Meta Warns Democratic Midterm Wins Could Make US Data Centre Construction Harder

Meta is preparing for a more difficult environment for building artificial intelligence infrastructure in the United States after Democratic victories in the midterm elections strengthened political opposition to data center projects, according to company president Dina Powell McCormick.

Speaking at the Wall Street Journal Tech conference on Wednesday, Powell McCormick said the election results would make it harder for technology companies to secure approval for new facilities, as communities become more concerned about the local costs of hosting energy-intensive infrastructure.

She also acknowledged that technology companies had failed to explain the economic and social benefits of their investments effectively enough to win public support.

"The visible sort of manifestation of AI is a data center, and it's something that people can say, 'I don't want this in my community,'" she said.

Powell McCormick argued that resistance to data centers goes beyond concerns about electricity consumption, water use and land availability. It also reflects broader public uncertainty about artificial intelligence and how the technology could affect people's lives and livelihoods.

"I think there's an underlying bigger issue, which is how people feel about AI in general," she said.

Her comments point to a growing challenge for Meta and other hyperscalers seeking to expand computing capacity. The companies need more data centers to train and operate AI models, but local opposition and political scrutiny are making access to land, power, and regulatory approvals more complicated.

The issue gained prominence during the midterm election campaign, when opposition to data center developments became part of local and state-level political debates. For technology companies that have committed billions of dollars to AI infrastructure, the outcome raises questions about the pace and cost of future expansion.

Powell McCormick said Meta believes it can improve public acceptance by demonstrating that data center investments deliver tangible benefits to the communities hosting them, rather than presenting the facilities primarily as infrastructure for technology companies.

She said the industry needs to explain how AI could contribute to areas such as medical research and education, while ensuring that access to the technology is broadly distributed.

"The stakes of who builds the technology that cures cancer, that gives access to education, that is also a very fundamental difference," she said.

Referring to Chief Executive Mark Zuckerberg's approach, she added that he was "very 'long human,'" arguing that the company's investment philosophy places greater emphasis on human potential than on replacing people with machines.

Powell McCormick said the principle behind Meta's position that the future should be for everyone is that people should not be excluded from the benefits of AI. She noted that wider access, fairness and transparency would be essential to making that possible.

Meta is also seeking to demonstrate that its infrastructure projects can provide direct economic benefits. Powell McCormick said the company has been building data centers for about 20 years to support its social networking services and now operates facilities in 23 US states.

Five planned projects, including one in Texas, are designed to have power capacity of at least one gigawatt each. The scale of these developments illustrates the electricity requirements of the next phase of AI expansion, but also explains why they are attracting greater attention from local authorities and residents.

She outlined several practices Meta uses when negotiating with communities. These include paying for the electricity consumed by its facilities, directing tax surpluses generated by data center projects towards teachers and first responders, employing local workers, and securing bipartisan political support.

The approach is intended to address concerns that communities bear the costs of infrastructure development while technology companies capture most of the economic benefits.

For local governments, the major concerns include whether a project will increase electricity bills, place additional pressure on water supplies, generate sustainable employment, and deliver sufficient tax revenue to justify the resources it consumes.

Meta's proposed model attempts to answer those concerns before projects encounter resistance. However, the company's ability to secure support will likely be determined by local residents considering the commitments credible and sufficient to offset the environmental and infrastructure demands of large computing facilities.

Powell McCormick expects future projects to face more extensive scrutiny, with elected officials paying closer attention to whether their constituents actually want the developments. That could make bipartisan support and community engagement more important components of data center investment decisions. It also introduces a political dimension into infrastructure planning that companies cannot resolve simply by committing additional capital.

Meta's concerns come as the company makes substantial investments in computing capacity to support its ambitions in artificial intelligence. Its consumer AI agent, Muse, currently has more than 10 million users, but Meta hopes to scale the product into a service capable of reaching billions of people.

That ambition is central to the company's long-term AI investment case. If Muse attracts a much larger user base, Meta could use its infrastructure to support consumer services and potentially develop new revenue opportunities.

However, the commercial returns from AI investment remain uncertain. Building data centers requires significant upfront spending on land, electricity infrastructure, cooling systems, networking equipment, and specialized computing chips. Those costs must eventually be supported by revenue from AI products, services or customers seeking access to computing capacity.

Powell McCormick indicated that Meta does not depend entirely on Muse becoming a mass-market business to generate returns from its infrastructure.

She said the company was already receiving enquiries from businesses interested in leasing its computing capacity.

"We're getting calls every day to lease that compute [we've built], and so one way or another, we're seeing a path to monetization," she said.

That option could provide Meta with an alternative source of income if demand for its own AI products develops more slowly than expected. Leasing computing capacity would allow the company to generate revenue from infrastructure that might otherwise be underutilized.

It also shows how the economics of AI infrastructure are evolving. Large technology companies are not only building capacity to serve their own applications but are positioning themselves to supply computing resources to other businesses.

However, the existence of demand for leased computing does not automatically guarantee attractive returns. The financial performance of these assets will depend on utilization rates, operating expenses, financing costs and the prices customers are willing to pay.

The political challenge could further complicate the economics. Delays in securing permits, electricity connections or community approval may increase project costs and push back the point at which facilities begin generating revenue.

Meta's immediate priority is therefore twofold: demonstrate that its AI investments can produce commercial returns and persuade communities that hosting the infrastructure is economically worthwhile.

The company has the financial resources and operating experience to pursue large-scale projects, but its expansion plans now face a constraint that additional capital alone cannot remove.