Something transformative is happening in software, and the numbers from Battery's 2024 OpenCloud report tell an interesting story. Most people are focused on technological advances in AI, but I think most of the folks are missing a deeper shift in how software creates value.
Traditional SaaS companies built tools that made humans more efficient - but, look at what's happening with cloud providers. Their CapEx growth has shot up 73% y/y in Q3 2024, far outpacing their revenue growth. Why spend so aggressively if you're just building better tools? Because they're not. They're building infrastructure for software that does the work itself.
This shows up clearly in the economics. Take call centers: there are 16 million agents globally with a $200B annual salary spend. Old software tried to make these agents more efficient. New software simply handles the calls. The addressable market just jumped from a fraction of the technology budget to the entire labor budget.
The same pattern appears in legal ($515B labor spend) and architecture/engineering ($270B labor spend). When your software does the work instead of helping with it, your market expands dramatically. That's why Battery estimates a $4T+ opportunity. It's not just eating software budgets anymore - it's eating labor budgets.
This shift changes everything about how software companies operate. Look at the revenue multiples as well — private AI companies are trading at a 3.2x premium to public SaaS companies. The market isn't just paying for growth anymore - it's paying for labor displacement (well..) potential.
Cloud-native companies that have made this shift are growing at a 34% CAGR even at scale; that's not just impressive growth - it's growth that defies traditional software scaling limitations. When your software does the work, you're not constrained by how many people can use your tools — and this IMO is such a nuanced, yet landmark idea shaping the new narrative.
But here's what's really interesting: only 37% of AI use cases are in production today. Most companies are still experimenting. This gap between potential and current reality explains why private markets are paying such high premiums. They see the flip coming.
The cloud providers' behavior is particularly telling. Public cloud providers have grown their combined run-rate revenue to $221B, but they're spending even faster on infrastructure. They're building for a world where software does more of the work directly.
Look at their revenue structure: legacy infrastructure pieces are growing at 19-21%, while AI-related services are growing at 44%. The market is shifting from computing-as-a-service to work-as-a-service.
I think we're witnessing a fundamental transformation in what software companies do. The old model was building tools for human workers. The new model is building systems that do the work. This isn't just automation - it's replacement.
This creates weird dynamics. When you sell tools, you want to make humans more efficient. When your software does the work, making humans more efficient actually shrinks your market. The incentives have flipped.
The companies that win in this new era won't be the ones with the best tools. They'll be the ones that figure out which parts of work can be eaten by software, and build reliable systems to eat them. That's why the ServiceNow story is so telling - they've grown from basic IT tools to $10B in revenue by steadily eating more types of work (from 11 month ago..).
This feels like one of those shifts that will seem obvious in retrospect but is non-obvious while it's happening. Just as the move to cloud changed everything about how software companies operated, this shift from tools to work will change everything again.
The interesting question isn't whether this will happen - the numbers show it's already happening. The interesting question is what software companies will look like when they're primarily in the business of doing work rather than helping with it. That's the great flip happening in software right now.