I've had a predictive maintenance startup. We managed to bring value for local gas distribution company, for mining machines provider and 100+ wind turbines owners. But in the end we needed to let it go - mostly due to lack of focus, long sales funnel and high costs of data scientists / software dev against to what value PdM brings. We dreamed that it will be repetitive Saas, but we always manages to implement a custom, hard to maintain thing.
