The same story as drugs, bigger market
Batteries, catalysts, superconductors, solar cells, fertilizers: materials problems are electron problems, and electrons are quantum. The reason your phone battery chemistry improves slowly is that predicting a new material's behavior from first principles is often beyond classical simulation, so labs iterate by expensive trial and error.
A quantum computer that could accurately simulate candidate battery electrolytes or catalytic surfaces would compress that loop: compute first, synthesize only the promising ones. That's the pitch, and unlike many quantum pitches it targets a genuine classical bottleneck rather than a solved problem.
Where the value would land
Three examples the field keeps citing. Better battery chemistry: simulating lithium compounds and next-generation cell chemistries accurately enough to guide design. Better catalysts: nature's nitrogen-fixing enzyme runs at room temperature while industrial fertilizer synthesis consumes a meaningful slice of world energy: understanding the enzyme's mechanism is a quantum-simulation problem. High-temperature superconductivity: still not fully understood decades after discovery; the underlying models are exactly the strongly correlated systems quantum hardware naturally represents.
Note what these share: the industries that would capture the value (energy, chemicals, automotive) are mostly not the companies building the machines. History says the users of a tool often profit more than the toolmakers.
The realistic state today
As with drugs: today's demonstrations run on molecules small enough for classical methods, so current partnerships are about method development and readiness. Carmakers have been early movers (battery chemistry is their existential problem) and several have standing programs with quantum hardware companies, documented below. The milestone to watch is identical to drug discovery's: a materials calculation that beats the best classical alternative on a question someone would pay to answer.
A ressalva honesta: High-value targets generally need error-corrected machines.