DeepSeek has thrown the AI world into disarray. Its new model, R1, isn’t just a technological breakthrough — it’s an economic one. Promising groundbreaking efficiency in AI compute, it should, in theory, reduce the demand for costly infrastructure. But history tells us a different story. The more efficient something becomes, the more people use it. This phenomenon, known as Jevons Paradox, is reshaping how we think about AI.
This post explores why efficiency in AI doesn’t curb demand — it amplifies it. From historical lessons to today’s booming AI industry, we’ll uncover why the race for better, faster, and cheaper compute is only fueling an insatiable appetite.
In 1865, economist William Stanley Jevons made a surprising observation. As steam engines became more efficient, coal consumption didn’t decrease — it skyrocketed. The reason? Efficiency made coal cheaper to use, which led to widespread adoption across industries.