AI Data Centers in Snohomish, Skagit and King County
Benefits, concerns, and potential real-estate impacts — a neutral look under the hood.
This topic focused on AI data centers in Snohomish, Skagit, and King County may be quite dry for many of you, but if you're monitoring the news as of late, it's a very popular topic of discussion and debate. It has also become quite political.
I wanted to create an article that that's neutral and factual, based on what we know TODAY (8/30/26). The market growth and evolution of AI and its associated brain and nervous system (i.e. data centers, networks, and electricity) is extremely fast moving, so what's factual and relevant today, may not be the case tomorrow.
PLEASE take the following information with a big grain of salt, and always conduct your own due diligence if you want to stay up to speed on this topic, and/or it may influence your situation (buying a home, potential AI data center construction near you, etc). I've included as many relevant links at the bottom of this article to help you dig deeper, continue to learn, and stay in the loop.
AI infrastructure, utility plans, development applications and local rules are changing quickly. A proposal can be revised, paused or approved within days. This article reflects the best available public information as 8/30/26. Please confirm current status through the linked government and utility sources before relying on it. |
A data center may look like a quiet warehouse with fewer parking spaces. Economically, however, it behaves more like a capital-heavy factory for computing: it needs dependable power, cooling, fiber, security and specialized equipment around the clock.
That can be good news, difficult news, or even both, depending on the community that hosts it.
One important distinction comes first: not every data center is an “AI data center.”
Data centers also support cloud storage, hospitals, 911 systems, universities, banking, streaming and ordinary business computing. Public permits rarely reveal the exact mix of workloads. The most accurate local discussion is therefore about data centers generally, while recognizing that AI is accelerating demand for larger, more power-dense facilities.
The short of it: Data centers can bring construction work, high-value taxable property and infrastructure investment. They can also create long-lived electricity, water, land-use, noise and neighborhood concerns, and potentially, real impact to local communities. The concerns and issues depend less on the label “data center” than on the site, design, utility agreement, incentives and local rules.
What is the current status of AI data centers in Snohomish, Skagit and King County?
“County status” is not one simple map. Counties regulate unincorporated land; cities regulate property inside their boundaries; utilities separately decide how and when they can serve large electrical loads.
There is no authoritative public inventory identifying every operating, proposed or AI-specific facility across counties (or states, or nationally).
Area | Verified status as of August 30, 2026 | What it means |
Snohomish County | Emergency Ordinance 26-026 created a six-month moratorium for data-center applications in all zones of unincorporated Snohomish County, effective June 24. Marysville adopted a separate six-month city pause. | This is a temporary planning period, not a permanent ban. Incorporated cities retain their own authority. Snohomish PUD says its large-load approach is intended to make “growth pay for growth.” |
Skagit County | Ordinance O20260005 imposed a six-month pause in unincorporated Skagit County on facilities over 2,000 square feet or with planned load of at least 2 MW. | Draft permanent rules were still under review. The proposed size and power limits were not final law at the cutoff, and cities and towns were outside the county pause. |
King County | There was no single countywide moratorium. Seattle Ordinance 127447 created a 365-day pause for facilities over 20 MVA. Federal Way and Burien enacted their own one-year pauses; Renton’s proposal was still awaiting final reading. | King County is a city-by-city and unincorporated-area patchwork. A utility inquiry or permit application is not the same as an approved project. |
What good can AI data centers bring to a community?
Construction and skilled-trade work. Large campuses can employ substantial construction crews and create demand for electricians, line workers, HVAC specialists, equipment installers, engineers and suppliers. The Bureau of Labor Statistics identifies AI data centers as one contributor to growing demand for electrical work. A detailed Virginia legislative study offers useful scale: a typical 250,000-square-foot building may have about 1,500 workers at peak construction but roughly 50 full-time workers after opening. That is an illustration, not a Washington forecast, but it explains why benefits are often strongest during the build.
Tax revenue. Expensive buildings and computer equipment can expand a local tax base. Washington’s preliminary Data Center Workgroup estimated the industry supported 8,990 direct jobs and contributed $1.8 billion in direct and indirect state and local tax revenue in 2023, net of more than $118 million in exemptions. The net local benefit still depends on the tax rules, incentives, public-service costs and infrastructure commitments in each deal.
Digital and energy investment. Data centers support services residents and businesses use every day. Major customers may also finance substations, transmission upgrades, clean generation, storage or water-reuse systems. Those investments can add regional capacity and spur innovation - but they are not automatically free benefits for everyone. Ownership, cost allocation and long-term use matter.
A possible productivity payoff. The Federal Reserve’s July 2026 review found that AI-related investment was supporting parts of the economy, while broad productivity gains had not yet become clear in aggregate data. In plain English: the buildout is real; the full economic return remains a work in progress.
What are the community tradeoffs?
Electricity demand and utility bills. The latest Lawrence Berkeley National Laboratory update estimates that data centers could use 11.8% of U.S. electricity in 2030, with scenarios ranging from 9.5% to 15.3%. That does not mean household rates must rise by the same amount, or even at all.
The result depends on how utilities price new service, assign infrastructure costs and protect customers if a proposed facility is delayed or leaves.
This is why rate design matters. Seattle City Light’s proposed large-load policy would place new facilities requesting at least 10 MVA into a separate rate class, require them to accept infrastructure costs and put them into a service queue.
Snohomish PUD similarly says project-specific growth costs should be borne by the requesting customer. Those policies reduce cost-shifting risk, although no policy can remove every forecasting or system-wide risk.
Water use. This is hot topic as of late on the media, and in the political ring. Sweeping claims such as “every data center consumes a city’s worth of water” are unreliable, and not always correct. In reality, it depends on several variables.
Cooling design, climate, server density and operating practices produce enormous variation. Virginia’s JLARC found that most facilities in its study used about as much water as, or less than, a large office building; a few used substantially more and some used less than a household. The right local questions are: What cooling system is proposed? Is potable or reclaimed water used? What happens during hot, dry periods? What wastewater is discharged?
Noise, generators and appearance. Cooling fans and mechanical equipment can create a constant low-frequency hum. Backup generators add testing noise and localized emissions. The Virginia review found most data centers produced no noise complaints because of their location or design, but about 10% of operating sites appeared to have generated noise residents considered problematic. Setbacks, sound modeling, equipment placement, berms and generator-testing schedules can make a major difference.
Construction pressure. Data-center projects compete most directly for specialized commercial trades and electrical equipment, not necessarily for the lumber, drywall or framing crews used in a typical home. The Federal Reserve has noted wage pressure in specialty trades, and shortages of transformers, switchgear and skilled workers can affect other projects locally. Evidence does not yet support a precise national “data-center surcharge” on home construction.
Land use and opportunity cost. A campus can occupy industrial, commercial or rural land for decades while providing relatively few daily customers or permanent workers. Supporting substations, transmission lines, battery systems and roads may affect areas beyond the data-center property itself. On the other hand, a well-screened project in a suitable industrial zone may create fewer traffic and public-service demands than some alternative uses.
Do data centers raise mortgage rates or lower nearby home values?
This is discussion point I'm hearing more and more - that corporate data center bond / loans may be driving up mortgage rates, due to their high yields, and volume. It's important to note that data-center corporate loans are not mixed into residential mortgage-backed securities.
The more credible connection is indirect. A Dallas Fed analysis estimates that AI-related investment-grade borrowing in 2026 could add as much as $360 billion of 10-year-equivalent duration, about one-eighth of Treasury duration supply. More long-term bonds competing for investors can place some upward pressure on long yields.
Separately, higher electricity prices can add modestly to inflation; another Dallas Fed model estimates roughly 0.04 to 0.13 percentage points of additional annual PCE (Personal Consumption Expenditures) inflation by 2030 is possible.
However, no credible study has isolated an AI-data-center effect on 30-year mortgage rates, at least yet. Freddie Mac’s national average was 6.66% on August 27, 2026, but there is no evidence-based way to assign a measurable portion of that rate to AI data centers. Treasury yields, inflation, economic growth, Federal Reserve expectations, federal borrowing, market volatility, and mortgage-bond spreads are the primary drivers. A reasonable conclusion at this point is that a small, but potentially growing influence, not the explanation for today’s higher mortgage rates.
Home values: no universal discount. The Virginia study found no established decline in nearby home values or sales speed, even though some residents reported serious quality-of-life concerns.
An individual property can still become harder to market if it has direct line of sight, nighttime hum, generator exposure, construction traffic or a large future expansion nearby. Distance, terrain, screening and the alternative permitted land use all matter. A blanket 5%, 10% or 15% adjustment is not supported by the evidence. This is yet another reason it's important to understand and research the area you're consider moving to and buying a home, or even if you're looking to sell a home in the near future.
Six practical questions for homeowners and neighbors
Which government has authority? Confirm whether the site is in a city or unincorporated county land.
What stage is the project in? Separate a utility inquiry, pre-application, SEPA review, permit application, approval and active construction.
Who pays for power and infrastructure? Look for a special tariff, minimum bill, deposit, service term and responsibility for stranded assets.
What will nearby residents experience? Ask for sound modeling, setbacks, cooling design, generator locations, lighting, traffic plans and all future phases.
What is the net public benefit? Compare permanent jobs and taxes with incentives, public-service needs, water demand, and the land’s other possible uses.
What development is proposed or approved nearby? Concerned about a particular piece of land or existing large facility or warehouse? First confirm whether the property is inside a city or in an unincorporated area, since cities and counties maintain separate permit records. Search the correct permit system by address, parcel number or nearby street, and review active permits, public notices and SEPA environmental decisions.
Do not search only for “AI data center.” A project may instead be described as a data center, technology facility, warehouse, industrial building or electrical substation. Also remember that “under review” is not the same as approved, and an approved land-use application does not necessarily mean construction will occur.
Useful search tools:
Not every commercial project requires public notice, so concerned residents can also contact the appropriate planning department with an address or parcel number.
The bottom line
AI data centers are neither a guaranteed windfall nor an automatic community burden.
They are unusually large, long-lived infrastructure projects whose benefits and costs can land in different places: construction workers may benefit regionally (during a relative short construction period), tax revenue may benefit a city, grid upgrades may serve a wider area, and noise or views may be concentrated on the nearest homes.
The best local decisions start with transparent project facts, enforceable utility agreements and site-specific safeguards, not assumptions based on the word “AI.”
I track AI closely and am here to help answer questions, review new information or point readers toward the right county, city, utility, or research source.
This article is a dated snapshot, not legal, financial, engineering or appraisal advice. AI technology, power forecasts, project applications, utility tariffs and local ordinances can change from one day or week to the next. Recheck the primary sources and the specific parcel’s current permit record before making a real-estate, investment or public-comment decision. |
Live Research Links: Local and National
Snohomish County
Snohomish County File 2026-1510 / Emergency Ordinance 26-026 — Official legislative history, scope and status of the unincorporated-county moratorium.
Snohomish PUD: Growth Without Compromise — Explains the utility’s approach to large loads, system studies and customer responsibility for growth costs.
Skagit County
Skagit County moratorium announcement — Plain-language official summary of the rural/unincorporated scope and the county’s rulemaking process.
Skagit County Ordinance O20260005 — Full six-month ordinance, including the 2,000-square-foot and 2 MW thresholds.
Skagit County 2026 docket notice — Tracks proposed permanent code amendments; useful for distinguishing draft rules from adopted law.
King County and its cities
King County permit and public-notice research — Starting point for projects on unincorporated King County land.
Seattle Ordinance 127447 — Official 365-day moratorium record and definition for facilities above 20 MVA.
Seattle City Light: Getting Ahead of Data Center Power Demands — Accessible explanation of the proposed large-load rate class and service queue.
MRSC: Washington local data-center rules and moratoria — Statewide index of local ordinances; always confirm against the linked city or county record.
Washington and National Perspective
Washington Data Center Workgroup — Official home for the state’s preliminary findings on jobs, taxes, energy, water and policy; the final report was still pending at the cutoff.
MRSC: What the Data Center Boom Means for Washington Local Governments — Nonpartisan local-government overview of benefits, infrastructure needs and evolving state policy.
Lawrence Berkeley National Laboratory: U.S. Data Center Energy Usage Report, 2025 Update — Latest national electricity-use scenarios through 2030 and the assumptions behind the wide range.
Virginia JLARC: Data Centers in Virginia — One of the most detailed public studies of jobs, taxes, power, water, noise and residential compatibility.
BLS: Construction employment growth — Federal labor analysis connecting data-center construction with demand for electricians and related trades.
Dallas Fed: AI debt financing and interest rates — Explains how corporate bonds and swaps can add long-term duration supply without claiming a measured mortgage-rate effect.
Dallas Fed: Data centers and electricity inflation — Scenario analysis estimating a modest but uncertain inflation channel through retail power prices.
Federal Reserve: The AI Buildout and the Economy — Tracks investment, adoption, productivity and labor evidence, emphasizing what is and is not yet measurable.
Freddie Mac Primary Mortgage Market Survey — Current weekly national mortgage-rate benchmark for readers checking the rate environment after publication.
Comments