Ed Zitron[1] has a lot of articles and podcast episodes on these deals. The nice thing about it is that he occasionally revisits the old announcements to check what happened with them. Apparently a lot of these deals just evaporate after prolonged contact with reality.
This stuff helps prevent the bubble popping, which means when they want the bubble to pop, they stop announcing these deals, giving them a great lever of profit.
Crazy how they can just lie to this extent without consequences. Or still get paid millions for making bad deals, meaning incompetence
Ed Zitron is a terrible source, he is so staunchly anti AI that he is effectively blind. Once in a while he'll be right by pure chance, but I wouldn't rely on any of it.
I think a tell that many of these deals likely aren't real and are basically just PR is the numbers are super round and digestable.
That's a clear signal that little analysis has gone into the numbers and, most generously, there's nothing but the shape of a deal the details of which will be ironed out and adjusted in practice.
I get that the amounts of funding and capital being sat on for the respective parties are collossal and lead to rounding that doesn't make sense from the point of view of an individual any more (what's a few million at this scale, just round up to nearest 10, etc) but deal sizes of literally round numbers of 100s start to stretch credibility on whether any real analysis was involved.
In fact it'd be a ridiculous coincidence if it had been. They're the kind of figures where you'd recheck your calculations to check it's right as it seems too perfectly round.