In every wave of technological disruption, we see a familiar pattern emerge. 

First, there are the infrastructure players, offering the “picks and shovels” of the new age—think OpenAI, Anthropic, or Google’s DeepMind. They’re gunning to build the foundational AI models everyone else will rely on. 

Critics might say these companies face fierce competition and risk quick commoditization. The recent DeepSeek announcement might be an indication that it’s already beginning, but even as the underlying models become more standardized, their early presence shapes the entire playing field, giving them leverage in how ecosystems develop.

Next, we have the application builders: the early adopters who integrate these AI models into existing business models. In healthcare, for example, we see many startups using AI to streamline parts of existing practices, such as: 

• Identifying health risks like collapsed lungs in imaging more quickly than human radiologists alone

• Helping doctors assess symptoms to improve diagnosis accuracy

• Transcribing conversations between clinicians and patients or to doing clinical documentation

While certainly valuable advances that could improve the practice of medicine and patient outcomes, these are akin to the earliest factories that replaced steam with electricity but changed little else. 

Skeptics might argue this incremental approach is both safer and more pragmatic—quick gains, measurable ROI, no wild future bets. And that’s valid. Not every company needs to reinvent its world. Yet, these plays often overlook the tech’s fundamental properties, treating AI like a feature upgrade rather than a catalyst for total reinvention.

History suggests that the big prizes don’t go to those who merely bolt on a new capability. 

In the early days of the internet, many tried to replicate offline models online, delivering marginal gains but failing to see the deeper shifts—like Pets.com simply moving a pet store onto the web. Meanwhile, Amazon and Google recognized the profound second-order effects: infinite shelf space, near-zero distribution costs, and global reach. 

If you look at AI today, many investors and founders question if it’s too soon to go “all-in” on rethinking everything. After all, regulations in fields like healthcare are complex, and the infrastructure itself is still evolving. But that hesitancy can be a missed opportunity. By the time the landscape settles, those who’ve anticipated the real shifts—those who’ve designed their companies to exploit AI’s unique economics—will have a significant head start.

This brings us to the second-order innovators—the companies willing to rebuild entire workflows and value chains around AI’s inherent advantages. Yes, it’s riskier. It demands vision well beyond what’s comfortable. And yes, some argue that the complexities of regulation, logistics, and customer adoption will slow radical change. But it’s precisely that friction that creates defensible opportunity. 

Consider clinical trials. Traditionally, exploring new drug candidates is expensive and time-consuming because each hypothesis requires substantial investments in lab work, animal studies, and eventually human trials. 

AI radically lowers this “cost of inquiry” by using computational models to screen vast numbers of compounds quickly and at minimal incremental cost. Instead of relying solely on the slow, linear approach of testing one candidate at a time, researchers can now run thousands of simulations in parallel—essentially shifting from sparse, guesswork-driven trials to dense, data-rich exploration.

This transformation becomes even more powerful when combined with digital twin technologies. Imagine a digital twin of a patient population or even an individual patient—an AI-driven, virtual model that captures countless physiological variables and responds dynamically to simulated treatments. 

With digital twins, researchers can test myriad drug candidates, dosages, and treatment protocols virtually before ever administering a single dose to a human. This approach not only accelerates discovery but also helps identify the most promising candidates earlier, reducing wasted resources on dead-end paths.

Yet, simply adding AI or digital twins on top of the existing clinical trial framework only tweaks the edges of a much larger opportunity. 

The real leap forward requires rethinking everything: logistics for rapidly manufacturing and distributing compounds, recruitment strategies for bringing the right patients into highly targeted trials, and seamless data aggregation that enables continuous feedback loops from both virtual and real-world outcomes. 

Soon, compounds will flow into trials at unprecedented speed and scale, and trial designs will evolve from static protocols to dynamic, adaptive systems. The winners will be those who prepare now for a reality where today’s bottlenecks—time, cost, patient availability—are no longer the defining constraints. 

Instead, success will hinge on the ability to orchestrate this explosive growth in inquiries and insights, leveraging AI and digital twins to navigate complexity and unlock entirely new models of innovation. 

In the end, the real differentiator isn’t just the technology—it’s the courage to envision and build for a future that looks nothing like today. 

At Novy, we’re not simply hunting for companies that can slot AI into existing processes. We’re scouting for those rare ventures that see second-order potential. We believe the biggest, most enduring wins will come from founders who don’t just optimize the old ways, but rewrite the rules entirely. Our investment thesis revolves around building companies who aren’t afraid to shatter assumptions and re-architect value chains from the ground up.

History shows that the lion’s share of value accrues to those who embrace the full magnitude of a disruption. Incremental improvements deliver incremental returns. Transformational change creates category-defining, ecosystem-shaping outcomes. 

By backing companies that aim higher, Novy positions itself at the frontier of lasting innovation—supporting entrepreneurs who are prepared to harness AI’s true capabilities and, in doing so, capture the world-changing opportunities that lie just beyond the bolt-on.