An innovation commons is a shared institution in which research and development capability is pooled and governed by the people who use it. The pooled resource is capability: laboratory space, equipment, staff time, access to capital, and the attention of people working in adjacent fields. Members hold rights to the resource and to the rules that allocate it.
The term adapts Elinor Ostrom's research on common-pool resources to the problem of where new work gets made.
Where the idea comes from
Elinor Ostrom spent her career studying irrigation systems in Spain and Nepal, inshore fisheries, and alpine grazing pastures in Switzerland. The received wisdom held that any resource held in common would be destroyed by the people who used it, and that the only remedies were private ownership or state control. Ostrom went and looked. She found systems that their users had governed successfully for four and five hundred years.
Governing the Commons came out in 1990. In 2009 she became the first woman to win the Nobel Memorial Prize in Economic Sciences. Her contribution was empirical rather than theoretical: she catalogued what the durable systems had in common, and produced eight design principles. Boundaries are clear about who is in and what the resource is. Rules fit local conditions. The people affected by the rules can change them. Monitoring is done by people accountable to the users. Sanctions escalate rather than starting at expulsion. Conflict resolution is cheap and close at hand. The right to organize is recognized by outside authority. Larger systems are built as nested layers rather than one central body.
What changes when the resource is capability
An aquifer can be drained. Research capability cannot, at least not in the same way. Two people using a laboratory does not leave less laboratory for a third, up to the point of congestion. So the governance problem changes shape.
The failure mode for a shared physical resource is depletion. The failure mode for shared innovation capability is capture. Whoever brings the most capital sets the research agenda, and whoever holds the legal claim takes the output. This is a well documented pattern rather than a hypothetical one, and it is the reason the governance layer has to be built deliberately instead of assumed.
The twentieth-century laboratories
Three institutions are cited constantly in this conversation, and it is worth being precise about what they show.
Bell Labs produced the transistor, the laser, and Claude Shannon's information theory. Its structural advantage was that theorists and experimentalists shared corridors and canteens, and that AT&T's regulated monopoly gave it patient money and no quarterly pressure. Xerox PARC produced the graphical user interface, Ethernet, and much of what personal computing became. The MIT Media Lab made antidisciplinary work an explicit structure rather than an accident, funded through a corporate consortium.
All three worked for the same underlying reason. They put people who would otherwise never have met in the same building and gave them long enough to get somewhere. And in all three the funder held the claim on what came out. The Xerox case is the famous one, where the commercial value of the personal computing work was realized largely by other companies. The less discussed version is that the public had no standing in any of it either way.
An innovation commons asks a narrow question about this history: can you keep the co-location and the patient horizon while moving the claim on the output to the community that produced it?
How to tell one from a coworking space
Two tests do most of the work.
Can members change the rules? A coworking space sells access at a price the landlord sets, and members have no standing in admissions, pricing, or what the space is for. A commons gives members defined rights in how the resource is allocated and who is admitted. Ostrom's third principle is the load-bearing one here.
Where does the output go? A coworking space has no claim on what its members produce, which sounds generous and means there is no mechanism to keep any of it public either. A commons defines in advance which share of what gets made stays in the commons.
An accelerator fails both tests differently. It selects companies, invests on standard terms, and exits at acquisition or IPO. Its horizon is its fund life, usually seven to ten years, which is a reasonable horizon for software and a poor one for materials science, longitudinal social research, or anything whose result is a public good rather than a company.
What makes the model hard
Three problems recur, and any honest account of the model should name them.
The pooled capability has to be substantial enough to be worth governing. If the shared resource is a room and a printer, the governance apparatus is overhead on a social club. The model only justifies its cost when the resource is expensive enough that no member could hold it alone.
Governance consumes member time, which is the scarcest thing members have. Ostrom's systems worked partly because participation was tied to a livelihood that depended on getting it right. A commons whose members have day jobs elsewhere has to design for much lower participation than its charter usually assumes.
And mission lock has to outlive the founders. Intent expressed in a founding document is worth very little in year fifteen. The mechanisms that actually bind are legal: a purpose trust that owns the mission independently of the operators, a mission-use restriction recorded against the property so it runs with the land, and revenue splits written into operating agreements rather than promised in a manifesto.
A worked example
Apollo Innovation Commons is building this model in San Francisco, opening in 2026. It runs eight lab clusters across AI, biotechnology and consciousness technology, human flourishing, climate, art and culture, civics and governance, robotics and hardware, and metatheory and moral inquiry, alongside education, incubation, design and capital deployment under one governance structure. It operates under fiscal sponsorship from the Flourishing Systems Foundation, a 501(c)(3), and holds its mission through a purpose trust and a recorded mission-use restriction.
Read the full description of how Apollo is structured, or read the white paper, which sets out the reasoning at length.
Common questions
What is an innovation commons?
A shared institution in which research and development capability is pooled and governed by the people who use it. The pooled resource is capability: laboratory space, equipment, staff time, access to capital, and the attention of people working in adjacent fields. Members hold rights to the resource and to the rules that allocate it.
How is it different from an incubator or accelerator?
An accelerator selects companies, invests on standard terms, and exits when they are acquired or go public. Its horizon is its fund life, usually seven to ten years. An innovation commons has no exit. It holds capability in place indefinitely and can support work whose payoff is a decade out, or whose result is something other than a company.
How is it different from a coworking space?
Two tests. Can members change the rules? A coworking space sells access on terms the landlord sets; a commons gives members standing in allocation and admissions. And where does the output go? A coworking space has no claim on what members make and no mechanism to keep any of it public.
Who owns what an innovation commons produces?
That depends on the legal structure, and it is the question the model exists to answer. Bell Labs, Xerox PARC and the MIT Media Lab each produced remarkable work by co-locating people who would otherwise not have met, and in each case the funder held the claim on the output. A commons uses a purpose trust, a recorded mission-use restriction, and revenue splits written into its operating agreements to keep a defined share in the commons.
Is an innovation commons a nonprofit?
Usually it is a combination. Discovery work that no investor will fund sits under a 501(c)(3) or a fiscal sponsor. Operations and revenue-generating labs sit in an LLC. The mission lock holds the two together, because it survives changes in who runs the operating entity.
How many innovation commons exist?
Few, and the term is not yet standardized. Related structures include community land trusts applied to research property, cooperative research organizations, and fiscally sponsored civic labs. The distinguishing features to look for are member governance rights and a legally binding claim on where the output goes.
Last reviewed 31 August 2026. Maintained by Apollo Innovation Commons, San Francisco.