The promise and perils of sodium-ion batteries

Sodium-ion attracts attention for three theoretical reasons, and Yao attaches an asterisk to each. Kann ranks the supply chain case the most certain of the three and safety the least; Yao agrees safety carries the biggest asterisk, but he hedges the supply chain case too, saying the geopolitical advantage holds in some cases and not in others. The raw spread is striking: soda ash, the sodium feedstock, runs about $200 a ton against roughly $10,000 a ton for lithium, which reached something like $60,000 to $80,000 at the 2022 peak. More than 93% of the world’s battery-grade graphite comes from China, which had recently imposed export controls on it. The drop-in manufacturing case is that a sodium-ion cell is the same physical sandwich as a lithium-ion one, built on the same equipment in the same order, so the capital equipment supply chain already exists. Yao narrows that: a small Chinese cell maker can genuinely switch a line, but a dedicated factory at the 10 to 35 gigawatt-hour scale is designed around a single product and, he says, you are most likely not going to experiment much at that scale. So the honest version is that a greenfield sodium-ion plant is a smaller leap than a genuinely novel chemistry, not that gigafactories get retrofitted. Safety gets the largest asterisk of all. A blanket claim that sodium-ion is safer is, in Yao’s words, objectively wrong, because safety tracks the specific cathode chemistry rather than the ion.

Kann then says what he thinks of all three, which is that none of them matter except through cost, and the rest of the conversation is about cost. Yao’s frame is an identity. Dollars per kilowatt-hour equals dollars per kilogram of material times kilograms of material per kilowatt-hour, and the industry celebrates sodium’s advantage on the first term in the same breath as admitting its disadvantage on the second, without multiplying the two together. The multiplication is unforgiving. His illustration: suppose it costs $3 a kilogram to manufacture LFP cathode powder and the minerals add $2, so $5 a kilogram, and that kilogram yields about 500 watt-hours. That is $10 per kilowatt-hour of material. Now take a sodium equivalent that costs the same $3 a kilogram to make but yields only 250 watt-hours. Even with its minerals free, it lands at $12. Cheap inputs lose to a bad denominator.

The same arithmetic explains where lithium-ion’s cost decline actually came from. Yao puts lithium-ion’s thirty-year learning rate around 20% to 22%, and says the larger share of it came from better cell design, meaning less inactive material per kilowatt-hour, rather than from cheaper inputs; lithium itself is only about 7% of a lithium-ion cell by mass. So when Kann asks whether sodium-ion is simply early on the same curve and needs scale, Yao says no, flatly. Scaling today’s sodium chemistries scales today’s materials intensity. The investment has to go into energy density instead, which he frames as the one piece of good news, since a developer is not trapped in the usual chicken-and-egg where low cost requires scale and scale requires low cost. The catch is that raising energy density far enough needs materials research and not only cell engineering. The most promising configuration he models is an anode-free sodium cell, which could reach $40 to $50 per kilowatt-hour and cross the line where LFP sits, but only if the underlying cathode capacity and anode-free plating both work out.

Kann then puts the uncomfortable version to him: a decade of research to arrive at roughly where LFP already is, an outcome that assumes LFP stops falling, which Kann flags as an unsafe assumption given that lithium-ion costs have kept surprising to the downside. Yao does not dispute the arithmetic. He offers two answers instead. The first is that the lithium-ion curve may bend upward rather than sodium’s bending down, with the graphite export controls the live candidate; their modelling shows a lithium-ion price shock genuinely accelerates the crossover, though he is clear that this is a kink in LFP’s curve and not progress on sodium’s. The second is to stop competing on cell cost at all, either on total cost of ownership, where a safety or cycle-life advantage lowers system-level capital cost, or by finding a performance niche, and he says US companies are currently split across all three strategies. On whether China is about to settle the question he is deliberately unsure: the builders today are tier-two and tier-three players, the ones to watch are the tier-one manufacturers CATL and BYD, and CATL’s chairman and chief executive had publicly said he expects sodium-ion to take up to half of LFP’s market. Yao’s own read, which he sources to the grapevine rather than to evidence, is that the announcement may function partly as leverage to keep lithium prices low.