Local voters across several US states backed restrictions on data center construction in 40 out of 41 municipal ballot contests during the Nov. 3 midterm elections, creating an unexpected advantage for Bitcoin miners that already control large energized sites.
The municipal pushback comes as tech conglomerates struggle to secure the massive electrical capacity needed to train artificial intelligence models. While greenfield developers face community resistance over utility rates and resource consumption, digital asset miners sitting on long-term power purchase agreements are positioning their facilities as ready-made hosting hubs.
Voters Slam the Brakes on New Grid Allocations
According to tracking data compiled by Ballotpedia, voters approved 27 distinct measures explicitly limiting or rejecting data center expansion. These actions included five blanket bans and 12 partial restrictions, alongside six-month development moratoriums approved by residents in three Maine municipalities.
The local backlash centered largely on environmental impact and household power affordability. In Grove City, Ohio, residents passed an ordinance requiring mandatory public referendum approval for any commercial project drawing more than 20 megawatts from the local electrical network.
Because municipal and county authorities hold jurisdiction over zoning, conditional use permits, and building approvals, local opposition can paralyze a planned development regardless of state-level enthusiasm.
“A company might have access to an attractive site and a significant amount of power capacity, but if local restrictions prevent it from expanding or upgrading that infrastructure, the opportunity becomes much more difficult to pursue,” said Jordan Knecht, head of strategic integrations at blockchain infrastructure firm Kresus.
The Premium on Energized Land
The widening regulatory bottleneck at the municipal level coincides with aggressive computing procurement by institutional AI firms. Publicly traded Bitcoin mining companies committed more than $70 billion to high-performance computing and AI contracts throughout 2025 and into early 2026, according to a first-quarter report by CoinShares.
For mining firms that already hold substation access and high-voltage grid interconnections, local zoning barriers elsewhere make those existing megawatts considerably more valuable.
“If a Bitcoin miner has not yet repurposed their electric contract and physical plant to AI, the case to do so just got more compelling,” said Jamie Brown, chief executive of Global Kiwi Labs, which operates hosting infrastructure for trading firms and cryptocurrency exchanges.
However, repurposing a crypto mining footprint into an enterprise tier-three data center involves a steep financial barrier. CoinShares estimates that constructing AI-capable facilities costs between $8 million and $15 million per megawatt, while a basic industrial Bitcoin mining farm costs between $700,000 and $1 million per megawatt.
“Transitioning from Bitcoin mining to AI workloads requires significant additional investment in infrastructure for cooling, networking, and reliability,” Knecht said.
The load profiles of the two operations also differ fundamentally. Bitcoin operations rely on interruptible power tariffs, meaning operators can power down thousands of ASICs within seconds when wholesale energy prices surge during peak hours. AI clusters handling model training require non-stop baseload electricity, eliminating the operational flexibility that power companies often prefer when balancing congested grids.
Federal Oversight and Rising Consumption
The local pushback mirrors broader projections from the International Energy Agency, which calculated that US data centers consumed 224 terawatt-hours of power in 2025. The IEA projects that figure will climb to 426 terawatt-hours by 2030, nearly doubling the sector’s draw on domestic power generation.
That surge is expected to invite tighter federal oversight from an incoming Democratic-controlled Congress, particularly over transmission buildout costs. Lawmakers are anticipated to examine whether residential consumers are subsidizing grid upgrades required by large computing campuses.
“There may be more pressure to ensure that large data center operators are paying their fair share of infrastructure costs rather than passing those expenses on to consumers,” Knecht noted.
Even with federal scrutiny and high retrofit costs, the financial returns of leasing energized capacity to AI operators continue to push mining executives away from pure-play proof-of-work operations. “Miners will swap to AI because it is a better ROI than mining,” Brown said.




