This paper starts the project of elucidating what a power-with looks like at a systematic long-term level and why it is preferable to a power-over. At an episodic, one-off level a power-over—that is, hoarding power and value—appears more profitable. However, the dynamic changes over the long term. Maximising our capacity to act (i.e. power) is inherently collective; enhancing our capacity to act requires combining with other things (e.g. people, objects, ideas). With human collaborators, future rounds of collaboration and value creation are jeopardised by past hoarding. Power and value hoarding erode reputation, trust, effective feedback, and partners’ capacity, limit innovation and resilience, and induce resistance and withdrawal from the partnership. In contrast, where value is shared (i.e. a power with) collaborators reinvest, and more value is created, a virtuous cycle. In this paper, alongside a simpler model demonstrating virtuous and vicious cycles of value creation or destruction, I demonstrate more complex systemic forms of power-with that I label ‘collective experimentation’ and ‘collaborative competition’. Collective experiments leverage diversity and, through experimentation, create new capacities (i.e. power) and resilience to change. They mitigate the groups’ risk of failure from experiments by sharing value from successes. Data is used as feedback from experiments rather than to judge and control. Facilitated by governing agents with protections against corruption and value hoarding, the capacities of all are enhanced, leading to collective flourishing. Real-world case studies will be provided and linked to theory to illustrate the applicability of these ideas to governments’ use of data.