New York has become the first US state to impose a statewide moratorium on large new data centers, creating an early regulatory test for Bitcoin miners that are rebuilding their businesses around artificial intelligence.
On July 14, Gov. Kathy Hochul signed an executive order directing state regulators to pause incomplete permit applications for new or expanding data centers capable of consuming at least 50 megawatts of power. The temporary halt will remain in effect while officials study the projects’ effects on electricity demand, water supplies, air quality, noise, and surrounding communities.
Applications declared complete before the order can continue, while local permits remain outside its scope. The measure therefore stops a portion of the development pipeline rather than every>imposed a two-year moratorium on certain air permits for fossil-fuel power plants supplying electricity directly to proof-of-work mining operations while officials conducted an environmental review.
The latest order expands the state’s scrutiny from a narrow group of crypto facilities to large computing projects that serve AI, cloud services, and other digital businesses.
While Bitcoin mining is absent from the current order, the facilities it covers closely resemble the infrastructure that an increasing number of miners hope to operate.
Over the past year, public $BTC mining companies have been converting sites built around large power connections, substations, and industrial land into campuses capable of hosting the graphics processors used for AI.
New York’s action therefore introduces a potential obstacle for an industry seeking to reduce its exposure to Bitcoin prices and the worsening economics of producing the cryptocurrency.
$BTC miners have tied their next growth cycle to AI
Bitcoin miners have committed billions of dollars to AI infrastructure, seeking more predictable revenue from the power-rich sites originally built to produce the top crypto.
Publicly traded miners have announced more than $70 billion in contracts to host AI and high-performance computing workloads. Matthew Kimmell, an investment strategist at CoinShares Valkyrie, estimated that AI could generate roughly 80% of public miners’ revenue by the end of 2026.
The opportunity is being driven by an unprecedented expansion in technology spending. Goldman Sachs estimates that annual AI capital expenditure could reach $765 billion in 2026 and rise to $1.6 trillion by 2031 as companies invest in data centers, chips, power generation, transmission infrastructure and cooling systems.
Bitcoin miners are positioned to supply some of the most constrained parts of that buildout. Many of these firms already control industrial land, large electricity allocations, energized substations and grid connections that can take years for new developers to secure. They also have experience operating power-intensive computing facilities around the clock.
Keel Infrastructure, formerly known as Bitfarms, illustrated the scale of the transition this week after officials in Sherbrooke, Quebec, conditionally approved a land sale tied to its proposed C$1.8 billion high-performance computing campus.
Keel plans to consolidate 96 megawatts of electricity currently distributed across three Bitcoin-mining facilities into a single AI>specialized machines used to mine Bitcoin generally cannot process AI workloads, forcing operators to install advanced graphics processors, networking equipment, backup power systems, and more sophisticated cooling infrastructure.
Miners are accepting those costs because AI contracts can run for 10 years or longer, offering revenue visibility that Bitcoin mining cannot provide. Mining income fluctuates with cryptocurrency prices, network competition, and periodic reductions in the block reward.
Those pressures intensified during the past year as CoinShares estimated that the average cash cost of producing one Bitcoin among publicly traded miners rose to about $79,995 in the fourth quarter of 2025, while revenue earned from each unit of computing power fell near multiyear lows.
AI, therefore, offers miners a way to convert electricity capacity into contracted infrastructure revenue.
Data center backlash spreads beyond New York
Meanwhile, the earnings opportunity that is drawing Bitcoin miners into AI is facing a widening political backlash as lawmakers respond to the electricity, water, and infrastructure demands of large data centers.
A Gallup survey conducted in March found that 71% of US adults opposed the construction of an AI data center in their local area, with 48% strongly opposed. About 70% said they were concerned about the facilities’ environmental effects.
Resource consumption was the most common source of opposition. Half of respondents who opposed local development cited excessive use of electricity, water, or other resources, while others raised concerns about pollution, higher utility bills, traffic, and the effects of large campuses on surrounding communities. Supporters most often pointed to potential jobs, tax revenue and broader economic benefits.
That public unease is beginning to shape legislation.
Lawmakers in 15 states had considered>said, with proposals still under consideration in Delaware, Georgia, Michigan, Pennsylvania, South Carolina and Vermont.
Pennsylvania lawmakers proposed a three-year pause accompanied by studies of the industry’s economic and environmental effects. A South Carolina bill would suspend local approvals until lawmakers establish a statewide oversight framework, while Vermont legislators proposed restricting new development until 2030.
The movement has also reached the US Congress, where Sen. Bernie Sanders of Vermont and Rep. Alexandria Ocasio-Cortez of New York unveiled the Artificial Intelligence Data Center Moratorium Act in March.
The proposal would halt the construction and expansion of AI data centers until the federal government adopts protections covering utility customers, workers, civil rights and the environment.
Still, most state efforts have yet to produce binding restrictions. Maine’s governor vetoed an 18-month moratorium, while proposals failed in Minnesota, New Hampshire, Oklahoma and South Dakota.
Those outcomes show that opposition has spread more quickly than statewide restrictions.
New York has now broken that pattern. Its action provides lawmakers elsewhere with a working model for restricting development while regulators study electricity costs, water consumption, and local infrastructure demands.
Wider moratoriums could raise the cost of $BTC miners’ AI pivot
If other states follow New York, Bitcoin miners could feel the financial effects before regulators permanently reject a single>AI projects that have yet to generate revenue.
The scale of the required investment leaves limited room for prolonged disruption. CoinShares estimates that Bitcoin-mining infrastructure typically costs about $700,000 to $1 million per megawatt, compared with roughly $8 million to $15 million per megawatt for AI facilities.
The difference reflects the advanced cooling, networking, backup generation and reliability standards demanded by AI customers. Bitcoin mines can reduce operations when electricity prices rise, or grids become strained, while AI tenants generally require near-continuous power and tighter service guarantees.
Miners unable to complete conversions on schedule could remain dependent on Bitcoin production for longer than planned. Their revenue would continue to fluctuate with the cryptocurrency’s price, transaction fees and network competition while capital remains tied to unfinished AI projects.
A wider set of restrictions could also narrow the number of jurisdictions available for development. Fewer viable sites would strengthen the negotiating position of utilities and local governments, which could demand larger contributions toward grid upgrades, taxes and community benefits.
New York’s order provides an early indication of how those additional costs could be imposed.
Hochul directed regulators to consider creating a Grid Acceleration Fund financed through upfront contributions from data-center developers. The money could support transmission upgrades, clean electricity generation, battery storage and protections against projects that fail to reach their proposed size.
The order also calls for a beneficiary-pays system that would place grid and infrastructure costs on the large customers creating them. Regulators may establish separate electricity-service classifications and require data centers to finance dedicated generation or storage capacity.
Those measures could increase the amount miners must invest before an AI facility begins producing revenue. Existing access to land, substations and power would remain valuable, but control of a grid connection may no longer shield developers from the broader cost of serving a large campus.
Companies with geographically diverse portfolios could redirect capital toward regions offering faster approvals and greater access to power, though a widening patchwork of state restrictions would make that flexibility more expensive.
As a result, $BTC miners could face longer development timelines, higher infrastructure contributions and a smaller pool of locations capable of supporting large AI campuses.
