The Open Discovery Corporation

The missing economics of fundamental discovery

Modern economies are extraordinarily good at exploiting possibilities they can already see. Once a technological opportunity becomes visible, companies can assess its value and mobilise capital and talent to pursue it. The prospect of economic return creates a gradient that pulls resources toward the opportunity. This is where markets are strongest.

Fundamental science operates before such a gradient exists. It does not merely pursue known opportunities; it creates altogether new ones. A theorem, physical principle or biological mechanism can open new scientific and technological domains, yet much of the economic value it ultimately enables only emerges later. The prospect of return is therefore too uncertain and distant to create a comparable gradient drawing resources into the search itself.

Government and philanthropy are society's principal solution to this problem. They fund work without seeking foreseeable commercial return. Because this funding substitutes for the missing market return rather than restoring it, there is little economic pressure for discovery capability to accumulate in a persistent organisation. Grants end, teams dissolve, and much of the expertise built in one project disperses. Science repeatedly pays to create discovery capability without building an institution to retain and compound it.

Scientific success, however, accumulates much more readily around the scientist than the institution. Success builds reputation, attracting funding and collaborators and increasing the scientist's capacity to do better science. The successful academic resembles a company of size one: an enterprise whose accumulated advantage remains attached to the individual.

By contrast, one of the corporation's great achievements is that productive capability can outlast the people who created it. Someone who joins a company inherits capabilities they did not themselves create. Because that accumulated capability makes the company more valuable, the company has an incentive to strengthen it further, while success provides the resources to do so.

Fundamental science remains surprisingly pre-corporate. Discovery capability remains attached largely to individual scientists rather than accumulating in enduring organisations. What is missing is an organisation in which discovery capability itself can accumulate and improve, while the discoveries remain open to all.

A new institutional category

Private enterprise and science have long sat on opposite sides of a familiar institutional divide. Companies are organised around private value; universities and non-profit institutes around public benefit. An open discovery corporation, or ODC for short, deliberately crosses that boundary. It is a for-profit institution devoted to fundamental discovery, with the knowledge it produces remaining open to everyone. Its defining principle is simple: private ownership of the institution need not imply private ownership of its discoveries.

Those who create and build the company can hold meaningful ownership and share in the value they help generate, while its discoveries remain published and freely usable. The constraint is on the privatisation of fundamental knowledge, not on the creation of private value. The crucial distinction is between knowledge and discovery capability. The knowledge can remain open while the capability to produce it remains scarce and valuable.

The corporate form gives this capability economic force. As better discovery capability generates revenue, the resulting resources can be reinvested to strengthen it further. The capability to discover, rather than the discoveries themselves, becomes the accumulating asset of the company.

That opportunity can already be pursued within existing corporate law. In the UK, an ordinary company can make fundamental discovery and open dissemination an enduring purpose rather than leaving them contingent on short-term shareholder pressure. No bespoke corporate form is needed to begin.

Discovery capability

Discovery capability has three components: direction, identifying the discoveries most worth pursuing; mobilisation, bringing the required capabilities to bear on them; and understanding, seeing how discoveries connect and what they make possible. Each can become a capability of the institution rather than remaining attached to individual scientists.

First, science does not progress simply by making discoveries, but by making enabling discoveries—discoveries that expand our capacity for further discovery. Brian Arthur describes technological progress as a recursive process in which technologies become building blocks for later technologies. Science progresses in much the same way: discoveries become building blocks for further discovery. Progress therefore depends not only on what is discovered, but on what those discoveries make possible next.

This makes choosing what to work on a central scientific problem. Science is highly developed at solving problems, but much less so at deciding which problems to solve. A scientist's research trajectory typically proceeds by local moves through idea space, rather than by systematic comparison across it. An ODC, by contrast, could ask which problems, if solved, would most expand the scope for further discovery, whether by removing bottlenecks, exploiting newly tractable problems, or uncovering connections between fields. AI makes this task more pressing by lowering the cost of probing possible directions, making the choice of where to concentrate effort even more consequential.

Second, established companies are geared to exploit existing products; exploring the products of the future is harder. Science has the opposite problem. Curiosity-driven research provides a powerful decentralised system for exploration. But when it reveals a promising opportunity, science is slow to concentrate effort around it, and often does so below the required scale. Because academia cannot readily redirect researchers toward new priorities, development proceeds mainly by raising grants and recruiting students and postdocs. Academic science tends to organise problems around people, rather than people around problems.

Because scientists join an ODC to participate in shared scientific missions, the organisation can assemble and reconfigure the capabilities a problem requires as the science develops. Such fluid collaboration may require a more collective model of scientific credit: ODC scientists would likely share credit for the body of science they produce, with papers carrying the institutional rather than individual names. Scientific reputation would shift from individual researchers and papers towards the group and its cumulative body of work. Ideas could then move between scientists, draw in new expertise and grow beyond the capabilities of their originators.

Third, knowledge consists of results; understanding lies in the relations between them. We may believe a theorem to be true without yet seeing its dependencies, analogues or consequences. Scientists develop detailed maps of their own specialised areas, but these remain local and dispersed across the scientific community. AI can now produce new results faster than their relationships can be understood, shifting the bottleneck from producing results to understanding them. Those relations matter because they reveal how results can become building blocks for further discovery.

An ODC can make understanding an institutional capability by building a relational map of science. Its scientists contribute deep local knowledge; the institution connects it across projects and fields, identifying dependencies, analogues and consequences beyond what any individual scientist could hold in mind. The map is not itself a discovery, but part of the institution's accumulated capability for making discoveries. Its global structure could reveal bottlenecks, newly tractable problems and connections between fields. Each new result would in turn refine the map, strengthening the institution's ability to see what can be discovered next.

Funding and ownership

An ODC can combine two funding models normally kept apart: the earned revenue of a company and the grants and gifts of a non-profit.

An ODC can earn revenue without restricting access to its discoveries. Companies and research organisations can commission it to identify problems worth pursuing or to solve specific problems themselves. What they pay for is the application of scarce discovery capabilities to problems they care about. Customers are paying to make a discovery effort happen, not to own the knowledge it produces.

An ODC can also earn revenue from the tools and infrastructure it develops for scientific discovery. AI systems, software, automated workflows and experimental platforms can remain scarce and valuable even when the discoveries they help produce are open.

Others can pay to use these capabilities without acquiring ownership of the resulting knowledge. In both cases, the ODC creates private value through its capability to discover rather than by restricting access to its discoveries.

An ODC already has access to a substantial range of research grant funding. In Europe, companies can host grants from Horizon Europe and the European Research Council. In the US, DARPA and the military research agencies regularly fund research carried out by companies, as does the UK's ARIA. By contrast, a for-profit ODC cannot currently lead or co-lead grants from the UK Research Councils, and Wellcome generally requires a non-profit lead.

Philanthropy can support an ODC for the same reason it supports a university or non-profit: open research delivers a public benefit. Individuals can give directly, while charitable foundations can fund companies whose work furthers their charitable purposes. An ODC would forgo some of the tax advantages of charitable status, but these matter less for an organisation operating globally: much large-scale philanthropy comes from foundations and overseas donors, to which UK Gift Aid does not apply.

For an R&D-intensive ODC, these forgone advantages can be outweighed by the R&D tax incentives available to a company. In the UK, the standard scheme provides a taxable credit equal to 20% of R&D expenditure, rising to a tax-free 27% if the ODC is loss-making.

Together, revenues, grants, donations and R&D relief can substantially reduce the need for outside equity to finance research and growth. Venture capital becomes a choice rather than a requirement, available where it is useful rather than necessary to sustain the company. This in turn allows ownership to remain concentrated among the people who create and build the company and institutions committed to its scientific mission. Limiting outside financial ownership also strengthens the case for philanthropic support: donors can fund open fundamental research without subsidising returns to external investors.

LIMS and Open Discovery: one mission, two models

I founded the London Institute for Mathematical Sciences as an independent institute devoted to fundamental discovery. Experience at LIMS shows the value of giving exceptional scientists unusual freedom and support. But it also suggests a more radical possibility: that an institution might itself accumulate discovery capability. I have now founded Open Discovery Corporation, a UK company, to put this model into practice. LIMS optimises the environment for exceptional scientists. Open Discovery asks whether an institution can compound its discovery capability.

LIMS is built around exceptional scientists and their scientific judgement, giving them unusual freedom to pursue the problems they consider most important. The institutional bargain is substantial scientific autonomy, but without the competitive pay or ownership upside available at a company. LIMS concentrates its institutional effort on creating the conditions for outstanding science rather than directing what scientists work on.

Open Discovery offers a different bargain: greater collective scientific power, competitive pay and meaningful ownership, in return for participating in more coordinated scientific missions. Working at Open Discovery means contributing not only to discoveries but to the discovery capability accumulated by the organisation. The first people at Open Discovery will help shape not only what it discovers, but how the institution itself learns to discover.

Some people will prefer the greater individual autonomy of LIMS, others the collective ambition and ownership of Open Discovery, and some may work across both. The scientific boundary between them should remain highly porous. They share a building and much of their institutional life: talks, visitors, collaborations and communication. Discoveries from either side can readily inform work on the other. While their legal and financial boundaries remain clear, their proximity creates a positive feedback loop. LIMS brings frontier problems, scientific judgement, leading collaborators and scientific credibility. Open Discovery brings the ability to identify promising directions, mobilise effort around them and build understanding across projects and fields.

LIMS can also be an anchor client for Open Discovery, commissioning research where a more coordinated approach may outperform additional in-house capacity. For the cost of another postdoc, does LIMS get more scientific progress by adding that capacity in-house or by commissioning Open Discovery instead? If Open Discovery repeatedly delivers more for the same resources, that is evidence that its model works.

Open Discovery would be owned principally by the people who create and build it, with LIMS Ventures holding a minority stake reflecting its role in Open Discovery's incubation.

If the experiment works

LIMS and Open Discovery begin as distinct institutional models. LIMS bets on exceptional scientists given exceptional conditions. Open Discovery tests whether discovery capability can accumulate in the organisation itself. The contrast between these models is the experiment.

If Open Discovery develops sufficiently powerful discovery capability, the distinction between the two models may diminish as their strengths are combined. What begins as two neighbouring organisations could eventually converge on a common institution. More importantly, the experiment would show that discovery capability can accumulate institutionally without making scientific knowledge private.

That experiment begins within existing legal structures, but those structures need not determine its eventual form. Companies and charities are themselves institutional inventions, and the rules around public funding, tax, ownership and governance can evolve as new arrangements prove their value. The experiment can precede the institutional form it may ultimately require.

Dr Thomas Fink is director of the London Institute for Mathematical Sciences.