When Software IP Stops Being the Moat

Salesforce can go headless because it holds your data. Can Adobe?

I first heard about ArtCraft last week and the coverage has spread across my feeds on news and technology. The story is interesting because it makes you think about how a previous idea of software IP as moat is changing.

As of October 12th, I see roughly 100K GitHub stars that make up the ArtCraft repositories. One of the products, PhotoCraft, is in alpha with about 8.9K forks and 250 contributors. If even 50% of the contributors decide to point Codex and Claude at the code base the speed of change will be extraordinary. This of course means the engineering focused on the code pipeline to keep up will need to be exceptional. Rigorous automated regression testing, quality standards and security scanning will need to be impeccable to keep momentum.

Adobe's moat is different from Salesforce's. Time will tell if it is enough. Free clones of Adobe's software have existed for years, but the speed that AI can replicate functionality gets better by the quarter. They still own the professional designer, patents, stock assets and plugins ecosystem. On the other hand will the enthusiast ever make the transition to professional with a product that can get to parity and can be programmed to use AI models that can do visual work?

Everyone should care because this story is not just about Adobe, it is also about every company you work for or own. Consider what happens when software intellectual property is no longer a primary component of the business moat.

This brings up interesting questions when evaluating business risk.

1. Is there a network effect because of the number of users interacting with the business?
2. Has the business built a network of people that provide services that would be difficult to manage?
3. Does the business own assets that make it hard to compete in their industry?
4. Do people have brand loyalty and have a good sentiment about the business?
5. Does the business have regulatory or capital requirements that make competition difficult?

Back in March I asked what SaaS survives when your team is five people and 20 agents. I made the point APIs are not enough. Salesforce announced Headless 360 in April by understanding that the new clients are agents. Headless means decoupling the logic from the user interface. This works for them since they have customer data, enterprise agreements, and migrations are expensive. I also think security and patching are more valuable than ever before with AI enhancing hacking and repository poisoning capability. Patching and security are part of the incumbent moat.

This also makes me wonder how new and existing companies succeed when data, network and existing relationships are not part of the equation.