That's an interesting point. Assuming that the Windows monopoly imposed hidden costs on the public in the First World, to what extent does Gates' philanthropy offset those costs? If it hadn't been funneled to Gates (and eventually, his foundation), where would it have been used? If the public had known during the antitrust "Micro$oft" era that Gates would donate most of his fortune to fighting disease in the third world, would attitudes have been different?
Unfortunately, I don't think it's possible to know the answers to these questions, but it's an interesting door to open.
Assuming Windows instead generated gains of historic proportion to the public in the first world, by consolidating a fractured ecosystem down to a dominant standard, helping to generate mass public adoption of computing, and in doing so driving productivity through the roof, then what Gates is doing is still one of the greatest things any human has ever done.
People often assume that if it weren't for Windows, something better would have existed in its place. I regard that as a fantasy. Windows was 85% good enough, cheap enough, and easy enough to use for its time. There's a perpetual and immense bias against Windows in hacker circles, but that says nothing of the non-hacker users that voluntarily made Windows successful. Microsoft didn't wake up one day with a magic monopoly, users voluntarily chose their software over other options for a decade before their monopoly position was finally in place.
Microsoft is criticized for not being innovators, instead just regurgitating ideas pioneered by someone else. There's truth in this criticism, but we can also view it differently. Microsoft's success was in figuring out how to commoditize the bleeding edge tech.
For example, I remember early in my career the industry trying to argue out CORBA standards, something that never achieved any widespread acceptance because the perfect solution they were after was endlessly being debated, and so complex. But Microsoft cut through that Gordian Knot, and put out their own DCOM technology. DCOM didn't achieve all the goals of CORBA - not by a long shot - but it put a viable object brokering technology into the hands of every developer, a fundamental requirement that the CORBA folks couldn't achieve.
Having lived through the computing renaissance of the '80s (that ended with Microsoft dominance through the '90s and 00's), I do not assume that something better would have come along.
Early 80s were a mess, with tens of competing hardware/os combinations (e.g. in 1985, you had C64, Spectrum, QL, BBC B+, AppleII, MSX, CPC664, Elan, Oric, Dragon, PC-XT, Lisa, Mac, Amiga, TI-99/4A, Atari {ST,XL,XE} and a few others I probably forgot or ever knew). Each of these was usable computer with its own ecosystem, and they were entirely incompatible with each other. Just like in the early 20th century, there were over a hundred car manufacturers.
Consolidation was imminent. But consolidating to a single vendor (Microsoft), I believe, was bad for the market. Three or four hw/os combinations for home computers would have been much better overall. E.g. the Amiga in 1985 had better sound and graphics than the majority of PCs in 1993. It also had a functional multitasking OS that worked better with 512KB of ram in 1985 than Windows did with 4MB of ram in 1992.
With microsoft becoming a single standard, there was no pressure to become better. And contrary to what some people here think, microsoft did NOT win the war fairly. e.g. Users did not choose BeOS because microsoft strongarmed PC makers into not including BeOS.
It seems unlikely that how charitable the company or person you're buying from would have a meaningful impact on consumer decisions. If that were so Newman's Own or similar companies would dominate over other food brands.
Unfortunately, I don't think it's possible to know the answers to these questions, but it's an interesting door to open.