>with a very small pay-off given the tiny percentage of users who develop on their Mac for iOS devices.

If they go that way, the pay off obviously wont be "helping iOS developers have faster simulations", but as the parent already implied, beta testing and driving forward the eventual mix of iOS apps and macOS. So the "faster simulations" would just be a step in that process helping with perfecting it in a low stakes situation.

Plus, if it indeed happen, as the article also states, it would be used for another payoff, always on (at least when on mains power) Hey Siri, plus controlling the touch strip perhaps.

At the rate Apple is increasing its iPhone SKUs every year, it won't be long until all iOS app companies will need to purchase like 20 different-generation iPhones. I think they'll need the emulator eventually.

Not sure why is this being downvoted. Apple is currently selling 8 models across 4 lines (6s, 7, 8, X and SE), and the iPhone 6 line is still being sold by carriers. This is pretty much the first time when 8 primary models are available from Apple + at least 2 models are still being sold.

If the model trend continues to grow as it has been since the iPhone 6, there might be actually a need for device model emulation for development...

Have you seen how many Samsung Galaxy variants exist in comparison?

https://en.wikipedia.org/wiki/Samsung_Galaxy#Phones

That's just one manufacturer.