We’re both correct.
LCOE is based on total operating costs of new electric power generation station over a 20+ year operation life. There are obviously a lot of assumptions in these sorts of analyses but Nat Gas is projected to become cheaper than Coal over the life of a new project, which some of that is expected to be due to carbon taxes.
LCOE has some flaws as a comparable number when comparing wind and solar to fossil fuels, but is good for understanding what will be cheapest to build of fuel based generation.
For current existing power stations, coal is cheapest of the fuels. The EIA numbers are here and here’s Statista research here on the historical cost of nat gas vs coal specifically which is frustratingly why coal phase outs have been so slow. Keeping existing coal plants operating is cheaper than building new almost anything.
And you are correct, price is specific to geography and availability of each. My blanket statement of “coal being the cheapest fuel” is over generalized and not universally correct.
Who cares? Because I assure you, Microsoft doesn’t.
20-25% of those webservers are running on Microsoft Azure hardware. Microsoft is the #2 cloud provider and has been slowly but closing their gap behind AWS in recent years. All of that is in large part due to them embracing Linux and open source support on their platform.
Software isn’t the battleground, and hasn’t been for a decade. The people behind Apache and Nginx aren’t making bank on their web server dominance. Microsoft and AWS still rake in money hand over fist regardless of what software runs on their servers.
The author of this article’s apparent attitude that this is some kind of indicator of Microsoft’s market failure is one of the most ridiculous conclusions I’ve heard in a while.