Biotech is drifting into a “fabless” model. The West still dominates late stage clinical execution and regulatory affairs, but a dangerous assumption has crept in upstream: that preclinical biology, process development, and early manufacturing are commodity services you can rent without losing the learning curve.
As biology industrialises, the loop becomes the asset. Data, automation, tacit know how, and the ability to run tight design make test learn cycles compound like an industrial base, not like a spreadsheet. Add AI and robotics and that compounding accelerates.
If you care about AI x bio, national resilience, or the next decade of drug innovation, you’re in the right place.
"Second, invest in the bio industrial commons. No startup should build a full stack lab from scratch. But we can build shared platforms that many startups can plug into. Biofoundries. Med chem cores. Assay centers. It is the physical infrastructure equivalent of what cloud did for software." 👏👏👏
Aren’t those choices rational - even inevitable - given the financing and governance stack?
Reading this, it struck me that many of the ‘reasonable’ decisions you list (asset-light models, renting capacity, outsourcing upstream biology) look less like misjudgements and more like a stable equilibrium under short fund lifetimes, exit pressures, and board incentives.
In that sense, outsourcing isn’t a mistake so much as a predictable outcome of a system that rewards short-term capital efficiency and systematically underprices long-horizon capability, resilience, and option value.
Which raises the harder question your piece points toward: if sovereignty and resilience are now strategic objectives, who in practice is meant to carry that long-dated risk? patient private capital, strategics, public underwriting, or some hybrid?
Totally agree. Corporate and venture funding is often driven by short term decision making, but the best roi in the long term comes from orthogonal high conviction bets. If you are brave and right you can cause a generational shift
If agentic AI and robotics truly democratize wet lab capability (as you suggest they will), doesn't that make vertical integration become unnecessary?
You can't simultaneously argue that
1. Automation will solve the labor cost problem and enable Western reshoring
2. Companies must vertically integrate to own the learning loop
If the loop can be reliably automated, then it can be distributed, and you’d maintain sovereignty through redundant, interoperable providers with low switching costs. I.e. if automation is good enough to eliminate China’s labor advantage, it’s also good enough to eliminate the need for vertical integration.
Your solution assumes biology remains artisanal enough that tacit knowledge requires physical proximity. I believe if automation actually delivers on its promise, the winning model looks more like AWS than Apple’s integrated hardware stack.
An important piece. There clearly is an ongoing attempt to decouple from Wuxi (even if it has been diluted of late). However I don’t know of any Western CDMO having a comparable breadth and scale-perhaps TMO comes closest. I also note the performance of the Bob Nelson backed “Resilience”.
One other point. As you point out there has been a stampede of Big Pharma players in-licensing Chinese assets and a few “newcos” have also been spun out.
The potential next phase in Chinas evolution (already happening) will be setting up shop in the West with a mandate to commercialise drugs for western population at a significant price discount.
"Second, invest in the bio industrial commons. No startup should build a full stack lab from scratch. But we can build shared platforms that many startups can plug into. Biofoundries. Med chem cores. Assay centers. It is the physical infrastructure equivalent of what cloud did for software." 👏👏👏
Aren’t those choices rational - even inevitable - given the financing and governance stack?
Reading this, it struck me that many of the ‘reasonable’ decisions you list (asset-light models, renting capacity, outsourcing upstream biology) look less like misjudgements and more like a stable equilibrium under short fund lifetimes, exit pressures, and board incentives.
In that sense, outsourcing isn’t a mistake so much as a predictable outcome of a system that rewards short-term capital efficiency and systematically underprices long-horizon capability, resilience, and option value.
Which raises the harder question your piece points toward: if sovereignty and resilience are now strategic objectives, who in practice is meant to carry that long-dated risk? patient private capital, strategics, public underwriting, or some hybrid?
Totally agree. Corporate and venture funding is often driven by short term decision making, but the best roi in the long term comes from orthogonal high conviction bets. If you are brave and right you can cause a generational shift
If agentic AI and robotics truly democratize wet lab capability (as you suggest they will), doesn't that make vertical integration become unnecessary?
You can't simultaneously argue that
1. Automation will solve the labor cost problem and enable Western reshoring
2. Companies must vertically integrate to own the learning loop
If the loop can be reliably automated, then it can be distributed, and you’d maintain sovereignty through redundant, interoperable providers with low switching costs. I.e. if automation is good enough to eliminate China’s labor advantage, it’s also good enough to eliminate the need for vertical integration.
Your solution assumes biology remains artisanal enough that tacit knowledge requires physical proximity. I believe if automation actually delivers on its promise, the winning model looks more like AWS than Apple’s integrated hardware stack.
You are assuming there is no cross over learning from automating one assay or dataset to discovery in another. Which is the core argument
An important piece. There clearly is an ongoing attempt to decouple from Wuxi (even if it has been diluted of late). However I don’t know of any Western CDMO having a comparable breadth and scale-perhaps TMO comes closest. I also note the performance of the Bob Nelson backed “Resilience”.
One other point. As you point out there has been a stampede of Big Pharma players in-licensing Chinese assets and a few “newcos” have also been spun out.
The potential next phase in Chinas evolution (already happening) will be setting up shop in the West with a mandate to commercialise drugs for western population at a significant price discount.