I'm sure it's just as you describe at many large companies. But it doesn't work that way at my university (which is quite large).
There you have two main classes of PC buyers: The administrators (HR, accountants, fund managers, etc.) and the researchers (faculty and their research groups).
The administrators purchase them using the budgets for their departments, buying them outright and continuing to use them until they need to be replaced. The researchers acquire their computers with grant money, and also buy them outright. When the faculty no longer need them (or leave the university), they are transferred to the administrators, who can continue to get use out of them.
There may be certain groups at my university that instead lease, but I don't belive it's standard practice.
There's a reason the term "companies" was specified rather than "Universities". Completely different goals, operating models, funding mechanisms, profit motives, liabilities, etc. You also conveniently left out the biggest purchases of computers in a higher education setting: the students. They don't care how the university procures computers because they're buying their own machines.
I strongly disagree with your assertion that "Outside of gaming ... the push to continually upgrade the GPU pointless."
High GPU performance is needed for video production work (which I believe is a large part of the target market for the Studio and MP), scientific visualization and GPU-based scientific computing, and AI and machine learning.
Even so, that is still a small percentage of the overall user base for computers, regardless of operating system in use. I'm not sure why you keep trying to push niche and edge cases as the overall trend, but it's not working.
That's not a functional argument, it's financial -- it's just illustrating why more laptops are sold (bigger base, more frequent replacement), and is already subsumed into the 80% figure referenced above.
But I've already stipulated to that. So let's set aside the financial question--no point in arguing about what I've already agreed with!—and focus on the functional, by asking this simple, big-picture question:
Suppose you are either a laptop buyer or a deskop buyer. Which one, on average, will have more of a functional need for the machine they're buying to have the lastest and greatest chip?
Your rhetorical question is based on several flawed premises. First, the only users who would benefit from the "LATEST and greatest" are those who would be pushing the system to its breaking point, which would be a very small percentage of the overall userbase. For most users, it doesn't matter whether they're using a 13th or 14th gen Intel system, a Ryzen 5000 or 7000 AMD system, or an M1 versus M3 Apple system. Second, as long as the machine meets the customer's needs (function), what's under the hood is irrelevant.
Talking about HOW and WHERE the computers are used is the very definition of "functional". Speaking of function, the ability to upgrade desktops when the laptop industry has increasingly moved towards Apple's model is a functional consideration with financial implications. Instead of replacing an entire desktop outright, it is possible to upgrade the RAM, GPU, CPU, etc. This has the functional effect of extending the lifespan of the machine while also providing a financial benefit to the consumer by only upgrading select components instead of replacing an entire machine.
Yeah, a lot of desktop buyers don't need it, but neither do a lot of laptop buyers. Indeed, I would argue that, on average, desktop user are more performance-focused (or, more precisely, the percent of desktop buyers that need ultimate performance is higher than the % of laptop buyers). Thus, if anything, it's the desktop buyers who are more in need of the latest chip.
The majority of both desktops and laptops sold are not performance-oriented models but those which fall more into the value category. This actually works against manufacturers because those systems carry lower margins (and therefore lower profits) than models targeted at the high performance end of the market. Additionally there are certain demographics who will always go with a desktop over a laptop not because they need a higher-end machine with the latest and greatest components, but because they have always used desktops, are comfortable with the usage model, and need/want a larger display for reasons of comfort rather than performance. As long as their machines can go online, send email, run Word, and access Google, the users are happy.