Pheonix Research - Editorial Index
Editorial analysis anchored to industries, segments, and the market intelligence behind them.
Liquid chlorine in Jiangsu averaged minus RMB 179 a tonne in Q2 2025. Strip that credit out and a tonne of PCl₃ costs more in yellow phosphorus than it sells for — a 3.5 per cent margin that a 10 per cent move in the element wipes out 2.7 times over.
A tonne of thermal phosphoric acid carries 4,109 kWh and RMB 7,688 of phosphorus, and sells for about RMB 7,940. That is why 1.79 million tonnes of Chinese furnace capacity sat idle in 2025 — and why the electricity tariff, not the capacity, is the real supply curve.
Ultium converted a Spring Hill EV battery line to LFP for stationary storage in under five months for USD 70 million. Against a USD 720 million, 50 GWh greenfield signed in March 2026 for first output in late 2027, that is the only capacity that can arrive before the 2028–2030 peak — and what it buys is a car cell asked to deliver 12,000 cycles.
The EV battery forecast says 19.03 per cent. The installations data says 16.3, and inside that number four second-tier Chinese suppliers are compounding at 35 to 37 per cent while two of the three largest non-Chinese suppliers shipped less than a year earlier. A total is a weighted average of populations that are not moving at the same speed — and once reconciled against the primary series, the report's own challenger volumes put those four ahead of the two largest non-Chinese suppliers by 28.52 per cent.
Battery-grade purified phosphoric acid is 4.09 per cent of the phosphate market's value and 49.1 per cent of purified acid volume — and it realises USD 1,040 a tonne against USD 1,250 for food grade. The hardest specification in the chain carries the smallest premium, and the pool that grows fastest is the one that cannot raise its price.
3.98 per cent is a fertiliser number. The 3.70 per cent of rock tonnage that goes into a furnace carries 8.70 per cent of the value and is priced by a tariff and a permit rather than a reserve — which is why three growth rates can all be right and only one of them is a decision.
Shandong Taihe lost 19.12 per cent of its revenue in 2023 and no customer left. In a market where the feedstock is 75 to 83 per cent of the product price, the forecast's real assumption is not demand — it is whether the chain can reprice a drum faster than a furnace changes its mind.
LiFSI capacity goes from 2,400 to 12,400 tonnes a year by 2027 — 5.17 times — inside an electrolyte market that compounds at 17.27 per cent. The value growth in this report depends on formulation complexity, formulation complexity depends on the additive, and the additive is the hardest thing in the chain to qualify.
Storage cost curves fell because cells fell. In 2026 the levelised cost of storage rose 36 per cent without the tax credit and the binding constraint turned out to be a semiconductor: an 18 per cent IGBT and SiC capacity gap and 30-to-45-week lead times will hold a fully contracted pipeline back faster than any cell price will move it forward.
Chinese yellow phosphorus ran at about 63 per cent of installed capacity in 2025 — which sounds like a third of the industry standing idle. Against the effective capacity band the same output is 83.33 per cent. The denominator is administered, and that is the entire supply thesis for every phosphonate above it.
The cobalt problem is not that there is too little of it. It is that the amount that may leave one country in a year has been written down to a number with a date on it — and 20,000 tonnes, roughly USD 1.1 billion, sitting against a 96,600-tonne cap is a supply risk that can be scheduled rather than forecast.
USD 91.7 billion becomes USD 369.5 billion by 2033 — a four-fold value claim about a product whose unit price falls every year, which means the volume claim is closer to 5.6 times. And the largest material inside it, LFP cathode, grows at 8.54 per cent and halves its share of the market that contains it.
Cell prices fell 20 per cent in 2024 while the delivered cost of utility-scale storage rose 36 per cent in 2026. The gap is not a chemistry problem — it is an eligibility test, a tax credit doing the work of an entire cost curve, and an inverter shortage that moves commissioning dates.
Two published figures for 2025 sodium-ion output differ by 5.8 times. Cathode tonnage per gigawatt-hour is a property of the chemistry, not of the factory — and it settles which number is real, why 435 GWh by 2030 is a capacity schedule rather than demand, and what 53 times the 2025 cathode tonnage would actually require in kilns.
LiPF6 tripled in six months and the electrolyte followed it up. But the salt's cost sits in a fluorspar mine, its price in a lithium contract and a sulphur chokepoint, and its specification now in a fire test. Three suppliers, and only one of them makes batteries.
Phosphoric acid is P(V) and priced off rock and sulfur. Phosphorous acid is P(III) and priced off an electric furnace, because it is the entry point to the carbon–phosphorus bond. Fifty-eight per cent of a tonne of HEDP is the yellow phosphorus inside it, and that one number reorganises the whole USD 855 million pool.
An LFP cathode maker invoices RMB 31,068 for a tonne and gets to keep a fifth of it after the lithium and the iron phosphate are paid for. That is not a materials business with a materials multiple — it is a conversion fee with a rock attached, and the whole eight-year forecast hangs on which of the two you think you are buying.
NaPF6 is 61.87% cheaper than LiPF6 per mole of charge carrier. Then the cell drinks 30% more of it, and the electrolyte ends up 1.82% dearer per kWh — while the salt's price is set by the lithium market it is supposed to displace.
Sodium-ion is meant to be the cheap chemistry. Its anode is 48.15% of the active-material cost from 31.55% of the tonnage — and substituting sodium for lithium is worth at most 29.6% of the cost gap the cell makers themselves quote.
The wet process beat the thermal route on cost for as long as sulfur was cheap. Then sulfur went to USD 820 a tonne, the advantage reversed by RMB 1,400 a tonne, and 2.48 million tonnes of new Chinese purification capacity meant the cost could not be passed on.
