And we all know the Giga Chad, 6'3 Einstein's on this board are definitely never fucking up and part of the problem at a workplace like 'ole Sanjay. I always forget how incredibly efficient, sexy, and intelligent this forum is.
If you have data on "major software issue causes X dollars worth of "damage"" compared to the labor costs of said company, I'd love to see it. I'll bet even millions of dollars worth of fuckups every X years still saves them more money than hiring Timmy the White Nerd who "knows his IRREPLACEABLE WORTH TO THE COMPANY!".
You will never see that data because bigger organizations tend to dissipate responsibility and atomize guilt for fuckups, especially big ones, because it would make management look bad. And management is like East Asia in that regard, you can't lose face. Especially to the public.
However, I have witnessed multiple major fuckups. One of the bigger ones wasn't even software related, but off-the shelf hardware: buying a cheaper file server with unproven technology (which was bought by a major hardware vendor with billions from a startup) instead of the one we recommended for their quite complex heterogeneous setup. Ironically it was developed by Indians, and we also found it funny that the mkfs for their "secret sauce" filesystem contained ext3 code strings.
Long story short, it was a total disaster, and didn't work as advertised. One of the middle managers tallied the costs (IT costs babysitting this thing, and costs due to loss of work). It was a very ugly number with 7 figures, and it was more than the cost for the setup that was recommended to them. So instead of saving the company money, they actively lost more than they could ever hope to save.
-- timeskip --
You'd think after a few years, when replacing such a system was in order, they would've learned a valuable lesson from all of this, right? No. They doubled down, and I quote: "Because we have to allow vendor to make it right."
So they brought in the newer replacement system, and it was even more of a shitshow than the last one. The vendor gracefully exited when they saw what a fuckup it was going to be (again), and said "engineering said you can't have this feature". A feature that had been in the required specifications since the very beginning of the first system. So in the end it was not the company that pulled the plug, but the vendor.
In the end, they had to go with the system we recommended. And that system never went down, not even a single time in the 4 years we had it. It was replaced by a bigger system from the same vendor, which also never went down once.