The prime example of burning massive amounts of money with which I'm intimately familiar:
Digital Equipment Corporation (aka DEC) spent over a billion dollars in the mid-80s (when that was real money!) developing a water cooled "IBM-killer" computer called
Aquarius, aka
VAX 9000. They sold under 48 of them - lifetime - at roughly a million apiece, before killing the entire program. The biggest program failure of the company, ever.
As was the practice at DEC, there were usually three different engineering teams - one in Marlboro MA that was developing
Aquarius, another in Littleton MA (can't remember what their doomed program was named) [edit: it was
Argonaut], and my group in Maynard MA - competing to develop platforms for each market segment, in this case the High End Computing segment, which wanted to go head-to-head with IBM's mainframes.
I was the Project Engineer in Maynard developing a RISC architecture called
Prism, with the first implementation called
Crystal, and that included Microsoft's Dave Cutler in the architectural specs. It eventually got starved for funding along with the Littleton group's CISC platform, when
Aquarius was picked...
https://en.wikipedia.org/wiki/VAX_9000
Cheers!