on volume and losing on value. The report models exactly that: 60.84 per cent volume growth against 45.40 per cent value growth, with revenue per tonne falling 9.60 per cent a year for eight consecutive years, and it says in the risk chapter that a faster-than-expected energy-density gain makes the volume line smaller, not larger.

Companion analysis and related research

The sodium-ion material stack is covered from the other two directions in this estate. Global Sodium-Ion Battery Materials sizes the cathode and the hard carbon, establishes the 2,450 tonnes-per-GWh cathode constant and the 3,579 tonnes-per-GWh total active-material constant used here, and deliberately excluded this line so that the two markets do not double count. Global LFP Cathode Active Material sizes the chemistry sodium has to undercut and carries the compaction-density and fee-ladder evidence that sets the competitive bar. Global Battery Electrolyte Production covers the lithium-ion electrolyte franchise this market is carved out of, including the solvent and LiFSI detail that this report does not repeat. Global Stationary Energy Storage Systems covers the end market that takes more than half of sodium-ion tonnage and is therefore the demand engine underneath the 2033 volume path.

Read together, the four establish a perimeter in which no tonne of material is counted twice and no molecule is counted once as two markets.

Table des matières

1. Value chain — the salt, the solvent, the additive, and where RMB 9,550 a tonne goes
2. Technology and innovation — dosage is the whole game
3. Competitive landscape — a 60% concentration sitting on 90% borrowed capacity
4. Regulatory landscape — the rebate that ends on 1 January 2027
5. Investment activity — RMB 1,084 a tonne of nameplate
6. Market risk — Natron, the price of a cell, and a fourfold disagreement about 2030

Paysage concurrentiel

This market is highly concentrated and the concentration is not a moat. Both halves of that sentence need to be held at once.

Three suppliers held more than 60 per cent of 2024 sodium-ion electrolyte shipments. Zhuhai Saiwei led with more than 30 per cent, Jiangsu Langu was second, and BYD held 12.9 per cent almost entirely captive to its own cells. Below them the field is thin and fragmentary: most second-tier producers ship at tonne scale from pilot and trial lines rather than from commercial plants, and several buy salt externally rather than making it. A 60-plus per cent three-firm concentration would, in most chemicals markets, describe a group with pricing power. Here it describes a group that is large relative to a market of 10,608 tonnes and small relative to the plants its competitors are about to build.

Roughly 90 per cent of sodium electrolyte capacity is retrofitted lithium-ion capacity. This single fact reorganises the entire competitive analysis. Sodium electrolyte lines are, to a first approximation, interchangeable with lithium lines — the same dissolution, blending, filtration and filling equipment, the same anhydrous handling, the same hazardous logistics. Which means supply is elastic and entry is nearly free: any of the twenty-odd Chinese electrolyte formulators can make sodium electrolyte if a customer asks, and several already do. The scarcity is not plant. It is the approved formulation and the cell platform it is designed into.

The listed entrants are arriving with dedicated capacity and they will change the concentration number in the opposite direction to the one most readers expect. Tinci received regulatory approval in March 2026 for a 500,000 tonne a year electrolyte project at Jiujiang Hukou — RMB 542 million, 450,000 tonnes lithium and 50,000 tonnes sodium, construction March to December 2026 — and disclosed in July 2026 that it is already in batch supply to CATL and other leading cell makers, offering an integrated package of NaPF6, NaFSI and formulation. Zhejiang NaChuang, Zhongxin Fluorine and Hongda Chemical signed a 50,000 tonne a year sodium electrolyte project in 2025. Yongtai Technology is converting LiPF6 plant, adding 30,000 tonnes of capability and forming a 50,000 tonne line. Two announced projects alone come to 100,000 tonnes — 9.43 times the entire 2025 world market, and 87.57 per cent of the volume this report forecasts for 2030. Concentration will fall as the market grows, because the entrants are larger than the incumbents.

The incumbent’s defence is design-in, and design-in is slow and sticky in both directions. A cell maker that has qualified a sodium electrolyte has qualified a formation protocol, a gas-generation profile, a cycle-life curve and a safety file against it. Switching costs a re-qualification cycle measured in quarters. That protects Zhuhai Saiwei and Jiangsu Langu for as long as their customers’ platforms live, and it is also why a single design-out can remove a third of a small market’s volume in one decision. The competitive question in this market is not who can make the cheapest tonne. It is who is written into the next NFPP platform, and there are only a handful of those.

The lithium franchise is the counter-example, and it explains why the giants are cautious. In June 2026 CATL signed long-term electrolyte agreements totalling 770,000 tonnes — 470,000 with Yongtai and 300,000 with Capchem on a tiered 50,000 / 100,000 / 150,000 schedule through 2028 — after Tinci’s previous agreement expired on 31 December 2025 without renewal. That is a market where scale buys volume and price is the currency. Sodium is the opposite: the volumes are four orders of magnitude smaller, the price is not the deciding variable, and a firm that wins lithium on cost has no automatic claim on sodium. Tinci’s own investor disclosure carried the warning in the same paragraph as the CATL supply news — the sodium electrolyte market is at an early industrialisation stage, demand has not broadly broken out, and the outlook is uncertain.

The salt makers occupy a different position in the same competition. Do-Fluoride’s 2025 annual report is the most complete primary disclosure of the sodium salt position available: 65,000 tonnes a year of LiPF6 capacity at 70–80 per cent utilisation, 80,000 tonnes of electronic-grade hydrogen fluoride at 30–40 per cent, 20 GWh of cell capacity at 50–60 per cent, and — in the company’s own words — kilotonne-scale lines for LiFSI and NaPF6, with NaPF6 already in commercial mass production and customers among mainstream sodium cell and electrolyte makers. Set the two side by side and the ratio is roughly 65 to 1. The sodium salt is not a business inside Do-Fluoride; it is an option on the fluorine balance, exercised when sodium demand justifies the switch, and the company says so without embarrassment.

What share is actually competed for. Not tonnage. Three things: first, the formulation slot on a qualified cell platform; second, integrated salt supply, because a formulator without its own salt has no defence against a market-following quote on its largest single cost line; third, the tolerance to absorb formulation churn — a supplier that scraps a batch every time a customer changes an additive cannot serve a market where packages differ by 30 per cent. On those three criteria the incumbents are strong and the entrants are unproven, and the entrant with the best claim is the one that already owns the lithium franchise and the salt above it.

What is absent. There is no announced Western sodium electrolyte capacity of consequence. The most visible Western sodium-ion manufacturer, Natron Energy, closed both of its plants in September 2025. Korea’s and Japan’s electrolyte houses have the capability and have not committed the capital. The competitive landscape through 2033 is a Chinese one, and the export-control question that governs the cathode — a published compaction-density threshold that sodium polyanion chemistry sits outside — has no equivalent on the electrolyte side at all.

Chaîne de valeur

A tonne of sodium-ion electrolyte is three inputs and one process, and the ratio between what each input weighs and what each input costs is the first thing that anyone entering this business gets wrong.

The stack. Take a representative tonne at January 2026 prices. It carries about 140 kilograms of sodium hexafluorophosphate and, in most current formulations, something under 15 kilograms of sodium bis(fluorosulfonyl)imide, the additive the industry calls NaFSI. The remaining 845 kilograms is a carbonate solvent blend — ethylene carbonate for the dielectric constant, propylene carbonate or ethyl methyl carbonate for the low-temperature viscosity — plus small amounts of film-forming additives such as fluoroethylene carbonate. In value terms the same tonne is RMB 11,130 of NaPF6, RMB 2,250 of NaFSI and roughly RMB 5,070 of solvent: RMB 18,450 of raw material, 65.89 per cent of a RMB 28,000 selling price, leaving a gross spread of RMB 9,550, or 34.11 per cent, before conversion energy, packaging, hazardous logistics and selling cost.

The inversion inside that stack. NaFSI is 8.04 per cent of the cost from 1.5 per cent of the mass, because it trades near RMB 150,000 a tonne — roughly twice what sodium hexafluorophosphate commands — while the solvent that is 84.5 per cent of the mass contributes only 18 per cent of the price. An electrolyte formulator is therefore not a blender who buys the cheap thing in bulk. He is a buyer of two very expensive minority ingredients and one cheap majority one, and the margin lives in how precisely he can dose the expensive two. This is why the industry describes its cost problem as an additive-constrained profit problem rather than a solvent problem, and why a 20 per cent move in NaFSI costs more than a 20 per cent move in the solvent blend.

The salt has no forward curve. SMM’s own description of the sodium hexafluorophosphate market in September 2026 is the most useful sentence written about this supply chain: public quotes of about RMB 70,000 a tonne, no unified pricing mechanism, strongly market-following, demand arriving in pulses because cell makers buy only after they have taken an order, no stable monthly volumes and no clear seasonal pattern, few stable mass-production enterprises, limited output from converted older lines, and some small producers buying the salt externally rather than making it. The industry produces to order and holds no stock. The consequence is that nothing in this chain can be hedged. A formulator quoting a cell maker a six-month price is taking a directional position on fluorine, on phosphorus and on the lithium salt market simultaneously, and the working capital tied up in inventory is a position rather than a buffer.

The shared fluorine balance, which is the real supply story. Do-Fluoride’s 2025 annual report sets out the sodium route without embellishment: NaPF6 is made from phosphorus pentachloride, anhydrous hydrogen fluoride and sodium fluoride, reacted in an anhydrous organic solvent and then purified, crystallised and dried. The lithium salt is made from the same PCl5 and the same anhydrous HF against a lithium source. These are not adjacent businesses sharing a supplier; they are the same fluorine and phosphorus balance sheet with a different cation bolted on. So when LiPF6 went from a RMB 49,000 low in July 2025 to roughly RMB 180,000 in November, the marginal tonne of HF and PCl5 went where the price was, and the sodium salt followed it up rather than continuing the deflation the sodium industry had budgeted. NaPF6 was expected by the market to approach RMB 50,000–60,000 in the second half of 2025. It quoted RMB 79,500 in January 2026 — above the top of that range — and only eased to about RMB 70,000 by September 2026.

That is the defining structural feature of this value chain: the input cost of the electrolyte that is supposed to displace lithium is set by the lithium market. A sodium salt line is a residual claimant on a shared fluorine balance, and residuals clear at the margin price, not at their own. The thermal phosphoric acid furnace in this estate’s phosphonate work is the same animal — a high-fixed-cost asset that runs when the price of what it shares is cheap and idles when it is not.

And the fluorine itself is a sulphur derivative. Hydrogen fluoride is made by reacting fluorspar with sulphuric acid, so the cost of the anion that makes both NaPF6 and LiPF6 what they are is exposed to a sulphur market that has nothing to do with batteries. This estate’s lithium electrolyte work records a sulphur feedstock disruption running from February 2026 off a shipping chokepoint, with China protecting domestic supply and lithium salt producers exposed to sulphuric acid costs. A sodium salt line does not diversify that exposure — it inherits it, one step further down a chain the sodium industry believes it is leaving behind.

Integration is therefore the only durable cost position. Tinci’s July 2026 disclosure is the clearest statement of strategy in the market: it offers customers an integrated formula solution covering NaPF6, NaFSI and the formulation, and says it will keep building sodium electrolyte capacity and upstream raw material capacity. A firm that makes its own salt internalises the spread between the salt and the electrolyte; a firm that buys it at a market-following quote with no forward curve absorbs the volatility and passes on what it can. Do-Fluoride sits on the other side of the same logic — anhydrous HF, PCl5-derived salts, 65,000 tonnes of LiPF6 capacity and a kilotonne-scale NaPF6 line that is already commercial — and its advantage is not the sodium line, which is small, but the fluorine above it.

What the chain is not. It is not capital-hungry. Tinci’s Jiujiang approval is RMB 542 million for 500,000 tonnes a year of electrolyte — RMB 1,084 per tonne of nameplate — split 450,000 tonnes lithium and 50,000 tonnes sodium, with construction scheduled from March to December 2026. Compare that with the cathode chapter of this estate’s LFP work, where a qualified line is a multi-year, kiln-by-kiln qualification with a cost per tonne of qualified output many multiples higher. Here the capital is almost free and the qualification is almost everything: a blending line with no approved formulation is worth its scrap value.

Where the value actually settles. Between the salt’s absence of a forward curve and the plant’s trivial capital cost, the residual is held by whoever owns the specification a cell maker has signed off. That is why the gross spread has held at roughly a third of the selling price through a period in which the salt’s price moved by tens of thousands of yuan, and why the deflation that reaches the customer is a deliberate transfer rather than a competitive necessity. The chain has one genuinely scarce asset, and it is a document.

Activité d'investissement

The capital requirement in this market is so small that it is not a barrier, and the investment question is therefore not how much but how defendable.

Start with the unit number. Tinci’s Jiujiang Hukou approval is RMB 542 million for 500,000 tonnes a year — 450,000 tonnes of lithium electrolyte and 50,000 of sodium — which is RMB 1,084 per tonne of nameplate capacity, with construction scheduled from March to December 2026. For context, the same estate’s work on LFP cathode active material describes a business where capital per tonne of qualified output is the strategic variable and where a 2.60 g/cm³ line requires a second calcination and roughly double the kiln time. Here a project with four times the 2025 world market in sodium capacity costs less than a mid-sized specialty chemical plant. Capital is not the gate.

What has been announced is already more than the market. Tinci: 50,000 tonnes a year of sodium electrolyte, approved March 2026. Zhejiang NaChuang with Zhongxin Fluorine and Hongda Chemical: a 50,000 tonne a year sodium electrolyte project signed in 2025. Yongtai Technology: a LiPF6 technical upgrade and expansion, adding 30,000 tonnes of capability into a 50,000 tonne line. Two sodium projects alone total 100,000 tonnes — 9.43 times the entire 2025 market and 87.57 per cent of the volume this report forecasts for 2030. That is not a capacity gap being filled; it is a capacity overshoot being scheduled, and the investment question becomes which of those tonnes will be qualified rather than which will be built.

The incumbent salt position is the other half of the investment picture, and its numbers are instructive about where the money actually is. Do-Fluoride’s 2025 annual report: revenue RMB 9.434 billion, with new-energy materials at RMB 3.134 billion, up 20.86 per cent, fluorine-based new materials at RMB 2.789 billion and electronic information materials at RMB 409 million, down 33.04 per cent, with net profit turning positive after a loss. Capacity and utilisation: LiPF6 65,000 tonnes at 70–80 per cent, electronic-grade hydrogen fluoride 80,000 tonnes at 30–40 per cent, cells 20 GWh at 50–60 per cent, 30,000 tonnes of LiPF6 and 10 GWh of cells under construction. The NaPF6 line is described as kilotonne-scale and already commercial. Read the utilisation column twice: in a chain where the plant is cheap, utilisation is the entire P&L, and a company running its lithium salt at three-quarters and its electronic-grade acid at a third does not allocate scarce fluorine to a kilotonne sodium line because the sodium story is good. It allocates it when the sodium price clears.

What to underwrite, and what not to. Three things are worth capital. Salt capacity with hydrogen fluoride integration, because the salt is the largest single cost line at 39.75 per cent of the electrolyte price and the only durable defence against a market-following quote with no forward curve is to own the fluorine above it. A formulation file tied to a named cell platform, because that is the asset that produces the 34.11 per cent gross spread and it cannot be bought at RMB 1,084 a tonne. And working capital sized for a market with no hedge, because inventory in this chain is a directional position on PCl5, HF and LiPF6 rather than a buffer against delivery risk.

Two things are not worth capital. Standalone blending capacity, for the arithmetic reason above: at RMB 1,084 a tonne of nameplate, a line with no approved formulation is worth its scrap value and its land. And second-tier tonne-scale pilot lines, which are how most of the field currently ships and which cannot survive a customer changing an additive package by 30 per cent.

The absence of M&A is itself a finding. In the phosphonate market covered elsewhere in this estate, PCBL paid INR 3,800 crore for Aquapharm at 1.86 times sales and 10.3 times EBITDA, and Dorf Ketal entered exclusive talks for Italmatch at a reported USD 1.6 billion, because what was for sale was an approved-vendor position and a formulation file. This market has the same asset and no transactions. It is too small to buy: a 30 per cent share of a USD 42.58 million market is USD 12.8 million of revenue, which does not clear anyone’s threshold, and the asset that would justify a premium — the formulation file — is attached to specific customers who would re-open the qualification if ownership changed. Expect consolidation to arrive with volume, not before it.

The counter-signal an investor should hold against all of this. In June 2026 CATL placed 770,000 tonnes of lithium electrolyte under long-term contract — 470,000 with Yongtai, 300,000 with Capchem — after Tinci’s agreement lapsed unrenewed at the end of 2025. That is the market that funds the plants, the R&D and the salt integration. Sodium electrolyte capacity is being built by firms whose lithium business pays for it, which means sodium investment decisions are made on strategic option value rather than on return on capital, and it means the capacity will be there whether or not the demand is — the classic condition for a price war in a product whose price is already the selling proposition. An investor underwriting this market should underwrite the spread, not the tonnage, and should assume the spread compresses the moment the 50,000-tonne lines qualify.

Technologie et innovation

Dosage is the whole game, and everything else in this chemistry is a detail. Sodium-ion’s cost case is made on the salt and lost on the kilogram count, and the kilogram count is a property of the cell rather than of the electrolyte.

The arithmetic is worth restating because it is the report. A sodium-ion cell absorbs roughly 1,300 tonnes of electrolyte per gigawatt-hour. The lithium chain runs at about 1,000, a figure that is not an estimate but a division that falls out of CATL’s 2026 procurement contracts — 770,000 tonnes of electrolyte against 770 gigawatt-hours of cells across the Yongtai and Capchem agreements. So on a per-kilowatt-hour basis the sodium electrolyte’s 52.82 per cent price advantage against the lithium salt is consumed almost exactly by a 30 per cent dosage penalty. At RMB 28,000 a tonne against LFP electrolyte at RMB 35,750, sodium costs RMB 0.0364 per watt-hour against RMB 0.03575 — 1.82 per cent more.

Why sodium needs more. Three reasons, and only one of them is fixable in the near term. First, energy density: a sodium cell stores fewer watt-hours per kilogram than an LFP cell, so for a similar electrolyte mass fraction the mass required per kilowatt-hour rises roughly in proportion to the density gap. Second, the anode interface: hard carbon forms a solid electrolyte interphase that consumes both sodium and electrolyte in the first cycles, so cells are built with excess electrolyte to survive the formation step — a direct, permanent tonnage penalty that a graphite anode does not impose to the same degree. Third, the salt is heavier: NaPF6 at 187.95 grams per mole against LiPF6 at 151.90 means that for the same molar concentration the mass loading is 23.7 per cent higher before any other effect is counted.

Dosage is also falling, and that is measurable rather than assumed. Reported electrolyte shipments of 5,060 tonnes in 2024 against 3.6 GWh of cells imply 1,406 tonnes per GWh. Reported shipments of 4,818 tonnes in the first half of 2025 against 3.7 GWh imply 1,302. A 7.4 per cent improvement in twelve months, from cell design and formation protocols rather than from any electrolyte innovation. The forecast therefore runs dosage down from 1,300 tonnes per GWh in 2025 to 950 in 2033, and the reader should understand that this is simultaneously a tailwind and a headwind: better cells make sodium cheaper per kilowatt-hour, which grows the market, and they make each gigawatt-hour buy less electrolyte, which shrinks the tonnage line the market is counted on.

The sensitivity is asymmetric and it points the wrong way for the salt. A 10 per cent cut in dosage to 1,170 tonnes per GWh takes the sodium electrolyte line to RMB 0.03276 per watt-hour — 8.36 per cent below LFP’s. Sodium wins the cost comparison not by making a cheaper salt but by needing less of it per unit of energy. Conversely, no plausible further fall in NaPF6 closes the gap while dosage stands at 1,300: a salt at zero would still leave the electrolyte at RMB 0.0234 per watt-hour once solvents and additives are counted, which is a 35 per cent saving and the absolute ceiling of what the salt can ever deliver.

The salt choice is a system decision, not a purchasing decision. NaPF6 is the incumbent solute because it is the direct analogue of LiPF6 — same hexafluorophosphate anion, similar conductivity, similar passivation behaviour, made on plant that already exists. It is also hydrolysable, generating hydrogen fluoride on contact with trace water, which is why cells are built to a water specification measured in parts per million and why the salt is handled under anhydrous conditions from reactor to drum. NaFSI at around RMB 150,000 a tonne is better on almost every electrochemical axis — more stable, more conductive, less prone to hydrolysis — and is used at one to two per cent because it attacks the aluminium current collector at high voltage. That constraint matters more here than it does in lithium: sodium does not alloy with aluminium, so a sodium cell can use aluminium foil on both electrodes and delete the copper. The cell’s cost advantage is partly a current-collector advantage, and the additive package has to protect it.

Formulation variance is the technology barrier. Sodium salt concentration and additive packages differ by more than 30 per cent between cell makers, and the variance runs across two cathode routes with different voltage windows: polyanion NFPP, which took 70 per cent of 2025 sodium cathode tonnage, and layered oxide, which runs at a higher average voltage and therefore a different anodic stability requirement. Two routes times three or four major cell designs times customer-specific formation protocols is a specification matrix rather than a product catalogue. There is no standard tonne of sodium electrolyte, which is precisely why a market of 10,608 tonnes can support seven or eight serious suppliers and a third of the selling price in gross spread.

What would actually move the arithmetic. Not a cheaper salt. Either a cell with materially higher energy density — which is a cathode and anode programme, not an electrolyte one — or a salt route that breaks the shared fluorine cost, which would require a different anion chemistry and a new qualification cycle. The industry’s announced innovation effort is concentrated in the first: NFPP and layered-oxide work, pre-sodiation to recover first-cycle losses, and hard-carbon porosity control. All of it reduces electrolyte tonnage per kilowatt-hour. None of it raises the electrolyte’s share of value. That is the technology chapter’s uncomfortable conclusion: every genuine advance in sodium-ion cell technology makes the electrolyte market smaller per gigawatt-hour and larger only if the gigawatt-hours arrive faster than the dosage falls — which the forecast assumes they do, by a factor of about three.

Risque de marché

The first risk is that the cost claim is a cell-level claim and the electrolyte cannot deliver it alone. At RMB 28,000 a tonne and 1,300 tonnes per gigawatt-hour, sodium electrolyte costs 1.82 per cent more per watt-hour than LFP electrolyte. That is not a stress case. It is the base case, at the prices the market actually printed in January 2026, and it persists for as long as the dosage penalty persists. The only lever that closes it is cell energy density, which no electrolyte supplier controls. An investor who buys the sodium electrolyte thesis is buying a bet on somebody else’s cathode and anode programme.

The second risk is that the most credible Western sodium-ion manufacturer has already failed. Natron Energy filed a WARN notice dated 28 August 2025 permanently closing its facilities in Holland, Michigan and Santa Clara, California, effective 3 September 2025, with 95 employees laid off across both states, 37 of them in Michigan. The notice is unusually candid about cause: the board concluded on 27 August 2025 that a follow-on equity round, a Series B process, a secured convertible note offering and a management-led buyout had all failed to raise enough working capital to execute purchase orders it had already received. Natron was a Prussian-blue sodium-ion manufacturer — the one route that sits outside the Chinese cathode specification, with a chemistry that uses no hard carbon and no NaPF6 in the same sense — and it stopped anyway, for capital reasons and not for technical ones. No Western sodium electrolyte demand of scale exists today, and the company that would have anchored it is gone.

The third risk is that the product is still more expensive than what it replaces, at the product level. Sodium-ion cells averaged RMB 0.41 per watt-hour in the first half of 2025, down from RMB 0.90 in 2021. LFP cells sit at RMB 0.35–0.40. The sodium cell is therefore 10 to 20 per cent above the cell it must beat, in the same quarter in which the material stack it is built from is supposed to be structurally cheaper. The chain has removed 54 per cent of the cell’s cost in four years and has not yet reached parity, and the electrolyte’s contribution to closing the remaining gap is, on the arithmetic above, negative.

The fourth risk is a fourfold disagreement about 2030. Published industry forecasts put global sodium-ion demand at 435 GWh by 2030, with electrolyte shipments above 500,000 tonnes. The Pheonix path — anchored to 20,000 tonnes of 2025 cathode shipments and 2,450 tonnes of cathode per gigawatt-hour — puts 2030 at roughly 114,000 tonnes of electrolyte and about 104 GWh of cells. The gap is a factor of about four and it sits entirely in the next four years; the endpoints converge later. This report does not dismiss the higher figure. It argues that the higher figure describes announced capacity and stated ambition while this one describes materials that shipped, and in a market where two 50,000-tonne projects are already scheduled against a 10,608-tonne base, the distinction between capacity and shipment is the whole difference between the two forecasts. If the consensus is right, this report is materially too low and every number in it is too low. If this report is right, the announced electrolyte capacity will run at a utilisation that makes the 34.11 per cent gross spread unsustainable.

The fifth risk is concentration, and it cuts in both directions. Three suppliers holding more than 60 per cent of shipments in a market of this size is not a moat; it is a small number of qualifications. A single design-out removes a third of the market’s volume in one customer decision, and there is no replacement platform behind it because the number of NFPP cell designs in commercial production is measured in single figures. The same concentration means the incumbents earn a spread that has no structural defence once the entrants qualify.

The sixth risk is that the salt’s price is not sodium’s to set. NaPF6 quoted RMB 79,500 in January 2026 against a market expectation of RMB 50,000–60,000, because LiPF6 went to about RMB 180,000 in November 2025 and took the shared fluorine and PCl5 with it. A second lithium squeeze would do the same again. On the evidence of the last cycle the formulator keeps the spread rather than absorbing the move, which means the increase lands on the cell maker — and the cell maker, already 10 to 20 per cent above LFP, passes it to a storage tender that will not pay it.

The seventh risk is that the demand is tender-priced. More than 60 per cent of sodium-ion tonnage goes to stationary storage, a market bought through competitive tenders by utilities and state-owned developers with published price benchmarks. A 34.11 per cent gross spread earned from a customer base that runs reverse auctions is a spread with a date on it.

The eighth risk is inside the forecast’s own mechanics. Dosage falls from 1,300 tonnes per GWh in 2025 to 950 in 2033. That is a 3.9 per cent annual improvement and it is generous, because the observed improvement from 2024 to the first half of 2025 was 7.4 per cent. If dosage reaches 900 instead, 2033 volume is 450,000 tonnes rather than 475,000 and value is about USD 806 million rather than USD 850.6 million. Every real advance in sodium cell technology makes this market smaller per gigawatt-hour. The forecast survives only because gigawatt-hours are assumed to arrive faster than the dosage falls, by a factor of roughly three.

Numbers this report refuses. It publishes no third-party sodium electrolyte market size; syndicated estimates surfaced during research and none is cited, used as a cross-check or reconciled to. It publishes no Chinese NaPF6 nameplate total, because the only primary disclosure is Do-Fluoride’s own phrase kilotonne-scale and circulating figures an order of magnitude larger are contradicted by the filing. It publishes no capacity split by producer. It sizes no NaFSI volume — the additive is priced at about RMB 150,000 a tonne and its loading varies by customer, so it is a cost line rather than a market in this perimeter. It states no per-customer formulation margin. And it does not quantify solid-state substitution: sodium-ion’s own energy-density trajectory is the nearer threat, and a solid-state cell that arrives at commercial cost would take the storage segment this market depends on without leaving a number to put in a model.

Paysage réglementaire

Two instruments matter in this market and neither of them was written about electrolytes. One is a tax notice. The other is a chemistry specification that stops short of the molecule this report sizes.

The rebate that ends on 1 January 2027. On 8 January 2026 China’s Ministry of Finance and State Taxation Administration issued Announcement No. 2 of 2026. It has three operative clauses. From 1 April 2026 the VAT export rebate on photovoltaic products is cancelled outright. From 1 April 2026 to 31 December 2026 the VAT export rebate on battery products is cut from 9 per cent to 6 per cent. From 1 January 2027 the battery rebate is abolished entirely. 271 product codes are affected, and the consumption-tax rebate rules for the same products are explicitly left unchanged. The rate that applies is the one on the customs declaration’s export date, not the contract date — which is the detail that decides behaviour, because it makes the cliff hard rather than soft.

For this market the notice operates through three channels, and the second is the one nobody has priced. First, it raises the realised cost of an exported cell by three percentage points of VAT-able value in 2026 and nine from 2027. A sodium-ion cell selling at RMB 0.41 per watt-hour against an LFP cell at RMB 0.35–0.40 has no margin to absorb nine points. The rebate had been functioning as a price discount to foreign buyers, which is exactly what the China Photovoltaic Industry Association said out loud when the announcement landed; removing it means either a price increase in export markets or a margin transfer, and neither is available to a product that has not yet achieved cost parity. Second, it pulls volume forward. Anything that ships before 1 April 2026 carries 9 per cent; anything shipping in the following nine months carries 6; anything from January 2027 carries nothing. A rational exporter ships early and a rational buyer accepts early delivery, which is why 2026 volume growth in this model is the strongest single year in the forecast — and why 2027 growth decelerates. Third, it discriminates between exporting a chemical and exporting a product. An electrolyte exported as a formulated chemical and a cell exported as a battery both sit in the battery product schedule, but the value added, the margin and the compliance burden are completely different. The notice gives no relief for the material and none for the cell; it simply makes the whole export chain more expensive at the same moment.

The control that stops short of the molecule. China’s MOFCOM and MOST Announcement No. 28 of 2025 (15 July 2025) added battery cathode material preparation technology to the restricted-export catalogue as item 252604X, written as a specification rather than a company list: LFP at ≥2.58 g/cm³ compaction, ≥160 mAh/g, ≥97 per cent first-cycle efficiency, LMFP at 2.38 / 155 / 3.85 V, and a precursor limb covering iron phosphate above 1.2 g/cc tap density and below 10 parts per billion magnetic foreign matter. No threshold was published for polyanion sodium cathodes and none at all for electrolyte formulations or sodium salts. That is an inference from the text, not a statement about intent, and the report makes it as an inference: the technology this instrument fences in is the cathode, and the chemistry carrying 70 per cent of 2025 sodium cathode tonnage sits outside the fence as written. On the electrolyte side there is no equivalent instrument in any jurisdiction this report can find.

The commercial consequence is worth stating plainly. If the electrolyte is not export-controlled, then the diffusion of sodium electrolyte capability is limited by customers, not by law — and the customer base is Chinese. That makes the legal environment permissive and the market environment restrictive at the same time, which is the opposite of the pattern in the cathode.

The real regulatory constraint is environmental, and it sits upstream. New hydrogen fluoride and fluorochemical capacity in China runs through a permitting process with lead times reported at around eighteen months, and the 2026 electrolyte price commentary attributes part of the solvent and additive tightness — ethylene carbonate moving from RMB 4,300 to RMB 6,200 a tonne, VC additive above RMB 230,000 — to exactly that constraint. For a sodium electrolyte producer, therefore, the binding permit is not on the blending line, which needs almost nothing, but on the fluorine above it. A market whose capital cost is RMB 1,084 a tonne is nevertheless a market whose input supply is gated by an environmental approval cycle two years long.

Two demand-side instruments do the work that material regulation does not. Sodium-ion cells pay no consumption tax from 1 September 2026 to the end of 2028 while lithium cells return to a 4 per cent rate — a cell-level advantage worth RMB 0.022 per watt-hour, or 14.7 to 22.0 per cent of the entire cost gap between the two chemistries, and one that expires inside the forecast window with 84.4 per cent of the 2033 tonnage arriving after it does. And on the other side of the Pacific the OBBBA’s replacement of the IRA sourcing test with a cost-weighted material assistance cost ratio counts technology licensing as assistance and tightens through 2030 — which is a live question for any Western sodium cell built on licensed Chinese process. Neither instrument names the electrolyte. Both change how much of it is bought.

Standards as an instrument, and as a moat. Do-Fluoride reports having led or participated in the revision of more than 130 national and industry standards, including two international standards, with electrolyte salt standards among them. In a market where the product is a customer-specific specification rather than a commodity, the ability to write the standard is a commercial position that no export-control list can replicate. It is also the mechanism by which the incumbent formulation houses keep the qualification barrier high while the plant cost stays near zero.