If AI can generate an application from a description, is software still worth anything?

I have lived with a version of that question longer than most.

I released Drupal for free more than twenty-five years ago, and later co-founded Acquia, which has grown into a large enterprise software company built around Drupal.

Granted, Drupal is free in a different way than AI-generated applications are free, but I'm not sure that changes the basic question of how to build a successful business around either one.

Open Source made code abundant by giving people broad rights to use, modify, and redistribute it. AI is lowering the cost of producing code. One lets you copy the software; the other makes it cheaper to recreate software.

§Free code changes what customers pay for

Because anyone could use Drupal for free, Acquia could never build a durable business around access to the code. From the start, we had to make money another way.

We built that business around helping enterprises build, run, and manage Drupal applications throughout their lifecycle. That includes hosting, but goes well beyond it: the tools and services needed to develop, deploy, secure, scale, monitor, and improve applications in production.

Proprietary SaaS typically bundles access to the application with the hosting and operations required to run it. With Open Source, organizations can run the software themselves or choose who hosts and operates it.

As AI makes applications cheaper to recreate, the traditional SaaS bundle of software and operations comes under pressure. Customers may become less willing to pay for access to application functionality without becoming any less willing to pay to run and manage applications in production. For Open Source businesses those economics are not new.

§Dependability becomes the product

Software can be free, or nearly free, without becoming cheap to depend on. The more people and organizations depend on a system, the more of its value comes from operating it securely, reliably, and at scale.

Once people depend on an application, the cost of its failure has little to do with how much it cost to build. An application that costs $1,000 to build can still cause a $10 million failure.

As AI makes enterprise applications easier to create, adapt, and integrate, they still have to be deployed, secured, scaled, monitored, and run reliably over time. As software cost comes down, dependability becomes a differentiator.

Linux is abundant; dependable cloud infrastructure is a service worth paying for. Drupal is abundant; dependable digital experience infrastructure is a service worth paying for.

Acquia has lived with those economics for nearly 20 years. Drupal made the code abundant, so we built our business around helping organizations build, run, and improve what they created with it. As AI makes code cheaper to generate, that business model may start to look a lot less unusual.

Either way, more software companies will have to answer the same question: if code and capabilities are abundant, what are customers really paying you for?