AI Will Fix Healthcare Costs, VC Insists, Despite All Evidence

Here we are again. PwC released a report this week estimating that medical costs will increase by 9%, and in a stunning plot twist that absolutely nobody saw coming, artificial intelligence is helping accelerate this trajectory. Not solving it. Not materially slowing it. Accelerating it. Yet somewhere in a Series B pitch meeting, a founder in a black turtleneck is still promising investors that their AI-powered prior authorization platform will "unlock unprecedented cost efficiencies across the healthcare ecosystem." The gap between what AI promised to do to healthcare costs and what it is actually doing has become so cavernous that you could fit an entire venture fund inside it.

For years, the Silicon Valley narrative around AI in healthcare has been almost religious in its fervor: machine learning will eliminate administrative waste, automate billing, reduce unnecessary procedures, and magically make care affordable. The theology was so pervasive that healthcare venture capitalists didn't even bother asking the foundational question: what if adding another layer of proprietary technology to the most complex, least efficient system in the developed world simply makes things more expensive? What if, hear me out, building AI tools that hospitals must license, integrate, maintain, and train staff on costs money? The PwC report doesn't just suggest this is happening—it quantifies it at a 9% annual increase. That's not a rounding error. That's not a transition cost. That's the new baseline.

This pattern is not new to the AI-healthcare complex. The industry has spent the better part of a decade funding companies promising to "disrupt" radiology, pathology, and drug discovery. Some have created genuine value. Many have created expensive databases. Yet the venture machine continues its eternal optimization: identify a problem, build an AI solution, raise Series A through Series D on the promise of solving it, achieve moderate adoption, eventually get acquired by a larger healthcare conglomerate at a modest multiple, and then slowly disappear into the Byzantine infrastructure that is American medicine. Rinse, repeat, collect management fees.

The language being deployed to defend these investments has become almost Orwellian in its disconnect from reality. "Efficiency" now means adding software costs. "Innovation" means replacing human bottlenecks with algorithmic ones. "Scalability" means spreading the expense across more hospitals faster. Investors speak of "unlocking value" in healthcare—which is a way of saying they're looking for new revenue streams in a system that's already cannibalizing itself. When PwC says costs are going up 9%, what they're not measuring is how many of those new dollars are flowing directly to AI vendors with zero evidence that their tools are solving the core problem.

The most damning evidence isn't in PwC's analysis—it's in what happens next. Healthcare venture funds will not slow down. If anything, they'll accelerate, because the 9% cost increase creates more perceived urgency, which creates more founder motivation, which creates more pitches about AI solutions. The feedback loop is self-sustaining. A rising tide of healthcare costs lifts all boats, especially the ones carrying software solutions looking for a problem. The system is broken, so we'll build more tools. The tools make it more expensive, so we'll need more tools. This is not disruption. This is a protection racket with a machine learning component.

What's truly remarkable is the immunity to contradiction. Every major AI-healthcare investment thesis of the last five years has faced the same headwind: healthcare costs continue climbing regardless. And yet the venture community treats this like weather—an external force, not a condemnation of their own thesis. PwC's 9% estimate isn't a wake-up call. It's just this quarter's number on which to build next quarter's funding round.