The world around us is a huge mess, unless you have some degree of understanding about the fundamentals that drive it. And lately Artificial Intelligence researchers have come to this conclusion too. They termed this the "foundation model", and it is the holy grail of AGI at the moment. But it needs huge quantities of computer power for now.
In my humble opinion, Tesla is currently best positioned to hit the jackpot on this: I should say rather, it is not a matter of luck, which a jackpot implies. The company has announced that they intend to start Dojo production in July 2023, and will have their compute up in the top five of supercomputers around February 2024, and presumably exceeding 100 exa-flops around October 2024. Can you imagine Tesla's computer power basically going times 5 in 7 months, and then times 3 in the next 8 months?