There isn't much money in writing libraries like NumPy. Why would a company hire a bunch of expensive devs to write some software to compete with Matlab? The market just isn't big enough to justify the cost.
Google hires devs to develop similar libraries... to power their machine learning efforts. There is tons of numerical work in finance, too. I don't think the money is to be made by selling copies of a Matlab-like piece of software, but in the application of the tools.
I haven't used Grumpy at all, and unless it starts supporting C extension modules like NumPy I doubt I ever will. Google's numerical computing / machine learning stack (e.g., TensorFlow) is based on Python/C++.
Ken Thompson didn't (still doesn't ?) have commit access because he hasn't been vetted by Google as a competent C programmer. I don't think Google is going to relax its hiring policies no matter who you are.
It's not about money. Apple has $200B just sitting in a bank doing nothing. It's about supply and owning the entire market. What happens when the entire market stagnates because those devs are writing great stuff for Apple exclusively?
Apple also used to buy the entire world supply of sapphire for phones and the entire world supply of flash memory for music players. Literally, nobody else could compete, because Apple contracted it in bulk, first.
Those days seem long gone now. Today Apple takes 5 years to update their Pro desktop, while at the same time, making fun of people for using 5 year old computers.