Mantis Biotech’s Digital Twins Are Solving the Wrong Problem

Here’s what nobody wants to admit: Mantis Biotech’s digital twin technology isn’t actually solving medicine’s data availability problem. It’s creating an expensive workaround for a problem we refuse to fix properly.

The company just announced they’re building digital replicas of human biology to simulate drug responses and disease progression. TechCrunch covered it like it’s the next big thing in healthcare AI. And sure, the tech is impressive. But we’re essentially admitting defeat on the real issue: we can’t get access to actual patient data, so we’re going to fake it with fancy simulations instead.

The Real Data Problem Nobody Talks About

Medicine’s data availability problem isn’t technical. It’s bureaucratic, legal, and political. We have mountains of patient data locked away in hospital systems that don’t talk to each other, trapped behind privacy regulations that were written before anyone imagined what AI could do with medical information, and hoarded by institutions that treat data like proprietary gold.

Digital twins are a Band-Aid on a bullet wound. Instead of fixing the broken data-sharing infrastructure, we’re building elaborate computer models that approximate what real patients might do. It’s like being too lazy to check the weather outside, so you build a supercomputer to simulate atmospheric conditions in your backyard.

Why This Matters More Than You Think

Don’t get me wrong—the technology itself is legitimately cool. Creating computational models of human biology that can predict drug interactions and disease outcomes has real value. But positioning this as the solution to data availability is misleading at best, dangerous at worst.

Simulations are only as good as the data they’re trained on. If we can’t access diverse, real-world patient data to build these digital twins, we’re just creating really sophisticated guesses. And in medicine, sophisticated guesses can kill people.

The AI healthcare space is already dealing with this problem. Recent coverage shows companies trying to use AI to address labor shortages in rare disease treatment. That’s another workaround for systemic issues—not enough specialists, not enough funding for rare disease research, not enough incentive for pharma companies to develop treatments for small patient populations.

What Actually Needs to Happen

We need data infrastructure reform. Real reform. That means interoperable electronic health records, clear frameworks for anonymized data sharing, and incentives for institutions to contribute to shared medical databases. It means updating privacy laws to distinguish between exploitative data harvesting and legitimate medical research.

Companies like Nephrogen are combining AI with gene therapy to tackle kidney disease. That’s the kind of direct approach that makes sense—using AI as a tool within the actual treatment pipeline, not as a replacement for missing data.

Digital twins could be incredibly useful as a complement to real patient data. Run simulations to generate hypotheses, then validate them against actual outcomes. Use them to explore edge cases and rare scenarios where you genuinely don’t have enough real-world examples. But treating them as a substitute for the hard work of fixing data access? That’s just kicking the can down the road.

The Uncomfortable Truth

Mantis Biotech is doing what any smart company would do—finding a market opportunity in a broken system. They’re not the villains here. But we shouldn’t pretend this is solving the fundamental problem. It’s solving a symptom.

The medical establishment loves technological solutions because they’re easier than institutional change. Building a digital twin platform is straightforward compared to getting fifty different hospital systems to agree on data standards, or convincing regulators to modernize privacy frameworks, or creating legal protections for data sharing.

But easy isn’t always right. And in healthcare, taking shortcuts with data can have consequences that ripple through entire populations. We’ve seen what happens when AI systems are trained on biased or incomplete datasets—they perpetuate and amplify existing disparities.

So yes, digital twins are impressive technology. Yes, they’ll probably help advance medical research in specific ways. But let’s stop pretending they’re the answer to medicine’s data problem. They’re a workaround. And the sooner we acknowledge that, the sooner we can start having honest conversations about what it would actually take to fix the underlying issues.

The data is out there. We’re just too scared, too lazy, or too invested in the status quo to go get it.

🕒 Published: March 30, 2026