By Daniel Goetzel
In previous posts, we introduce the major opportunities coming from huge investments by tech companies into AI infrastructure and offer a playbook for local and regional policymakers when negotiating with OpenAI or other big tech firms. Here we discuss major ongoing and upcoming AI investments and relevant innovative public-private partnership models that have recently launched in the US and around the world.
The original vision of OpenAI, the non-profit, was to create a future world where AI was a force multiplier for good and shared prosperity. Even as they began raising capital from investors, they pitched a “capped profit” structure, where investors agreed to a maximum return on their investment, with any profits exceeding that cap going back to the non-profit entity. That same capped profit structure can be applied to these place based economic development projects, where the value created is shared with the community as opposed to exclusively shipping the wealth, intellectual property, and talent they create back to Silicon Valley. Emerson Collective and Urban Institute have proposed a compelling model for how you could structure real estate deals so community members can participate in the projects’ upside called
Some of this work shaping future AI Hubs is already underway, happening organically. This gives me hope that this new model is executable with a strong business case behind it for all parties.
New Jersey: A by NJ, Microsoft, Coreweave, and Princeton to create an AI hub via R&D, startup support, incubation space, and the into a datacenter for Coreweave.
Virginia: A public-private partnership with , a seven year old startup spun out of MIT, for a grid-scale fusion plant, with only $2 M state/county incentive contributions. However, the state did broker partnerships with local municipalities and the state’s largest utility, Dominion Energy, to bring this project to life. While it won’t come online until the early 2030s, it is a forward looking investment to diversify energy sources to meet AI’s unrelenting energy appetite.
Illinois: The state announced a , a leading quantum computing company to anchor its new mega campus at the old site of US Steel. PsiQuantum commits to sharing space in their new Quantum Computer Operations Center with university researchers and academics, with the state covering the cost of shared cryogenic facilities, equipment labs, and research spaces. This innovative model attracted DARPA, IBM, University of Illinois and other private companies to the campus, creating an emerging cluster. The state and PsiQuantum are making joint workforce and neighborhood level investments to uplift the historically underserved community surrounding the campus (full disclosure: I am advising P33, who is a partner on this project on a strategy to develop more energy efficient data centers).
Ireland: Ireland is one of the fastest growing data center hubs in Europe. The town of Tallaght in South Dublin to re-use the excess heat from the data center to heat homes and buildings in the community, drawing on co-generation models pioneered on university campuses.
Takeaways
There is a perfect storm brewing where the capital expenditures on data centers by hyperscalers (Meta, Microsoft, Google, OpenAI) are surging to meet the demand and pace of AI innovation. The tech giants are in pitched battles with their competitors to secure the equipment, real estate, chips and regulatory approval needed to capitalize on the moment. This creates an opportunity for state and local government leverage. But it also creates an opportunity to innovate upon old frameworks, constructing new partnership models that create win-wins for the private sector and the communities in which they work.
The negotiations start today, if not yesterday, and it's time to replace the outdated data center strategies with a new AI Hub model. Together, we can launch a place-based strategy that works in concert with a broader national industrial strategy for the artificial intelligence revolution.
