"Energy is too cheap" is only a problem when the people who own whatever equipment is generating energy are interested in a direct return on their investment. From the perspective of Kardashev, "Energy is too cheap" is a fantastic opportunity for a civilization.

While yes, too much surplus capacity would eventually indicate overinvestment, for now it seems like great news for the compute and software industry in China. If you reliably have times when energy is better-than-free, that's the ideal time to scale up and generate embeddings from your stored data, or train LLMs, or do all sorts of other batch-able but intensive tasks.