Where does economic value accrue as AI continues its exponential progress?

We’ve been thinking for a while that a massive amount of the value created by AI will accrue to services companies. And as AI has progressed, that view has only been reinforced.

The SaaSpocalypse six months ago was one signal. The conversation then focused on the threat AI poses to SaaS. Software stocks have rebounded sharply since, but to me the more interesting implication is what happens on the other side: as AI reshapes the economics of software and dramatically increases productivity, who captures the value?

I think services companies are particularly well positioned, for two reasons:

First, the markets are much bigger. Healthcare, legal, accounting, consulting and other services represent trillions in economic activity. The software markets serving them are a fraction of the size.

Second, services companies have every incentive to keep the massive operating leverage AI creates for themselves. If AI lets a company do twice the work with the same people, why give much of that value to a technology vendor? Vendors can try to capture more through outcome-based pricing, but outcomes are notoriously hard to define, attribute and price.

You can get a glimpse of how this may play out in healthcare. The Affordable Care Act accelerated the shift toward value-based care, where providers can be rewarded for improving quality and outcomes while controlling costs. But 15+ years later, adoption has been slow. Measuring outcomes, establishing benchmarks, attributing results and aligning financial incentives are all complicated.

Some of the most interesting models in healthcare sidestep part of this problem through vertical integration. Kaiser Permanente combines insurance and care delivery. Devoted Health combines a Medicare Advantage plan with its own technology, navigation and medical services, alongside external provider partners. They participate much more directly in the economics of the care they deliver. When technology helps them deliver better outcomes at lower cost, they have the opportunity to capture much more of that value themselves.

AI could bring that same logic to many other industries. Maybe the SaaSpocalypse was the first-order question. The next is what happens when enormous services markets start to have software-like economics.