If done well, team specific data warehouses could be one form of a data mesh, but I doubt that's the usual implementation companies will go for, especially since on the consumption side of data we have machine learning engineers as well for whom data warehouses are a very poor form of interface for anything production like (again - usually).

Data mesh = a simple form of decentralizing things in a tech company. More specifically it's the decentralization of owning, transforming & serving data. (in MY words, not sure that's that the ThoughtWorkers also have in mind).

That's all, any specific implementation might feature team specific data warehouses or whatever the team chooses to use. Or what the larger unit decides to use, since decentralization does not necessarily mean into the team, it just means getting it away from a central unit.

Hope that helps.