Jevons Paradox in Simulation

Earlier this year, 19th century economist William Stanley Jevons had something of an extremely-late-career renaissance.

Jevons studied the coal industry. He described the counter-intuitive effects of improved coal-consumption efficiency: In short, he postulated that cutting down the coal requirements for existing use-cases would not decrease demand but would instead increase it. Through what economists now call a rebound effect, efficiency would drive new applications — new use-cases, new users, potentially entire new industries — that would more than offset any short-term reductions in consumption.

This effect — the Jevons paradox — re-entered the Zeitgeist via this X post from Microsoft CEO Satya Nadella:

"Jevons paradox strikes again! As AI gets more efficient and accessible, we will see its use skyrocket, turning it into a commodity we just can't get enough of."

Setting aside the implications for AI writ-large, this discussion of the Jevons paradox really touched on a question we’d be considering within Mach42: As we drive improved simulation efficiency, what does this mean for the semiconductor industry? More directly: Are simulation providers going to benefit from our presence? Will more efficient simulation usage open up new methodologies that weren’t previously possible? We strongly believe the answer is yes — and we’re already observing it.

Take verification, for example — the process of ensuring semiconductor designs conform to expectations. Previously, designers were forced to restrict tests to a small subset of potential issues due to the time costs of simulation. The efficiency provided by our platform is now allowing them to broaden the net with far greater coverage per simulation run, automating detection of conditions that cause their designs to fall out of spec. Rather than just running through their old process using fewer simulation runs, our tech has opened up the possibility of far more intelligent and extensive testing within the same time frame.

The result isn’t less demand for simulations; the result is better, more robust designs within an equivalent budget. In every discussion we have with semiconductor designers we uncover more possibilities for innovation. Better insights into design performance, faster turnaround during phases of the design cycle, and greatly reduced risk of design failure — all enabled by the acceleration we provide.

These results underpin our vision for the future of semiconductor design. At its heart is the Jevons paradox: We’re not just in the business of more efficient coal engines — we’re in the business of powering an industrial revolution.