White Slough Data Center: A Second Look Needs Better Numbers
LodiEye — November 2026
At a Glance
Test the offer, not the slogan. A data center’s potential public benefit depends on its size, full service costs and binding commitments. The examples below deliberately vary electrical capacity and operating assumptions. They do not establish available power, surplus water, achievable profits or promised bill credits.
Customer bill credits are worth examining. They should depend on verified project costs and enforceable commitments—not assumptions about surplus water or available power.
Former mayor Jack Sieglock’s White Slough data-center pitch combines developer-funded construction, recycled-water cooling and customer bill credits, according to Steve Mann’s November 16 About Town post, “Datacenters.” Mann describes Sieglock’s position with the words, “There would be no noise or pollution.” That is Mann’s account of Sieglock’s argument, not an independently authenticated verbatim statement by Sieglock.[1]
The appropriate response is neither an automatic rejection nor a promise of cheaper electricity. It is a public feasibility test: What would be built, how much electricity and water would it require, who would pay every attributable cost, and what would actually remain for residents?
Start with the missing number
Megawatts describe electrical demand or capacity. Kilowatt-hours measure electricity consumed over time. A data center’s computing-equipment demand is also different from the demand of the whole facility, which includes cooling and other electrical systems; Berkeley Lab treats those separately in its national energy analysis.[2]
LodiEye’s scenarios below use total facility capacity at the electricity meter, not computing capacity alone. They test four hypothetical sizes and three electrical load factors—the ratio of average demand to the stated capacity. A 75% electrical load factor does not mean that servers are performing computing tasks 75% of the time.
| Facility capacity | 50% load factor | 75% load factor | 90% load factor |
|---|---|---|---|
| 5 MW | 21.90 | 32.85 | 39.42 |
| 10 MW | 43.80 | 65.70 | 78.84 |
| 25 MW | 109.50 | 164.25 | 197.10 |
| 50 MW | 219.00 | 328.50 | 394.20 |
Source: LodiEye calculations. Values are million kWh per year. Capacity in MW × 1,000 × 8,760 × assumed load factor. Illustrative cases—not forecasts or findings of available capacity.
The definition matters. At an assumed power usage effectiveness, or PUE, of 1.20, a project described as 10 MW of computing load would require 12 MW for the entire facility. Berkeley Lab defines PUE as total facility electricity divided by computing-equipment electricity.[2]
Before discussing a benefit package, the city should require a developer to specify both numbers and distinguish initial demand from ultimate buildout. Nothing in this scenario exercise establishes a 10–15 MW local ceiling or proves that a larger installation is feasible.
ILLUSTRATIVE SCENARIOS • Total facility electricity, not computing load alone.
Chart loading. All plotted values appear in the accompanying table.
Source: LodiEye calculations in Table 1. The 5, 10, 25 and 50 MW capacities are inputs—not findings of local service availability.
Sales are not customer savings
An electricity customer can generate substantial gross utility revenue without generating the same amount of money for bill credits. The relevant question is what remains after accounting for the costs attributable to serving that customer.
LodiEye’s model therefore uses an assumed net distributable amount per facility kWh, not the retail electricity rate. This amount is defined here as what remains after power purchases, capacity and transmission obligations, operations, infrastructure financing, administration and appropriate reserves. Separately negotiated land-lease income would need its own accounting, avoiding double counting.
| Facility capacity | 1¢ net per kWh | 3¢ net per kWh | 5¢ net per kWh |
|---|---|---|---|
| 5 MW | $328,500 | $985,500 | $1,642,500 |
| 10 MW | $657,000 | $1,971,000 | $3,285,000 |
| 25 MW | $1,642,500 | $4,927,500 | $8,212,500 |
| 50 MW | $3,285,000 | $9,855,000 | $16,425,000 |
Source: LodiEye calculations at a 75% electrical load factor. Margins are hypothetical amounts after all attributable costs and reserves. Actual results could be zero or negative.
For an example of the potential household-scale benefit, assume 28,000 eligible accounts, a deliberately rounded modeling assumption—not a verified current customer count. If all the modeled money went into equal credits, a 10 MW facility at a 75% load factor would produce about $1.96, $5.87 or $9.78 per account per month at the respective 1-, 3- and 5-cent margins. Allocating money to the general fund or another purpose would reduce those credits.
Across the modeled 50%–90% load factors and positive 1–5 cent margins, that same 10 MW case yields approximately $1.30–$11.73 monthly per assumed account. This is a sensitivity range, not a prediction. At a negative margin of 1 cent per kWh and a 75% load factor, the facility would instead create a $657,000 annual shortfall; the agreement would need to determine who bears it.
ILLUSTRATIVE SCENARIOS • 75% electrical load factor; 28,000 assumed eligible accounts.
Chart loading. All plotted values appear in the accompanying table.
| Capacity | 1¢ net/kWh | 3¢ net/kWh | 5¢ net/kWh |
|---|---|---|---|
| 5 MW | $0.98 | $2.93 | $4.89 |
| 10 MW | $1.96 | $5.87 | $9.78 |
| 25 MW | $4.89 | $14.67 | $24.44 |
| 50 MW | $9.78 | $29.33 | $48.88 |
Source: LodiEye calculations. 75% load factor and 28,000 assumed eligible accounts. Equal allocation of all modeled net funds; no verified account count or promised savings.
Source: LodiEye calculations. These margins and the account count are modeling assumptions, not utility forecasts.
The scale required for very large benefits is equally important. Producing $20 million annually from electricity service alone would require approximately 61 MW at a 5-cent net margin, 101 MW at 3 cents, or 304 MW at 1 cent, each at a 75% load factor. Those calculations test a hypothetical target; they do not establish available capacity, achievable margins or an actual city revenue forecast.
Paying upfront is not enough
LodiEye’s proposed safeguard is straightforward: a developer should cover not only initial construction, but also the continuing costs and financial risks created by its demand. Contracts should address delayed occupancy, lower-than-promised consumption, early closure, unpaid obligations and infrastructure that remains after a customer leaves.
This is not solely a local concern. Reuters reported on disputes over allocating grid costs; the industry argues that reducing peak demand can lower shared costs.[3]
CalMatters also reported the Little Hoover Commission’s recommendation that data centers pay their full infrastructure and grid-service costs. That is policy context—not a project-specific cost finding or a statement of enacted legal requirements.[4]
A White Slough analysis should identify the serving utility, service territory, necessary facilities and power-supply arrangements in writing. A nearby generating plant should not be treated as proof of a dedicated supply commitment. This article assigns neither an infrastructure price tag nor an opening year without those findings.
Recycled water needs a budget
The adopted 2025 White Slough water-board order, finding 17, distinguishes permitted average dry-weather flow of 8.5 million gallons a day from municipal influent of approximately 4.5–4.6 million gallons a day. Those are conditions described in that order, not a newly measured November 2026 flow.[5]
The order documents separate municipal and industrial collection systems and existing agricultural, power-plant and fish-rearing uses. Finding 29 also describes preliminary hydrogen-project planning—not an established operating demand or verified current contractual commitment. The record does not establish an unallocated water supply for a data center.[5]
For this proposal, recycled wastewater is more precise than “grey water.” The city should disclose a seasonal water balance showing existing uses, commitments, water quality and the amount reliably available for a new customer. Annual averages alone would not answer whether adequate water is available during the hottest days.
Cooling changes the answer
Cooling design can substantially change both electricity and water requirements. Berkeley Lab describes a trade-off between evaporative cooling’s water consumption and the electricity needed for dry heat rejection; recirculating liquid within a system does not, by itself, make the overall facility water-free.[2]
For illustration, take the 10 MW total-facility case, a 75% load factor and an assumed PUE of 1.20. That produces 54.75 million kWh of computing-equipment electricity annually. Applying three assumed site water usage effectiveness values, or WUE values, gives the following results:
| Assumed site WUE | Annual consumption | Average daily consumption |
|---|---|---|
| 0.10 liter per computing kWh | About 1.45 million gallons | About 4,000 gallons |
| 0.36 liter per computing kWh | About 5.21 million gallons | About 14,300 gallons |
| 1.80 liters per computing kWh | About 26.03 million gallons | About 71,300 gallons |
Source: LodiEye sensitivity calculations. 10 MW total capacity, 75% load factor, PUE 1.20 and 54.75 million computing kWh annually. WUE uses computing-equipment electricity. The middle case approximates Berkeley Lab’s modeled 2023 U.S. average; others are illustrative—not verified designs or universal bounds. [2]
These estimates concern on-site consumption, not necessarily total water delivered, and exclude water consumed in generating electricity elsewhere. The modeled daily figures are annual averages, not hot-weather maximums. A developer would need to document its cooling system, maximum-day supply requirement, discharge characteristics and water-quality treatment needs.
Test the environmental assurances
LodiEye’s recommendation is to replace an absolute assurance about noise and pollution with measurable design requirements. Cooling equipment, emergency power systems, fuel handling, lighting and construction should receive site-specific evaluation, with enforceable operating limits and monitoring where warranted.
The California Energy Commission’s description of the unrelated RB Inyokern proposal, for example, includes diesel backup generators. That does not establish what White Slough would use; it demonstrates why the backup-power design must be disclosed before environmental assurances are accepted.[6]
Flood resilience also deserves attention. The adopted White Slough order identifies the treatment facility and some land west of Interstate 5 as lying within the Delta’s 100-year floodplain. That finding is not a flood determination for an unspecified data-center parcel, but it supports requiring parcel-specific review.[5]
The county has extended its pause
The county’s October 7 announcement reported an initial 45-day moratorium on land-use approvals for new data-center development in unincorporated San Joaquin County.[7] Its November 17 announcement reported an extension of 10 months and 15 days for hyperscale data-center development in unincorporated territory.[8]
Those are county announcements, not a substitute for the enacted ordinance’s definitions, exemptions or parcel-specific applicability. Whether a particular White Slough site and proposal would fall under the restrictions requires a written jurisdictional determination. City ownership alone should not be used to assume either applicability or exemption.
This article does not calculate a precise moratorium expiration date, claim that a White Slough project is legally barred, or rely on unverified summaries of state bills. Those are separate legal questions.
What a public offer needs
Before negotiating customer credits, LodiEye recommends publishing a project-specific feasibility package containing:
- A defined project: Parcel, operator, phased buildout, total facility MW and computing-equipment MW.
- A written service determination: Serving utility, available capacity, required upgrades, supply obligations and scheduling dependencies.
- A complete financial model: Gross receipts, all attributable costs, conservative operating cases, independently reviewed assumptions and clearly separated lease income.
- Enforceable financial protection: Upfront payments where appropriate, adequate security, minimum obligations tied to committed costs and responsibility for early closure or default.
- A water and environmental plan: Existing water commitments, seasonal and peak-day needs, cooling design, discharges, noise, backup power and parcel-specific flood resilience.
- A customer-benefit policy: Verified eligible-account count, allocation rules, treatment of any general-fund share and public reporting of actual results.
The decision standard should be net public benefit after costs, enforceable protection against losses, and acceptable site-specific impacts—not gross electricity sales or an assumed monthly credit.
Modeling note: All scenario figures are hypothetical calculations, not a developer proposal or forecast. The annual electricity calculation uses a standard 8,760-hour year. The 28,000-account example replaces the earlier unverified exact denominator with an expressly illustrative assumption. Definitive construction-cost estimates, opening dates, statutory provisions and court holdings have been omitted because the necessary underlying records have not been validated.
LodiEye is the original civic-reporting and analysis arm of Lodi411.com, a citizen-run civic data and transparency platform serving Lodi, California and San Joaquin County. LodiEye gathers information of public interest, applies editorial judgment to public records, meetings, and data, and publishes original explanatory reporting for its readers — the work of a newsroom, and a representative of the news media as that term is defined under federal law. Our reporting emphasizes primary sources, public data, and full source transparency so readers can check every claim. LodiEye complements, and does not replace, the other outlets covering this region; for additional reporting on Lodi, San Joaquin County, and the broader region, we also encourage readers to consult the Lodi News-Sentinel, Stocktonia, The Sacramento Bee, CalMatters, and other established news organizations. Our full editorial standards and news-media-status statement is published at lodi411.com/editorial-standards.
This LodiEye report was prepared with artificial intelligence tools under human editorial direction. Lodi411 uses multiple AI platforms in its research and publication workflow, including Anthropic’s Claude and Perplexity AI. The work on this article included:
Source Discovery
- Sources include Steve Mann’s November 16 column, the adopted White Slough water-board order, Berkeley Lab’s national report, county moratorium announcements, Reuters, CalMatters and a California Energy Commission project description.
Credibility Validation
- The article distinguishes opinions from findings, permitted flow from actual flow reported in the 2025 order, and county announcements from ordinances and approved minutes. Detailed statutory claims, definite costs and fixed opening dates were removed where underlying records had not been validated.
Analysis and Synthesis
- Electricity, revenue, credit and water figures are reproducible sensitivity calculations. Assumed capacity, load factor, margin, account count, PUE and WUE are labeled. They are not measured operating data, a developer’s plan or a utility forecast.
Presentation
- AI tools drafted and formatted the article, calculated the tables and prepared the Kendo UI charts and accessible tables. External links identify supporting sources.
Final Review
- This is a publication draft for the editor’s final review. No independent engineering study or legal opinion is claimed. Feasibility, parcel jurisdiction, full ordinance terms and contract protections remain matters for documentary verification.
Lodi411 believes transparency about AI use serves its readers. Readers who spot errors are encouraged to write editor@lodi411.com so corrections can be made.
Sources and calculation notes
Numbered links point to these records. Source limitations are identified below.
-
Steve Mann, “Datacenters,” About Town, November 16, 2026 (column heading: About Lodi, November 17).
Source for the proposal and attributed environmental assurance; not a project application.
-
Lawrence Berkeley National Laboratory, 2024 United States Data Center Energy Usage Report.
PUE and WUE definitions, cooling systems, and modeled 2023 U.S. WUE benchmark.
-
Reuters, “US data centers slash electric bills with brief demand cuts, raising questions over who pays,” November 16, 2026, via Yahoo Finance.
National cost-allocation context; not a White Slough project finding.
-
CalMatters, “AI data centers could hike California electricity bills,” March 4, 2026.
Reporting on Little Hoover Commission recommendations, not enacted law.
-
Central Valley Regional Water Quality Control Board, Order R5-2025-0046, White Slough Waste Discharge Requirements, 2025.
Findings 10, 16, 17, 29, 32 and 42: separate systems, flow, uses, preliminary hydrogen planning and floodplain context.
-
California Energy Commission, RB Inyokern Data Center project description.
Unrelated project illustrating the need to disclose backup-generation design.
-
San Joaquin County, initial data-center moratorium announcement, October 7, 2026.
County announcement of a 45-day pause; not the ordinance or approved minutes.
-
San Joaquin County, moratorium-extension announcement, November 17, 2026.
County announcement of an additional 10 months and 15 days; not a parcel-applicability determination.
Calculation methods
- Annual electricity: facility MW × 1,000 × 8,760 hours × electrical load factor.
- Annual net amount: annual facility kWh × assumed net dollars per kWh.
- Monthly credit: annual net amount ÷ 28,000 assumed accounts ÷ 12; equal allocation and full distribution assumed.
- Computing electricity: annual facility kWh ÷ assumed PUE.
- On-site water consumption: computing kWh × WUE in liters per kWh ÷ 3.785411784 liters per U.S. gallon. Divide by 365 for an annual-average daily figure.