There is a lot of evidence that "primitive" static type systems add value via enabling better tooling.

Low cost of adoption + clear minor tooling benefits is usually enough evidence to adopt Java/C# style static types.

Haskell style static typing has a very high cost of adoption and so must conclusively show a strong benefit in order to be adopted by industry.

At the moment I feel Haskell companies can get away with using Haskell because PL enthusiasts are willing to absorb the training costs.