If you zoom out and map the trajectory of biotech over the past few decades, what emerges is a field defined by paradox. Scientific tools have never been more powerful. Our ability to interrogate biology — from single-cell sequencing to CRISPR editing, from AI-driven protein folding to synthetic biology — has seen exponential growth. And yet, the process of turning those insights into approved drugs remains stubbornly slow, inefficient, and unforgivingly expensive.
Clinical trials now cost upwards of $500,000 per patient — compared to just $10,000 in the early biotech era — with total average development costs exceeding $2 billion per successful approval. This isn’t just unsustainable, it’s a signal of something structurally broken. All the scientific acceleration in the world doesn’t help if the machinery that turns discovery into medicine stays locked in a framework designed for a very different time.
We’ve spent a lot of time thinking about this at Dellecod. Our work lives at the intersection of complex systems, biology, and software — so this friction between what should be possible and what’s actually happening is something we feel. There’s an opportunity here, not just to build more tools, but to rethink entire pathways. And the cracks in the system aren't subtle anymore.
The U.S. biotech ecosystem, for all its historic ingenuity, is going through a reset. Post-COVID exuberance has cooled. IPO windows closed. Seed-stage fundraising hit historic lows. Many public biotech companies are trading below the value of their cash — a pretty brutal signal from the market. Momentum has shifted.
At the same time, China has moved decisively to fill the void. Their regulatory infrastructure is now objectively faster and cheaper. While U.S. startups often wait months for an FDA green light, companies operating in China benefit from an “implied approval” model — if there’s no reply within 30 days, you start. Human trials can begin there 5 or 6 times faster than in the U.S., often at a fraction of the cost.
This change isn’t just about speed. It's about leverage. China is starting to out-execute. In some corners — gene editing, CAR-T constructs, certain immunotherapies — they are pulling ahead. That changes the competitive landscape at a foundational level. The U.S. used to be the undisputed center of biotech because it was the best place to invent and test. But if invention stays here while testing moves abroad, the flywheel starts to slip.
Compounding this is a regulatory system that hasn’t meaningfully evolved. Barring rare exceptions (like during the AIDS crisis), the FDA’s requirements — while well-intentioned — now contribute heavily to trial costs. These aren’t just bureaucratic drags. They create real-world barriers to testing new ideas in humans. And that biases the industry toward incrementalism and optimization, not invention.
This matters most in the fields we should be sprinting into — like aging. Right now, there are no FDA-approved drugs for aging itself. Drugs like GLP-1s have opened an exciting window — not just for obesity or metabolic syndromes, but potentially for age-related diseases — yet huge parts of the system aren’t ready. Medicare doesn’t even cover obesity drugs. And there’s still no agreed-upon way to define or quantify biological aging in a clinical trial.
What we need isn’t moonshot optimism. It’s practical reform. Introduce something like Orphan Drug Designation, but for aging or similarly neglected, chronic diseases. Make it easier and faster to run investigator-initiated trials. Track “cost per patient” in clinical trials as a core metric worth improving — not just a footnote in the burn rate.
And then there’s AI. The promise is massive. Five years from now, nearly every biotech company will be AI-native by default. Virtual screening, protein structure prediction, automated lab workflows — these tools will reshape the discovery phase. There’s a real chance that, for the first time in decades, E-Room’s Law — the idea that drug development is becoming exponentially harder and more expensive — could be reversed.
But so far, most AI efforts stop before humans are involved. They optimize molecules in silico, or in cells, but don’t break through the critical wall between preclinical promise and clinical success. Phase II trials — the ones that test whether something actually works in people — still fail at staggering rates. If AI wants to change biotech’s economics, it needs to engage with biology at this level of messiness and unpredictability.
We see this as a frontier calling for new modalities and new infrastructure — tools that compress the time from concept to clinic. Think platforms that combine AI, delivery systems, and sequencing — not as siloed tech stacks, but as integrated systems for building medicines from the ground up. If you look at the biggest wins in recent biotech history — mRNA vaccines, gene editing platforms — they were built on this kind of convergence.
The future here isn’t a return to past paradigms. It’s a shift — from Castle Biotech to Networked Biotech. From monolithic, one-shot programs to platforms that generate a pipeline of approaches. From layering new tools onto old processes, to rethinking the map entirely.
This restructuring won’t happen overnight. But we’re optimistic. Beneath today’s down cycle there’s something stirring — a quieter, more foundational reinvention. Yes, costs are too high, trials too slow, incentives misaligned. But these are solvable problems. What makes biotech extraordinary is its capacity to do the impossible — to grab hold of a deeply complex system and reshape it.
If we can apply that mentality not just to the science, but to how we bring science into the world, we think the next generation of breakthroughs won’t just be in molecules. They’ll be in the systems themselves.
Clinical trials now cost upwards of $500,000 per patient — compared to just $10,000 in the early biotech era — with total average development costs exceeding $2 billion per successful approval. This isn’t just unsustainable, it’s a signal of something structurally broken. All the scientific acceleration in the world doesn’t help if the machinery that turns discovery into medicine stays locked in a framework designed for a very different time.
We’ve spent a lot of time thinking about this at Dellecod. Our work lives at the intersection of complex systems, biology, and software — so this friction between what should be possible and what’s actually happening is something we feel. There’s an opportunity here, not just to build more tools, but to rethink entire pathways. And the cracks in the system aren't subtle anymore.
The U.S. biotech ecosystem, for all its historic ingenuity, is going through a reset. Post-COVID exuberance has cooled. IPO windows closed. Seed-stage fundraising hit historic lows. Many public biotech companies are trading below the value of their cash — a pretty brutal signal from the market. Momentum has shifted.
At the same time, China has moved decisively to fill the void. Their regulatory infrastructure is now objectively faster and cheaper. While U.S. startups often wait months for an FDA green light, companies operating in China benefit from an “implied approval” model — if there’s no reply within 30 days, you start. Human trials can begin there 5 or 6 times faster than in the U.S., often at a fraction of the cost.
This change isn’t just about speed. It's about leverage. China is starting to out-execute. In some corners — gene editing, CAR-T constructs, certain immunotherapies — they are pulling ahead. That changes the competitive landscape at a foundational level. The U.S. used to be the undisputed center of biotech because it was the best place to invent and test. But if invention stays here while testing moves abroad, the flywheel starts to slip.
Compounding this is a regulatory system that hasn’t meaningfully evolved. Barring rare exceptions (like during the AIDS crisis), the FDA’s requirements — while well-intentioned — now contribute heavily to trial costs. These aren’t just bureaucratic drags. They create real-world barriers to testing new ideas in humans. And that biases the industry toward incrementalism and optimization, not invention.
This matters most in the fields we should be sprinting into — like aging. Right now, there are no FDA-approved drugs for aging itself. Drugs like GLP-1s have opened an exciting window — not just for obesity or metabolic syndromes, but potentially for age-related diseases — yet huge parts of the system aren’t ready. Medicare doesn’t even cover obesity drugs. And there’s still no agreed-upon way to define or quantify biological aging in a clinical trial.
What we need isn’t moonshot optimism. It’s practical reform. Introduce something like Orphan Drug Designation, but for aging or similarly neglected, chronic diseases. Make it easier and faster to run investigator-initiated trials. Track “cost per patient” in clinical trials as a core metric worth improving — not just a footnote in the burn rate.
And then there’s AI. The promise is massive. Five years from now, nearly every biotech company will be AI-native by default. Virtual screening, protein structure prediction, automated lab workflows — these tools will reshape the discovery phase. There’s a real chance that, for the first time in decades, E-Room’s Law — the idea that drug development is becoming exponentially harder and more expensive — could be reversed.
But so far, most AI efforts stop before humans are involved. They optimize molecules in silico, or in cells, but don’t break through the critical wall between preclinical promise and clinical success. Phase II trials — the ones that test whether something actually works in people — still fail at staggering rates. If AI wants to change biotech’s economics, it needs to engage with biology at this level of messiness and unpredictability.
We see this as a frontier calling for new modalities and new infrastructure — tools that compress the time from concept to clinic. Think platforms that combine AI, delivery systems, and sequencing — not as siloed tech stacks, but as integrated systems for building medicines from the ground up. If you look at the biggest wins in recent biotech history — mRNA vaccines, gene editing platforms — they were built on this kind of convergence.
The future here isn’t a return to past paradigms. It’s a shift — from Castle Biotech to Networked Biotech. From monolithic, one-shot programs to platforms that generate a pipeline of approaches. From layering new tools onto old processes, to rethinking the map entirely.
This restructuring won’t happen overnight. But we’re optimistic. Beneath today’s down cycle there’s something stirring — a quieter, more foundational reinvention. Yes, costs are too high, trials too slow, incentives misaligned. But these are solvable problems. What makes biotech extraordinary is its capacity to do the impossible — to grab hold of a deeply complex system and reshape it.
If we can apply that mentality not just to the science, but to how we bring science into the world, we think the next generation of breakthroughs won’t just be in molecules. They’ll be in the systems themselves.