SaaS is a zombie on a treadmill

Software as a Service is dead. It just takes a long time for a zombie to fall over.

For the last fifteen years, the Software-As-A-Service business was lucrative for investors. Build a piece of software once, host it centrally, and charge people a monthly rent to access it. The marginal cost of adding a new user rounds to zero. Wall Street loved the recurring revenue, so valuations went to the moon.

But the math stopped working for the people actually using the software.

We are drowning in subscriptions. Companies are paying $30 a month per seat for note-taking apps, project managers, CRM tools, communication platforms, and specialized HR portals. It's death by a thousand paper cuts. And the software isn't getting better. It's getting bloated. Founders add pointless features to justify raising the monthly price, making the core product slower and harder to use.

What kills SaaS won't be a sudden crash. It will be a thousand frustrated employees vibe coding software thats better than the bloated crap your managers licensed for them. It will be bottom up, the gradual realization that most of these tools don't need to be rented anymore since almost nobody is using them enough to justify the license.

As open-source alternatives get better and in house software becomes incredibly cheap, the tolerance for paying perpetual rent for basic utility software will vanish. Why pay a monthly fee for a web app on top of a database when you can deploy a perfectly good custom or open-source equivalent to a cheap, stable cloud server in three days? Atlassian, Github, Salesforce, ServiceNow, are all living on borrowed time.

The future of SaaS is selling tokens for your own models, to let customers make what they need.

What about Microsoft? I'm not sure. Azure is pure gold, but MS software is not so special anymore. I can run any custom AI written software on Linux just fine. MS Teams and the 'new' Outlook are written in Javascript anyway, they don't even need Windows. I see why MS is so desperate to push co-pilot (all 93 versions of it).

The companies building complex, data-heavy, genuinely useful software will survive. The companies renting out basic database interfaces wrapped in pretty CSS will not. The market is just waiting for the current generation of SaaS contracts to expire.