Web Scale has changed.

Writing an application that supports millions of users is much more difficult than writing one that supports hundreds. Historically, B2B SaaS apps aimed to grow big enough to need to worry about supporting that many users. Thousands of companies with thousands of employees. Having done that work gave you a moat, it was hard to replicate.

If that requirement is gone, the moat might be as well.

Let's think about some example apps here. JIRA and Slack, or their variants, are likely in the tech stack of most companies working with technology. These products have growth to support truly massive user bases and had the valuations to show it. Unless you're an enterprise user you get mostly the standard feature set, possibly customizable through extensions or add-ons.

Every team that purchased a JIRA or Slack subscription was adding users to the pool. Uptime got harder for Atlassian and Salesforce because most users means more potential for things to go wrong. Engineering teams grew. SRE becomes a major focus. Certifications like ISO27001 and SOC2 become essential because you're trusting a third party with key company data.

Of all the things AI can do, cloning a big app in a small way is high on it's capability list. Some apps, like Spotify or Netflix, can't so easily be cloned because their value isn't in the software. Your ability to create a web app that plays audio or video isn't a threat to their business model.

For JIRA and Slack though, being able to replicate a mini-version might be an issue.

I'm not claiming here that you could rapidly spin up a JIRA competitor that could handle the same usage and growth. I am claiming you probably don't need to. If you're building a single-use JIRA just for your company, you probably have a tiny fraction of the user base. Building an app to that scale is a whole lot easier.

This has effects across the software supply chain. If we don't need heavily replicated data bases on every continent, AWS bills will go down. If the product lives only within the company VPN, the barrier for security gets an inherent boost. Not a free pass by any means but IT teams have become very good at keeping things within walled gardens.

Not every company is going to have the desire or the resources to build their own JIRA or Slack. Especially if they're happily being used right now. There's a long history of purchasing software that needs to be overcome.

This does mean there's likely an opportunity here for new smaller vendors. Especially for vendors willing to take a simple core version of a product and rapidly customize. Enterprise teams are very unlikely to shift in a hurry but there's thousands and thousands of smaller deals out there waiting to be won.

Personally, I find this exciting. I am hopeful we'll see a wave of new ideas, new business models and new remixes of old software in a way we hadn't before. I hope that some portion of that happens enough in public to encourage standing on the shoulders of others and making things that genuinely help people. The moat created from being the largest player in the space always seemed like the most satisfying to erode.