Finding a new candidate does not, by itself, produce a usable technology. Even cells that work cannot become a treatment if they cannot reach the place where they are needed.
What caught our attention in this issue was how the work that comes next is becoming an investment opportunity in its own right.
BoSS is not seeking to develop a new drug. It is developing materials and processes that allow cells to be stored at room temperature and recover their function when needed. If successful, these technologies could ease constraints shared by multiple cell-based products, rather than a single treatment. BoSS overview ↗
A similar development is emerging around scientific AI. After AI identifies an interesting sequence or target, someone must make it testable, measure it under consistent conditions and turn the findings into results that can support further development.
There are also plans to organize this work as a market, rather than leave each researcher to arrange it individually. IGoR, announced by ARPA-H this May, aims to connect AI-assisted experimental design, shared experimental protocols and a network of research facilities that perform the work. The system is still being built, but its official FAQ envisages participation by contract research organizations and experiments that include manual work, and explicitly provides for payment for services. IGoR plans ↗, Official FAQ ↗
Experiments and logistics have always mattered. The sign of change is the effort to rebuild this work as infrastructure shared across many research projects, rather than as a supporting task attached to one project or product.
It is too early to conclude that the spread of AI will make experimental service providers profitable. A business that performs requested experiments quickly and cheaply has a different revenue model and risk profile from one that turns the results into development assets and retains rights to them. Schrödinger, discussed in “If AI Can Solve It, What Do You Sell?” ↗, is pursuing the latter as well.
If the cost of proposing candidates falls and candidates alone become abundant, other parts of the research process will become the constraints. Can a candidate actually be made? Can its function be demonstrated? Can it be stored? Work that resolves these constraints could become relatively more valuable.
From this perspective, the companies that gain strength may not be those proposing the most candidates. They may be those that can take a candidate and develop it to the point where others can decide to pursue or adopt it.
Owning equipment, however, is not enough either. What performance will a company commit to delivering? On what terms will it hand over the results, and which rights will it retain? The design of the contract, too, will distinguish one business from another.
Discovery is not losing its value. What is coming into sharper focus is who takes responsibility for the work that follows. That is the signal we see in this issue’s news of investments and partnerships.