You could spend the money, but you also need to consider whether that money is well spent. Batteries do not last forever. Maybe that money is better spent on R&D to develop better batteries first. Also natural resources and environmental impact needs to be considered. Batteries take natural resources to build and also occupies a lot of space.

20 years ago, we also have the technology to run AI workloads. Except we probably had to deploy billions of CPUs to match the capability of today’s GPUs. We have the technology then, but it is not practical. And that money was much better spent in the R&D that lead to today’s GPUs. So similarly our batteries probably needs to be a few magnitude better than what we have today before it is practical to use.