If so - they could not make it work as a business. The next owner could not make it work. The latest owner is taking on a big gamble - but they probably have a better chance since they have a longer runway and the ability to parallel it with Fusion 360.
I'm a software guy who tinkers with hardware, rather than the other way around, so I (think I) have a useful perspective as a longtime EAGLE user. Basically, Autodesk bought the name EAGLE. They didn't receive any software IP of any particular value going forward. Trying to "cloud-ify" a 20+ year old program like EAGLE while bringing it up to modern UI standards would be an Augean task. If the necessary changes were easy ones to make,
CadSoft would already have made them. Instead, the code base has been passed from one acquirer to the next like a
hot radioactive potato.
So if Autodesk thought they were getting an actual finished product to sell, or even one that could be profitably maintained through incremental development, then they made a big mistake, and by now they'll have realized that. I doubt that's the case, because the Autodesk guys are not newbies in the software industry.
If they did know what they were getting into, then we can assume that they allocated a team of strong, experienced CAD developers to the product at the time of acquisition. This team will have been dedicated to the task of rewriting the package from scratch, with nothing retained except the .brd, .sch, and .lbr formats. They will be working entirely behind the scenes, and will continue to do so for some time to come. The resulting product will carry the name EAGLE forward for the next 20+ years. However, unlike the current monolithic executable, it will be engineered with modern tools and practices to grow and evolve from day one.
And that's the point at which a subscription-based licensing model begins to make sense. What they are selling
now is a stopgap measure which doesn't. A screwed-up library manager is the least of the grief that users can expect if Autodesk continues down this path.