Test report DSG-9220 · Rev E · tested November 17, 2026

AI Datacenter InfrastructureDevice under test

Voters Back Data Center Restrictions in 40 of 41 Elections

Voters backed restrictions on data centers in 40 of 41 November 3 elections. Amazon, Alphabet, Microsoft and Meta plan $720–$740 billion in 2026 capital spending, but permits may delay sites.

Read
4 min
Words
747
Node
65nm
Operator
Priya Raman

Spec summary

  1. 40 of 41 November 3 local elections backed data-center restrictions.
  2. Amazon, Alphabet, Microsoft and Meta plan $720–$740 billion in 2026 capital spending.
  3. AWS backlog was $578 billion; Google Cloud backlog was $561 billion; Microsoft commercial RPO rose 73% to $677 billion.
  4. The Energy Department projects data centers could use up to 12% of U.S. electricity by 2028.
  5. Microsoft reduced dock-to-live time for new GB300 systems by about 20%.
Voters Just Rejected AI Data Centers in 40 of 41 Elections. Is the $1 Trillion Boom in Trouble?
Fig. AVoters Just Rejected AI Data Centers in 40 of 41 Elections. Is the $1 Trillion Boom in Trouble? — 10512837 / pixabay

What did voters decide?

Forty of 41 local elections held on November 3, 2026, backed restrictions on nearby data-center development, according to Ballotpedia, leaving Harwood, North Dakota as the sole exception.

Ballotpedia counted 27 measures approving restrictions or opposing development, while voters defeated 13 pro-development measures. Voters approved five outright bans, 12 partial bans and three temporary moratoriums (construction pauses).

Those three detailed categories total 20 measures; the source does not itemize the other seven. Local objections centered on electricity bills, water consumption, land use and road strain, with opposition crossing party lines.

How large is the spending commitment?

  • Amazon expects approximately $220 billion in 2026 cash capital spending (capex).
  • Alphabet guides to $195 billion–$205 billion.
  • Microsoft expects approximately $170 billion for calendar 2026.
  • Meta plans $135 billion–$145 billion.

Together, the plans total $720 billion to $740 billion in 2026, establishing the scale of the four companies' stated spending programs without requiring the headline's broader $1 trillion framing. That range is more than 22 times Alphabet's 2023 spending of $32.3 billion.

The source frames the wider buildout as a $1 trillion boom. The four disclosed 2026 plans do not, by themselves, establish that broader figure. These are spending expectations, not completed investment, and the source assigns no site-by-site construction schedule.

Alphabet reported negative $9.07 billion in free cash flow (cash remaining after operating and capital spending) in the prior quarter. It also issued $28.7 billion in notes (debt securities). Meta's free cash flow was negative $2.12 billion.

Over the past month, Alphabet shares fell 17.77% and Meta shares fell 10.62%. The source does not establish that election results caused either decline.

Does contracted demand still justify the buildout?

The order books remain large, but the source mixes backlog and remaining performance obligations, which may use different accounting definitions and should not be treated as one market total.

  • AWS reported a $578 billion backlog (future contracted revenue).
  • Google Cloud reported $561 billion in backlog.
  • Microsoft's commercial remaining performance obligations (RPO, contracted revenue not yet recognized) rose 73% to $677 billion.

Andy Jassy said big customers reserve “hundreds of megawatts or sometimes up to a gigawatt of capacity” in advance. These reservations establish customer commitments, but they do not guarantee local permits, grid connections or construction dates for every project.

What can grids and communities absorb?

The Energy Department projects that data centers could use up to 12% of U.S. electricity by 2028, concentrating new loads on local transmission and distribution grids. One Denver campus may use up to 805,000 gallons of water per day.

Amazon's earnings call offered no specific management commentary on local opposition or permitting. The company nevertheless announced a $1 billion “Built Together” investment in U.S. data-center communities.

That commitment puts a disclosed amount behind community relations, although the announcement does not identify approved projects or offset power, water and road constraints.

Who bears the cost of a siting delay?

All four companies can move or postpone projects. Smaller developers may lack that flexibility. For a debt-funded project, a 12-month delay can materially alter the economics, while developers with secured power, permits and financing gain ground on rivals still awaiting site approval.

Microsoft cut dock-to-live time (the interval from hardware arrival to live operation) for new GB300 systems by about 20%, extracting more capacity from sites it already operates. Alphabet is placing AI chips in partner and customer facilities. Thomas Kurian said this “offsets the need for us to have to prebuild facilities.”

These responses raise utilization or reduce the need for prebuilt capacity. They do not remove municipal opposition, connection queues or the cost of finding an acceptable site.

What signals matter in 2027?

Amazon plans to spend meaningfully more in 2027, while Alphabet expects capex to increase significantly; neither company supplied a new dollar range in the report. The next fourth-quarter results should be tested against three signals: 2027 capital-spending guidance, local permitting in risk disclosures and any further delay to Microsoft third-party capacity.

Microsoft's third-party capacity has already slipped to the second half of the year. Another delay would connect voter resistance to revenue timing and project costs, although the source provides no model quantifying that effect.

Voter opposition now implies slower and costlier siting for the four companies. The more acute risk falls on developers that lack both power and permits.

via 247wallst.com (Source)

Filed under

  • data-centers
  • ai-infrastructure
  • data-center-permitting
  • energy-demand
  • hyperscaler-capex
Share this article:

More from Priya Raman

Priya Raman

Show full bio

Correspondent covering money, funding and market shifts around hardware layer of artificial intelligence at Die Signal.

661 articles

Same lot · LOT-362D

« Previous articleNext article »