Growing unease surrounds the artificial intelligence industry regarding potential infrastructure overbuilding and the issue of who ultimately bears the financing risk. With these concerns in mind, how are GPUs genuinely funded in practice? 🧵