Meta says election backlash may complicate data-center buildout
Meta President Dina Powell McCormick said Democratic wins in the midterm elections could make it harder for technology companies to build data centers, as opposition to such projects featured in local political campaigns. Speaking at the WSJ Tech conference, she linked the resistance not only to individual facilities but also to public attitudes towards artificial intelligence.
Meta operates data centers in 23 US states, Powell McCormick said. The company has five projects planned at one gigawatt or more, including a project in Texas. She said Meta has built data centers for 20 years to support the computing requirements of its social-networking products.
Data centers become a visible proxy for AI
“The visible sort of manifestation of AI is a data center,” Powell McCormick said, describing why residents can oppose a proposed facility in their community. In her view, the debate reflects a broader question of how people feel about AI, rather than a response to buildings and equipment alone.
She said Meta must better explain both the local benefits of data-center investment and the value it sees in AI. The executive framed the company’s position around broad access to the technology, saying it should be democratized fairly and transparently rather than leaving people behind.
Meta’s conditions for local projects
Powell McCormick said projects are more likely to succeed where communities see a direct benefit. Meta’s approach includes paying for its own electricity, directing tax surpluses associated with a data center to teachers and first responders, retaining a local workforce, and securing bipartisan support.
She expects projects to receive greater scrutiny as elected officials weigh proposals against constituent preferences. The model that works, she said, is one that states want, but the company anticipates a harder environment for new construction.
Compute capacity retains commercial value
Meta is investing in AI infrastructure in support of its consumer AI agent Muse, which Powell McCormick said has 10 million users today and is intended to scale far beyond that level. The company also sees value in the underlying compute even if the agent does not reach the scale it envisages.
That view aligns with Meta’s plan to monetize surplus AI capacity because excess AI compute can be leased to outside customers, offering a route to monetisation beyond Meta’s own products. Powell McCormick said Meta receives daily calls from parties seeking to lease the compute it has built.
For businesses planning AI capacity, the message is that site selection cannot be separated from community benefits, power costs and political support, while spare compute may retain value through leasing demand.