> all sorts of SLAs and support contracts, all sorts of infrastructure for supporting all of that, etc
We're talking about different verticals, here. Product businesses can get away with this. Service businesses probably can't.
Even in service industries, though, there's a simple hack: when #2 acquires #3, they keep #3 running for a few months, but they make #3 push a change to the SLA, where service provided by #2 in lieu of service provided by #3 will be considered an acceptable service-level. They only begin liquidation of #3 once all #3 customers have signed onto the new SLA. Then, when the lights are turned off on #3's infra, all #3 customers are temporarily just provided with #2 services.
Again, this is mostly a thing with product businesses, or, say, logistics providers, where the product/service is "pushed" to the customer by the provider, and the customer doesn't need to change how they do things, they just need to take receipt of the alternate product/service from the new provider for a while. "MomNPop beer is unavailable for a month, but here's a truckload of existing AB InBev 'indie-branded' stock to put in its place until we get our new marque spun up."
This is less of a thing if we're talking about a "demand"-driven business, like, say, a SaaS API provider, where the customer has to communicate to #2 or #3 through an API, and they have different APIs. This can still be made to work, but it's all about how long the transition period takes. Google's purchase of Nest looked like this, with a transition period of two years. But it wasn't temporary; with a transition period like that, you may as well just shutter the #3 "API" entirely.
(Example I can think of where this does happen in a service industry: cell-service MVNOs getting acquired by their own infrastructure provider. If ISP #3, an MVNO who uses ISP #2 as their network, gets bought by #2, there's nothing in #3 that #2 wants or needs, other than their customers, and maybe #3's brand value. So they just tell #3 customers that they're going to be #2 customers for a while (i.e. their phone will be "roaming" onto #2's network all the time, but they'll be charged regular #3 prices for it.) Either every customer from #3 is then pivoted into being a "real" #2 customer; or, if #2 keeps the #3 brand around, their #3 service starts working "normally" again a while later, as just a marque of #2 (if they stuck around with #3 instead of checking the box on their monthly bill that offers to switch them to #2.)
> and immediately fired everyone who knows how to manage it all!?!?!? And what, then just immediately hire several thousand people to replace them 24 hours later, and have every single one of them up and running and productive in a jiffy?
You don't hire the "new"-from-#3 employees to service the #3 accounts. #2 employees service the from-#3 accounts. #3 employees come in at the bottom, as if they were outreach hires.
> followed by #2 being erased from the marketplace by lawsuits within a year or two.
I mean, not if #2 doesn't own #3 when it liquidates, they don't. A company doesn't have to literally acquire another company; they can just pay it to commit suicide, and then pick up the pieces. Who do you sue, if your provider-of-choice #3 just decides to fold?
Or, even more sneaky, #2 can just figure out a way to enter into some kind of financial arrangement with #3, where it then slowly squeezes #3 to death... and then picks up the pieces. (Do you know what happened to Target in Canada? Do you know why? Hint: all the land the Targets operated on was leased to them... by Walmart! And now, after Target set them up nice and neat, and then got squeezed out of the market, they are Walmarts.)