When costs drop, people don't buy less—they spend the same and buy more.
This counterintuitive truth has played out across industries for centuries. And it's about to play out again in software development.
The Jevons Paradox
In 1865, economist William Stanley Jevons observed something strange: as steam engines became more efficient and coal cheaper, Britain consumed more coal, not less. Efficiency improvements didn't reduce consumption—they increased it.
The same pattern appears everywhere. When semiconductors got cheaper, we didn't use fewer chips—we put them in everything. When cloud computing costs dropped, we didn't reduce our usage—we built more applications. When storage became affordable, we didn't store less data—we stored everything.
The Open-Source Revolution
Consider what happened in software development from 1998 to 2005. The cost of building software dropped dramatically.
In 1998, there was no AWS. You had to buy and host your own servers. Python and Ruby weren't really viable yet. For your database, you'd pay Oracle huge sums—MySQL had just come out but wasn't production-ready until version 5.0 in 2005.
By 2005, everything changed. AWS launched. Python and Ruby became mainstream. MySQL was viable. Open-source tools were everywhere.
This didn't kill software as a business. It led to a lot more software being built—but it didn't make software companies less valuable. It made them more valuable.
The On-Premise to SaaS Transition
The same pattern played out with the shift from on-premise software to SaaS. In 1998, the largest software companies were Microsoft, Oracle, SAP, Computer Associates, Adobe, Intuit, Novell, Autodesk, and Sybase.
Some of those names are irrelevant today. But it's striking how many of the largest names in software and SaaS today are the same names from 1998. Microsoft, Oracle, SAP, Adobe, Intuit—they all successfully transitioned.
How did they do it? Microsoft built Office 365, shifting from boxed software to subscriptions. Oracle moved its database and enterprise software to the cloud. SAP launched SAP Cloud Platform. Adobe transformed from selling Creative Suite licenses to Creative Cloud subscriptions. Intuit pivoted QuickBooks from desktop software to cloud-based accounting. Each company took a different path, but they all made the transition.
It took many of these companies years to execute the transition, but they had time. The market, in its shoot-first-ask-questions-later mentality, underestimated how long companies had to adapt.
What This Means for AI
Today, we're seeing the same dynamic with AI and software development costs. As coding models make it cheaper to build software, some worry this will erode the value of software companies.
But history suggests the opposite. When development costs drop, companies don't spend less on software—they build more of it. They raise the bar for what's expected. "What do you mean you don't integrate with our CRM?" customers will ask. "What do you mean you don't connect to our accounting system?" "How can you not have a feature that handles our specific workflow?"
As the cost of building software comes down, we're going to build massively more of it. New market entrants will have tools to ford existing moats? The new moats will stretch a hundred miles in every direction.
The Pattern Holds
The cost paradox isn't unique to software. It's a fundamental economic principle: when something becomes cheaper, we consume more of it, not less. When refrigerators got cheaper, households didn't buy one—they bought multiple. When TVs became affordable, families didn't stick with one—they put screens in every room. When air travel costs dropped, we didn't fly less—we flew more often and to more places.
Cheaper software development won't kill software companies. It will make them more valuable—because they'll build more, compete on more dimensions, and create more value for customers. The companies that understand this and execute on it will win.