It's early. And companies and governments achieving some sort of competitive advantage would be playing these applications close to their chest to keep their advantage.
QC has led to significant prevention of loss-of-life, loss-of-equipment, and territory gains in battlefield applications. Encrypt/ Decrpyt was already mentioned as a use case. None of these dual-use practical applications are public.
QC is good at "analog" computing: applications where you are more interested in potentials than in 0s and 1s. That lends itself to material design problems. There are applications in battery design and catalytic converter. F1 teams have some interesting fluid dynamics and laminar flow problems where QC was instrumental to test hunches fast. It's not so much about exact compute of solutions, but how to verify a first guess fast to then use traditional software. One could argue this is a niche application, but not if you consider marketing and streaming rights and advertisement impact. A similar approach is used for drug discovery.
None of these "practical" applications are talked about - they are a competitive advantage of commercial enterprises. But if you care for a proxy you could look at any significant hires 1-2 years ago with PhDs, compute, math experience in relevant QC fields. And you would be surprised to see some companies hiring more than 20-30 people in that field (and not just Google or Amazon etc.). That's an FTE expense that could just be hedging of a large company; but could also be product design; and is definitely more than a test balloon or R&D pet project. Some of these companies are known for industrial adhesives, specialty cement, or "green" base oils for lubrication and cooling and cutting. so not all tech companies or big pharma.