By Jenna Somers
AI data centers seem to be cropping up everywhere, often in rural farming communities, where crops are grown, where livelihoods depend on their yield and where their yield depends on groundwater, the cost of energy and infrastructure, and the health of the environment.
Increasingly, these communities, in states like and , want to know the impacts of AI data centers on the cost of utilities and the local economy. But getting this information has not been easy. For , public pressure in a rural Wisconsin community forced one public utility partner, contracted to operate a Meta AI data center, to refile their contract with the state, so that it would offer more transparency about the impact of the center. Even so, an environmental group鈥檚 lawsuit to gain more transparency on the center鈥檚 energy demands remains open.
Scenarios like this, of schisms in public-private partnerships due to opaque contracts, are not outliers, according to , a leading scholar of public-private partnerships and University Distinguished Professor of Leadership, Policy, and Organizations at Vanderbilt of education and human development. In , a new book published by Oxford University Press, Heinrich and her co-authors at Indiana University, the University of Oxford and the University of the Arts London analyze two case studies of public-private partnerships, one in the U.S. and one in the U.K., to reveal how traditional, formal government contracts can fail to deliver public value and how a new contract model, rooted in relationships, can succeed.
鈥淚f contracts are being developed well, they should address challenging circumstances and things like rapidly evolving technology, but the government tends to rely on very old, rigid, boilerplate-type contracts that aren鈥檛 well suited for the problems we have today,鈥� Heinrich said. 鈥淭hey need to be adaptive鈥攁s one of my colleagues said鈥攁 鈥榣iving instrument鈥欌�攂uilt on relationships of trust and transparency for protecting the public interest.鈥�
Two partnerships, two outcomes: what makes the difference?
The two public-private partnerships highlighted in the book were created to serve vulnerable populations. In the U.S., the state of Indiana contracted with IBM to automate the state鈥檚 welfare services. Similarly, in the U.K., the government contracted with three organizations to create the Kirklees Better Outcomes Partnership, which helps people at risk of homelessness to remain in their homes.
The IBM-Indiana partnership became embroiled in conflict over rigid performance metrics, and the formal contract lacked mechanisms for adapting expectations and communicating problems. Instead, both parties used mechanisms of the contract punitively. The partnership ended in a costly decade-long legal battle that, ironically, consumed precious public resources that could have gone toward serving vulnerable members of the public whom the partnership was originally intended to help.
Conversely, the Kirklees Better Outcomes Partnership聽experienced disruptions caused by the Covid-19 pandemic; but from the outset, the contracts between the organizations structured relational mechanisms to address unforeseen problems. For example, they held regular meetings and forums to discuss data and engage in collaborative problem solving. Early on, they realized a need to modify accountability metrics鈥攕omething not possible in the IBM-Indiana partnership. When circumstances changed, the relational structures of the partnership were strong enough that they could adapt the formal contract components to reflect new realities.
“The difference wasn’t geographical context,” Heinrich said. “It was whether the government and private organizations intentionally built relational infrastructure alongside the formal contract鈥攎echanisms for trust, transparency, and collaborative adaptation.”
Formal-relational contracts for effective contracting
Heinrich and her co-authors suggest that public-private partnerships establish relational mechanisms early, such as structured data-sharing protocols, bench-learning meetings to discuss data and whether plans need to change, a definition of public value, adaptive frameworks for unforeseen circumstances, and transparency requirements.
Additionally, the authors emphasize that contracts are not a one-size-fits-all approach. Governments should think about the varying needs of different domains. For example, a 10-year research partnership should have different formal and relational structures in place compared to a two-year technical services contract. Likewise, formal and relational needs exist on a spectrum, with some partnerships needing more formal than relational structures and vice versa.
The urgent need for reform
Governments, dealing with constrained budgets, are increasingly turning to public-private partnerships to deliver for the public interest. These partnerships are expanding the private sector鈥檚 influence over various domains, including infrastructure, social services, education, prisons and immigration detention facilities. According to Heinrich, without a new approach to government contracts that prioritizes public value, these partnerships risk wasting public resources, harming communities and creating power imbalances that benefit the private sector over the public good. 鈥淲ith more at stake than ever for public well-being, we can鈥檛 afford to continue our outdated, business-as-usual approach to public contracting,鈥� Heinrich said.
The co-authors of are
- Deanna Malatesta, associate professor at Indiana University’s O’Neill School of Public and Environmental Affairs
- Eleanor Carter, UK Research and Innovation Future Leaders Research Fellow at the University of Oxford Blavatnik School of Government and academic co-director of the Government Outcomes Lab
- Michael Gibson, DPhil student in public policy and research and a policy associate with the Government Outcomes Lab at the University of Oxford Blavatnik School of Government
- Nigel Ball, director of Social Purpose Lab at University of the Arts London.
