On the night of July 13, the city council in Lansing, Michigan, passed a six-month moratorium on new data center development, joining a growing list of local pauses around the country. The next day, New York Governor Kathy Hochul went further, signing an executive order imposing the country’s first statewide moratorium on hyperscale data centers, pausing new permits for up to a year while the state writes rules governing energy use, water consumption, and community protection.
On the morning of July 15, across the street from Lansing City Hall, the state government moved in the opposite direction: Governor Gretchen Whitmer launched the Michigan Affordability and Responsible Growth Pledge, the first state-level commitment framework of its kind. Within 24 hours, Anthropic, Google, Microsoft, OpenAI, Oracle, and Verrus had signed.
Four days, four announcements, and two different theories of how institutions should absorb a capital surge they did not design for.
Many see this as a contest between states, between political leaders, between companies, and between different visions of the future. Whatever the motivations behind each actor, focusing only on their goals obscures what is actually being tested.
Institutional Gear-Grinding
Alphabet, Amazon, Meta, and Microsoft are on pace to spend roughly $700 billion in combined capital expenditure this year. Data center construction spending year-to-date is more than four times higher than 2025’s record-breaking pace.
For the communities receiving this capital, the scale and speed are far more tangible. A data center project may arrive as a single application larger than any project in the local jurisdiction’s history, far out of scale with everything else on the local docket.
A single hyperscale campus can draw as much electricity as a mid-sized city, and one new project routinely becomes its host town’s largest taxpayer. A planning board that spent last year reviewing subdivisions and a new grocery store is handed a project the size of its entire downtown, with a review process built for the subdivision, not the campus.
When a technology moving in quarterly cycles connects to institutions moving in decadal ones, the interface grinds. The grinding shows up as stalled projects, emergency legislation, rate shock headlines, and packed city council meetings.
Tracking by Data Center Watch puts the value of projects stalled by community opposition at $156 billion in 2025, with more than $130 billion added in the first quarter of 2026 alone. Heatmap Pro polling finds that more than seven in ten Americans now oppose new data center construction near their homes, up from just over four in ten last fall, and a majority say they would support a nationwide halt.
This problem cannot be reduced to a binary yes/no question, nor can a state hope to close this speed gap. It cannot be closed, even at the national level. Model release cycles will not slow to match rate cases, and utility commissions should not approve billion-dollar grid investments at the pace of software product launches. The question is how to connect two systems that must run at different speeds.
As deployment accelerates, everyone in this system is feeling greater torque, just at different points of contact. Residents feel it as rising bills. City councils feel it as overflowing hearing rooms. Developers feel it as projects that stall after capital is committed. Governors feel it as a choice they can no longer defer.
The Clutch
Mechanical engineering has spent centuries managing the transfer of force between systems moving at different speeds, and its vocabulary is useful.
A clutch disengages the gears entirely. Stop the transfer of force, stop the grinding, and use the pause to make adjustments.
That is what New York has done. The moratorium halts new permits for large facilities while the state develops standards for energy, water, and community impact. It is a blunt instrument, and it is meant to be. New York is effectively acknowledging that its existing regulatory framework was not built for gigawatt-scale private loads, and that approving projects under rules written for a different era risks locking in outcomes the state cannot later unwind.
Viewed through the institutional lens, this is a defensible position. A pause is what a system does when it lacks the machinery to engage safely. What happens during a well-used pause is unglamorous but specific. In Lawrence, Kansas, which approved a two-year moratorium the same day as New York’s order, the city’s land development code currently allows large data centers by right in several zoning districts, with no public process through which residents can object. The moratorium exists so planning staff can study water, power, and land use impacts and draft data center-specific regulations; the city commission initiated a code amendment the same night. That is gear-cutting work: rewriting ordinances, defining thresholds, building the review machinery that engagement requires.
The cost is real. A clutch stops good projects along with bad ones, and capital that cannot wait will route elsewhere. But the logic is coherent. Buy time, build capacity, then re-engage.
The open question is what machinery New York can build during the pause. Governor Hochul directed Empire State Development Corp. to, within 60 days, develop a Community Investment Framework to provide localities with “a roadmap for balancing economic development with community interests.”
A moratorium that produces a clear, durable rulebook within its one-year window might prove to have been a reasonable price for institutional readiness. A moratorium that expires into continued uncertainty will have paid the cost without the benefit.
The Differential
A differential does something more sophisticated. It is the mechanism that lets a car turn a corner: the outer wheels must travel farther than the inner ones, and the differential lets its two output shafts turn at different speeds while power continues to flow through both. Neither side has to match the other’s pace. The mechanism absorbs the difference.
Michigan is building a differential, and it is worth being precise about both the design and its current state of assembly. The pledge itself, ten commitments covering full cost responsibility for energy and grid upgrades, additive clean generation, long-term contracts with minimum capacity payments, freshwater protection, local hiring, and public transparency, is the visible layer.
Beneath the pledge sits a mandatory layer that predates the announcement, assembled through Michigan Public Service Commission-approved tariffs and project-specific contracts: minimum billing demand, termination fees, credit and collateral guarantees, and emergency curtailment provisions that put data centers first in line for load shedding. Governor Whitmer is simultaneously asking the legislature to codify those safeguards into statute.
Michigan Affordability and Responsible Growth Pledge
Full cost responsibility for energy demands
Build, bring, or buy additive power
Long-term contracts with minimum capacity payments
Growth that lowers system costs
Grid reliability support
Environmental compliance and oversight
Freshwater protection
Local jobs and expanded tax base
Public transparency and reporting
No cost-shifting to families or small businesses
Full text and signed pledge: Gov. Whitmer Secures Six Company Commitments to Michigan Affordability and Responsible Growth Pledge
The sequencing matters. Michigan could deploy an alignment layer because it had already machined the gears underneath. The pledge works not because it is voluntary but because it is voluntary on top of rules that are not. Its design also tracks what opposition research shows. Heatmap Pro’s project tracking finds that lack of transparency consistently ranks among residents’ top five concerns about data center projects, whether those projects end up approved or canceled. A framework that puts company commitments on the public record, in advance, is aimed at the grievance that actually organizes opposition.
What the six signatures prove is narrower than the headline suggests. The signatories represent the companies behind the hyperscale projects already under development in the state, some already under construction. For them, the pledge largely formalized commitments already in motion, so the marginal cost of signing was likely low. That does not diminish the framework, but it does define its test.
A differential retrofitted onto a moving drivetrain is still a differential. Putting the current wave’s commitments on the public record does more than document the status quo: companies that have publicly pledged to specific terms are poorly positioned to lobby against codifying those same terms, which is leverage for the statutory push now underway. The test is whether the framework shapes the next wave, in Michigan and in the states now studying it.
This is also why the two states’ choices are less contradictory than they appear. New York has not yet assembled an equivalent mandatory layer on which an alignment framework could rest. The clutch and the differential are not competing philosophies; they are different points on the same institutional maturity curve. The remaining question is durability. A pledge binds reputationally, not legally, until the legislature acts, and commitments made in Lansing do not automatically translate to a planning board in Marquette. But as a piece of institutional machinery, it is the most complete one any state has produced.
Drivetrain Design as Strategy
Bessemer Venture Partners estimates $1.5 trillion in capital is stuck in the permitting pipeline. The word pipeline flatters it: little of that capital is actually flowing, and what moves it is not processing speed. It is earlier alignment with the institutions that hold the keys, before positions harden into moratoriums, ballot measures, and rate cases.
Utility commissions, planning boards, and governors’ offices are part of the capital stack whether developers acknowledge it or not. The past week made that impossible to ignore: in one state, institutional posture paused billions in potential deployment; in another, it unlocked commitments from six of the largest companies in the world in a single day.
More states will now choose between the clutch and the differential, and many will try each in sequence. The developers and investors who fare best will be those who treat frameworks like Michigan’s not as compliance burdens to negotiate down but as the operating system they should be seeking out, and helping to build, in every market where they intend to stay.
Infrastructure and institutions will keep running at different speeds. The challenge ahead is how to refine the machinery that connects them.
A car makes it around a corner only because its wheels are free to turn at different speeds while power keeps flowing to both. The decade ahead is a corner.
David Gilford is the Managing Partner of PolicyAlpha. He advised Verrus on its participation in the Michigan Affordability and Responsible Growth Pledge. David writes about infrastructure, private capital, and the institutional systems that determine what gets built.