The infrastructure advantage: build the shovel, not just the goldmine

Founders: I'm writing to advocate that you look at your technology through a new lens. Don't just ask, "Does my product solve today's problem?" Instead, ask, "Does my technology lower the barrier for someone solving tomorrow's problem?" Let me explain.

‍

The value of infrastructure

We're all very familiar with the AI boom, bubble… whatever we're calling it these days. All sorts of VC money is pouring into AI companies. OpenAI and Anthropic alone accounted for 43% of all startup funding in the first half of 2026, illustrating just how concentrated investment has become. However, the heroes of this market have been companies providing the infrastructure that supports the AI boom. NVIDIA is a great example because it sells a capability (on-chip memory capacity) that thousands of AI companies build on. That's the unique power and value of infrastructure; NVIDIA is selling shovels while everyone else fixates on digging for gold.

Infrastructure lowers costs for innovators. It can reduce uncertainty in the development process. It allows for specialization, in a particular capability for example, instead of forcing companies to spend time and resources reinventing the wheel. Infrastructure creates common standards and standard practices. This does not sound very appealing from a glitz-and-glamor perspective, but it is critical to building the backbone of an ecosystem. Lastly, infrastructure enables feedback loops to propagate, enabling iterations of innovation heretofore unimaginable.

‍

Infrastructure in biotechnology

Before I dive into ways to recognize what this technology looks like, I'd like to use an example in order to illustrate a new way of thinking about infrastructure in biotech. For decades, the most common path to success in the sector has been blockbuster drugs. Just look at what GLP-1s have been doing recently: sales through the roof for a drug initially intended for Type 2 diabetes that is now the highest-grossing weight-management medication on the market. But this model has proven unsustainable for an entire industry, especially given the recent shift in biotechnology investment. There will always be Big Pharma and its standard approach to developing the next therapeutic. Still, when it comes to biotechnology innovation, VCs are increasingly hesitant to invest because breakthroughs aren't predictable by an algorithm. And a pipeline that takes 10 years to realize is a very expensive way to lose money if the treatment doesn't pan out.

This brings me to a new flavor of biotechnology that is emerging: tools for building the next generation of therapeutics. And specifically, organoids. These small clusters of human tissue cells may represent the infrastructure upon which biotechnology will build tomorrow's solutions. Organoids, although commonly misunderstood as "mini organs," are really just cells representing different functions found in those organs, coupled with enormous predictive power by design. The scientists building these models are thinking about the questions other scientists are trying to solve and engineering solutions that enable them to test those questions. It's not about building the model that most accurately represents a human liver or colon. It's about generating an organoid that models the specific processes occurring in those organs and enables researchers to test different ideas within those functional environments.

Okay, without diving any further down the scientific rabbit hole, let's test whether organoids have the potential to become infrastructure for the next generation of biotech. One hallmark of infrastructure is that the product has many downstream users. In the case of organoids, this checks out. There are many ways to design models that achieve unique predictive capabilities, for example, within the human gastrointestinal tract. But more broadly, anyone working with therapeutics or supplements needs to know how their product interacts with the human gut.

Building a model, an intestinal organoid, for example, also lays the groundwork for standardized outputs. Critical infrastructure enables standardization so that every user knows how to use the tool and achieve comparable results. Then there's the potential for use across industries. Within biotechnology, there are many different applications and approaches for developing chemical and biological solutions to disease. Organoid models have the potential to unlock testing capacity across all those therapeutic areas. Which proves the next point about infrastructure: it becomes more valuable as adoption grows. As it lowers the barrier to testing and innovation, it becomes increasingly difficult to replace, entrenching itself within the system.

‍

Seeing infrastructure all around you

You might take it for granted, but standardization, reproducibility, and scalability are core tenets of the technology underpinning our biggest breakthroughs. So let's address some other assumptions, such as the idea that therapeutics are the real drivers of value.

They are. Sometimes.

Most of the time, though, they're moonshots. Platform technologies, by contrast, build entire ecosystems. That's a kind of longevity a blockbuster drug doesn't have. Another assumption is that research tools are an expense. They are… for someone. But for you, why not make the tool everyone has to buy to do their work? Then you're making money, not shelling it out. And finally, there's the overarching assumption that success comes from selling that one product: the unicorn. It certainly can. But success can also come from enabling the production of thousands of products.

My argument here is simple: let's not get caught up in the frenzy of building the next big product. Instead, consider all the value that could be created by a tool that enables an entire generation of breakthroughs. It's just the kind of 'outside-the-box' thinking that we're going to need in biotechnology to pull the sector out of the current nosedive and right our course towards a sustainable future.

‍