Who Pays to Keep the Lights On? Data Centers, EVs, and the Grid's New Era in California and Lodi

Who Pays to Keep the Lights On? Data Centers, EVs, and the Grid's New Era in California and Lodi

LodiEye — May, 2026

Summary

After two decades of flat electricity demand, the American power grid has hit an inflection point. Artificial-intelligence data centers, a wave of electric vehicles, and a generation mix tilting toward solar and wind are changing both how much power the country needs and the architecture of the system that delivers it. The hardest question is not whether the lights stay on, but who pays to keep them on — and whether ordinary households can still afford the bill. Nowhere is that question sharper than in California, and the answer looks different depending on which side of a city limit you live on.

~134 GW

Projected U.S. data-center grid demand by 2030, up from roughly 53 GW in 2023

~30¢

Average California residential rate per kWh, about 74% above the U.S. average

14.28¢

Lodi's Tier 1 residential rate per kWh — less than half the California average

Demand wakes up after twenty flat years

For roughly two decades, U.S. electricity demand barely moved, growing at well under one percent a year as efficiency gains offset population and economic growth. That era is over, and three forces are driving the change at once.

The largest is the data center. Analysts at S&P Global's 451 Research project that U.S. data-center demand on the grid will reach about 75.8 gigawatts in 2026, climbing to roughly 108 GW in 2028 and 134 GW by 2030 — up from approximately 53 GW in 2023, driven largely by artificial intelligence.¹ Deloitte's outlook runs higher still, with data-center demand potentially reaching 176 GW by 2035.² By several estimates, data centers account for nearly half of all projected U.S. power-demand growth through the end of the decade.³

U.S. data-center electricity demand on the grid (gigawatts)

Source: S&P Global / 451 Research Datacenter Services & Infrastructure forecast (2025–2026).

The second force is transportation. As drivers switch to electric vehicles, the load they add is modest per car but enormous in aggregate — and flexible, since an EV can charge at almost any hour, which makes it either a problem or a solution depending on when it plugs in. The third is electrification of buildings and industry, as heat pumps and electric equipment replace gas. Together they have pushed utilities' five-year peak-demand forecasts up sharply — one tally found the total of utilities' published five-year growth projections jumped from 38 GW in 2023 to 128 GW in 2024.⁴

A grid built for a different era

The change is not only about more megawatts; it is about a system whose physical logic is being rewritten. Three architectural shifts matter most.

First, demand is concentrating. Data centers cluster where land, fiber, and power align, so a handful of regions absorb a disproportionate share of growth, straining specific transmission corridors rather than the system evenly. Second, the shape of supply has flipped: the grid is moving from large plants that run on demand toward solar and wind that run when the weather allows, backed by batteries that discharge for a few hours. As coal and older gas plants retire, most new capacity is variable. Deloitte estimates the U.S. will see roughly 104 GW of coal and gas retirements by 2030, offset by about 209 GW of new capacity — but only around 10 percent of those additions are firm, dispatchable power that can run whenever it is needed.² Third, the binding constraint has moved from generation to interconnection and transmission: the queue to connect new projects has become the bottleneck.

One cautionary thread runs through every serious forecast: a large share of announced data-center demand may never materialize. When Ohio regulators required data centers to pay for a portion of the capacity they request even if it goes unused, American Electric Power's Ohio pipeline shrank from more than 30 GW to about 13 GW.⁵ That collapse is the clearest evidence available that who pays determines how much of the boom is real — the question at the heart of the affordability debate.

California's high-cost squeeze

California is not, primarily, a data-center story; it ranks in the middle nationally for that load, far behind Virginia and Texas. Its affordability problem comes from enormous fixed costs recovered through the per-kilowatt-hour price. California residential rates reached about 30.3 cents per kWh in January 2026 — roughly 74 percent above the national average and second only to Hawaii.⁶ Wildfire prevention and recovery spending by the investor-owned utilities, which has run into the tens of billions, is a major driver even when a utility is not at fault,⁷ alongside distribution rebuilding, clean-energy mandates, and exposure to natural-gas prices. The economist Severin Borenstein, who chairs the board of the state's grid operator, notes that of the roughly 40 cents per kWh an average customer of the three big private utilities pays, only about 12 to 15 cents is genuinely open to competition; the rest is fixed cost that does not fall when a household uses less.⁸

Average residential electricity rate, cents per kWh

Sources: U.S. EIA-based state averages for January 2026 (national and California); City of Lodi Electric Utility residential rate schedule for Lodi tiers (2026). Tier 1 covers the first 391 kWh in winter (Nov–Apr) or 481 kWh in summer (May–Oct); Tier 3 applies above 782 kWh winter / 962 kWh summer. A state energy tax of $0.00030 per kWh is added to all usage.

On the supply side, solar has reshaped the grid. The famous "duck curve" has deepened into what analysts now call a "loon" — the California Independent System Operator regularly runs negative net load in the middle of the day, when utility-scale solar exceeds total demand.⁹ The state has answered with a historic battery build-out, from about 500 megawatts of grid storage in 2020 to more than 13 gigawatts by early 2025.¹⁰ But batteries discharge for only a few hours, so the evening ramp, heat waves, and wildfire-driven shutoffs still call for generation that can run on command.

CAISO grid-connected battery storage capacity (gigawatts)

Source: GridStatus / CAISO data, 2020–2025.

Electric vehicles cut both ways. A UC Davis study found that 67 percent of the state's distribution feeders will need upgrades by 2045 — about 25 GW of work costing between $6 billion and $20 billion — yet because EVs add electricity sales, that growth can push the per-kWh rate down by one to six cents if charging is managed to avoid the evening peak.¹¹ Whether EVs raise or lower bills depends largely on when the cars charge — a recurring lesson: new demand can either burden ratepayers or help them, depending on the rules.

The deeper dive: San Joaquin County

San Joaquin County is home to roughly 824,000 people, with Stockton as its seat.¹² Most of the county — Stockton, Manteca, Tracy, Lathrop, and the unincorporated areas — is served by Pacific Gas and Electric, so most residents are fully exposed to the statewide rate pressures above. The region is also no longer only a consumer of power made elsewhere. At the Port of Stockton, Nautilus Data Technologies runs a roughly 7-megawatt data center that cools its servers with San Joaquin River water rather than potable supply,¹³ a reminder that the demand and infrastructure trends transforming the national grid are already present in the Valley.

Lodi's municipal advantage

Lodi sits inside this county but outside PG&E's territory, and that single fact changes its exposure to the entire story above. Since 1910, the city has run its own provider, Lodi Electric Utility — a customer-owned, city-operated system serving about 27,400 accounts across a 14-square-mile territory on an annual budget just north of $100 million.¹⁴ It is a member of the Northern California Power Agency (NCPA), a joint-powers cooperative through which member cities buy and generate power together.¹⁵

The structural differences from an investor-owned utility are significant. Lodi's rates are set by the elected City Council, not the California Public Utilities Commission. The utility is non-profit and returns roughly $7.4 million a year to the city's general fund — money that supports parks, police, and fire.¹⁵ Because municipal systems do not carry the same wildfire-liability and shareholder-profit burdens that dominate the investor-owned utilities' rates, they have generally kept prices well below PG&E's. Lodi Electric's residential schedule sets the Tier 1 baseline at $0.14280 per kWh — less than half the California average and below even the U.S. average. Tier 2 rises to $0.15810 per kWh for usage above the seasonal baseline of 391 kWh in winter (November–April) or 481 kWh in summer (May–October), and a high-usage Tier 3 rate of $0.33660 per kWh applies above 782 kWh (winter) or 962 kWh (summer), designed to encourage conservation. A state energy tax of $0.00030 per kWh is added to all usage. For households on baseline use — most of them, most of the year — the savings versus an investor-owned-utility customer just outside city limits are substantial.¹⁴

What a municipal utility changes

A city-owned utility is governed locally and exempt from CPUC rate regulation, which gives Lodi direct control over rate design and over the terms it offers any large new customer. It also concentrates accountability: the people who set the rates are the people residents elect. The trade-off is scale — a 14-square-mile system has less buying power than a statewide utility — which is why Lodi pools resources through NCPA and guards its credit rating to keep wholesale power costs low. The advantage is real but not unlimited: after nearly a decade without a rate increase, the City Council in February 2026 adopted an updated Electric Utility fee schedule with authorized annual, cost-based adjustments,¹⁶ a sign that municipal systems still feel the same rising wholesale-power and capital pressures.

A new high-voltage backbone for Lodi

The structural advantages above are only as good as the wires that deliver the power, and Lodi's existing connection to the bulk grid has been running out of room. As far back as its 2012–2013 Transmission Planning Process, the California Independent System Operator identified five PG&E 60 kV lines between the Lockeford and Lodi substations as suffering thermal overloads and high voltage deviations — serious enough to flag North American Electric Reliability Corporation (NERC) compliance issues — and the 2017–2018 planning cycle reaffirmed the need to fix them.²⁸

The fix is the Northern San Joaquin 230 kV Transmission Project, a roughly 10.6-mile build of new double-circuit 230 kV transmission line that loops PG&E's existing Brighton–Bellota 230 kV corridor through Lockeford Substation and runs a new line to a new PG&E Thurman Switching Station on the eastern edge of Lodi.²⁹ Lodi Electric Utility will build a new 230/60 kV substation of its own — the LEU Guild Substation — between the Thurman station and LEU's existing Fred M. Reid Industrial Substation, with transformers stepping the 230 kV down to 60 kV for distribution across the city's system.²⁸ Once the Guild Substation is operating, PG&E will disconnect its 60 kV system from LEU's, giving Lodi a cleaner, dedicated tie into the bulk transmission grid. The city's share of the project is approximately $30 million,²⁷ with PG&E's CPCN application filed with the California Public Utilities Commission in September 2023, the final environmental impact report issued in June 2025, construction targeted for 2026–2027, and operation expected by 2029.³⁰

What the 230 kV upgrade buys Lodi

A higher-capacity connection to CAISO's bulk transmission grid through the Brighton–Bellota corridor; resolution of long-standing thermal overloads and the NERC compliance issues that came with them; reliable headroom for summer peak demand, residential and commercial growth, and electrification; independence of LEU's 60 kV distribution from PG&E's 60 kV system; and tighter integration with NCPA's wholesale power and the Lodi Energy Center across the regional grid. This is the local mirror image of the national transmission bottleneck the new demand era has exposed — exactly the kind of unglamorous but indispensable infrastructure investment that determines whether a utility can deliver on its affordability and reliability promises.

Lodi's other asset: generation at White Slough

Lodi's position is unusual in a second way: it helps own the power plant in its own backyard. Beside the city's White Slough Water Pollution Control Facility sits the Lodi Energy Center, a roughly 300-megawatt (nominally 296 MW) combined-cycle natural-gas plant that NCPA opened in 2012 and operates on behalf of nine member agencies and four other public entities.¹⁷ NCPA describes it as one of the cleanest and most efficient gas-fired systems in the country and the first in the nation to use "fast-start" technology — the ability to ramp generation up and down quickly to counterbalance variable wind and solar.¹⁵ That is precisely the dispatchable capability California is short of, and the plant pairs it with a water-energy synergy, using the city's treated wastewater for cooling rather than drawing down drinking-water supplies.¹⁵

The plant is also a step into the future fuel debate. After its original turbine failed in 2020, NCPA replaced it with a Siemens Energy model engineered to burn up to a 45 percent hydrogen blend — far above the roughly 11 percent blended at a Hawaii plant or the 20 percent common internationally.¹⁸ Burning hydrogen in place of part of the natural gas lowers emissions while preserving on-demand reliability.

Hydrogen share a power plant can blend with natural gas (percent)

Sources: NCPA and City of Lodi materials; Lodi News-Sentinel. The Lodi Energy Center's turbine is rated for up to a 45% blend today; the 100% figure was a project roadmap target whose funding is now in question.

The hydrogen vision and a $35 million setback

NCPA, the City of Lodi, Siemens Energy, PG&E, and partners proposed a Lodi Hydrogen Center at the site: an electrolyzer producing roughly 24 tons of clean hydrogen a day from recycled wastewater and renewable electricity, with a roadmap that once targeted running the plant on up to 100 percent hydrogen by 2028.¹⁹ The project was a Tier I component of ARCHES, California's $12.6 billion statewide hydrogen hub backed by up to $1.2 billion in federal funds.²⁰

In October 2025, the U.S. Department of Energy canceled the nearly $35 million in federal funding earmarked for the Lodi project, part of a roughly $8 billion rollback across 223 energy projects nationwide; DOE later moved to eliminate the entire $1.2 billion ARCHES award.²¹²² Backers had pitched the Lodi plant as lowering energy bills, creating more than 200 jobs, cleaning the air, and boosting grid reliability, with the Port of Oakland interested in hydrogen-fueled trucks; it would have been Northern California's only facility of its kind.²¹ The plant's existing 45 percent blend capability stands, but on-site production is on hold. Representative Josh Harder and Senators Alex Padilla and Adam Schiff have pressed to restore the funds, with critics noting the canceled projects fell overwhelmingly in states that did not vote for the president in 2024.²¹²³

What it would take to keep power adequate and affordable

The pivotal insight ties the whole story together: rates and bills move differently, and new demand can actually lower everyone's rates by spreading fixed costs over more kilowatt-hours — but only if that new demand pays its own way. Otherwise it becomes a subsidy from households to the new load. Cost allocation, in other words, is the whole ballgame, and a practical agenda follows from it.

A practical agenda for residential affordability

1. Make large new loads pay their full cost. The strongest protection for households is a large-load tariff that requires data centers and other big users to fund the transmission and generation they need, with minimum-use commitments and exit fees if projects stall. Pennsylvania adopted a statewide model tariff in 2026, Texas requires large loads to share in shortage curtailment, and California's Senate Bill 57 directs the CPUC to determine by January 2027 whether new data-center loads shift costs onto other customers.²⁴ A 2026 White House pledge had major hyperscalers commit to building or buying their own new generation, on the stated principle that the public should not foot the bill.²⁵

2. Reform how fixed costs are recovered. Recovering wildfire and distribution costs partly through fixed charges, with lower per-kWh prices, makes electrification cheaper and rates fairer — though the design, including income-graduated charges, is genuinely contested.

3. Invest in flexibility, not just capacity. Storage, demand response, virtual power plants aggregating home batteries, and managed EV charging that shifts load into the solar-rich midday are usually cheaper than new plants and directly tame the evening ramp that drives peak costs.

4. Fix transmission and interconnection — the true bottleneck — including grid-enhancing technologies that wring more capacity from existing lines at a fraction of new-build cost.

5. Keep firm, clean supply in the mix. California reversed course in April 2026, with federal regulators renewing the Diablo Canyon license for another two decades; the plant supplies nearly a fifth of the state's clean electricity.²⁶ Dispatchable, lower-carbon generation — from nuclear to fast-start gas moving toward hydrogen — fills the gap batteries cannot.

The demand inflection is real and permanent, but rising household bills are not inevitable; the outcome turns almost entirely on who pays for the new infrastructure. Most of San Joaquin County will have that question answered by a state commission and an investor-owned utility. Lodi, with a non-profit municipal utility that answers to local voters, a new high-voltage transmission backbone now under permit, and a co-owned dispatchable power plant on its own land, holds more of the answer itself — an advantage worth using deliberately as the grid's new era arrives.

About This Report

LodiEye is the investigative research arm of Lodi411.com, a citizen-run civic data and transparency platform serving Lodi, California and San Joaquin County. LodiEye is not a traditional news outlet. It does not employ professional journalists or reporters, and the people behind it do not hold journalism degrees or have professional newsroom experience. LodiEye is best understood as civic research and analysis — not peer journalism — and is not a substitute for the local and regional news organizations that do this work professionally. For traditional reporting on Lodi, San Joaquin County, and the broader region, readers are encouraged to consult the Lodi News-Sentinel, Stocktonia, The Sacramento Bee, CalMatters, and other established news outlets staffed by credentialed journalists.

This LodiEye analysis was produced using artificial intelligence tools under the direction and review of the founder. Lodi411 uses multiple AI platforms in its research and publication workflow, including Anthropic's Claude (primarily Opus and Sonnet models) and Perplexity AI across a variety of large language models offered by each. These tools were used in the following capacities:

Source Discovery: AI-assisted search and retrieval identified national grid-outlook material from S&P Global, Deloitte, the U.S. Energy Information Administration, and the World Resources Institute; California rate, wildfire, and grid-operator data from the California Public Utilities Commission, CAISO, GridStatus, CalMatters, and the UC Berkeley Haas Energy Institute; San Joaquin County and Lodi-specific records from the City of Lodi, the Northern California Power Agency, the California Energy Commission, the Lodi News-Sentinel, Stocktonia, CBS News Sacramento, the office of U.S. Senator Alex Padilla, and earlier LodiEye coverage. Perplexity AI was used for initial source discovery and real-time data retrieval; Claude was used for deeper analysis of identified sources.

Credibility Validation: AI cross-referenced claims across multiple independent sources, prioritizing primary materials — CPUC filings and environmental impact reports, NCPA documents, official city of Lodi rate schedules, EIA data, federal agency announcements, and on-the-record statements from officials — over secondary commentary. Multiple AI models independently checked key figures (data-center demand projections, California rate levels, Lodi tier rates, the $35 million federal cancellation, and the 230 kV project's scope and timeline) for consistency across sources.

Analysis and Synthesis: Claude Opus and Sonnet assisted in connecting the national demand and architecture shift to California's distinctive fixed-cost rate structure, to San Joaquin County's PG&E exposure, and to Lodi's three structural assets — the municipal utility, the new high-voltage transmission backbone, and the co-owned generation hub at White Slough — under a single cost-allocation thesis ("who pays") that runs from the AEP Ohio anecdote through the closing agenda.

Presentation: Claude assisted in drafting the article text, building the four inline Kendo UI charts (data-center demand growth, comparative residential rates including Lodi's tiered schedule, CAISO storage growth, and hydrogen blend capability), structuring the section flow from national to local, and formatting the document to the LodiEye HTML specification.

Final Review: Multiple AI models reviewed the completed draft for factual consistency, source attribution accuracy, balanced presentation of contested points (rooftop-solar cost shift, partisan characterization of federal funding cuts, the value of dispatchable gas as a bridge fuel), and adherence to LodiEye's editorial conventions on titles, attribution, and no-self-promotion.

Lodi411/LodiEye believes transparency about AI use serves both readers and the broader information ecosystem. Readers who spot errors are encouraged to write editor@lodi411.com so corrections can be made.

References

1. S&P Global / 451 Research, "Data center grid-power demand to rise 22% in 2025, nearly triple by 2030" (2025); Data Center Dynamics summary (2026). spglobal.com
2. Deloitte, "2026 Power and Utilities Industry Outlook" (peak demand, retirements, firm-baseload share). deloitte.com
3. KilowattLogic analysis of S&P / 451 Research data (2026). kilowattlogic.com
4. World Resources Institute, citing Grid Strategies, on rising utility five-year peak-demand forecasts (2025). wri.org
5. S&P Global / Data Center Dynamics, on the AEP Ohio data-center pipeline reduction following a new tariff (2025–2026). datacenterdynamics.com
6. EcoFlow / U.S. EIA, "U.S. Electricity Rates by State: 2026" (April 2026). ecoflow.com
7. Straight Arrow News, "Wildfires and a 'black box' of utility spending drive California's record electric rate hikes" (December 2025). san.com
8. Severin Borenstein, UC Berkeley Haas Energy Institute / CalMatters (January 2026). haas.berkeley.edu
9. GridStatus, "In CAISO, Solar Generation Jumps Again While Batteries Reshape Demand" (2025). gridstatus.io
10. Yes Energy, "The Duck Curve Explained" (February 2026); EticaAG CAISO battery analysis (April 2026). yesenergy.com
11. Y. Li and A. Jenn, "Impact of electric vehicle charging demand on power distribution grid congestion," PNAS (2024). pnas.org
12. U.S. Census / Wikipedia, "San Joaquin County, California" (2025 estimate). wikipedia.org
13. Site Selection magazine (2021); Nautilus Data Technologies, Stockton facility. nautilusdt.com
14. City of Lodi, Electric Utility (residential rate schedule and accounts/territory). lodi.gov
15. Northern California Power Agency, "Lodi Electric Utility" and "Lodi Energy Center" (general-fund contribution; fast-start technology; wastewater cooling; participants). ncpa.com
16. Lodi411, "Lodi City Council Meeting — February 18, 2026" (updated Electric Utility fee schedule with annual cost-based adjustments). lodi411.com
17. California Energy Commission, "Lodi Energy Center" (capacity and configuration). energy.ca.gov
18. American Public Power Association, "NCPA plans hydrogen-fueled power plant" (turbine replacement and 45% blend); Lodi News-Sentinel, "Lodi to be base for hydrogen pilot program" (2022). publicpower.org
19. American Public Power Association, "APPA grant helps NCPA analyze path to clean hydrogen production facility" (Lodi Hydrogen Center electrolyzer; Siemens roadmap toward 100% hydrogen). publicpower.org
20. ARCHES / Office of the Governor of California, "California launches world-leading Hydrogen Hub" ($12.6 billion agreement, up to $1.2 billion federal). gov.ca.gov
21. CBS News Sacramento, "Federal funding cut from massive Lodi clean energy project" (nearly $35 million canceled; jobs, bills, air-quality claims; project on hold; Rep. Josh Harder and Mayor Pro Tempore Mikey Hothi). cbsnews.com
22. Stocktonia News, "Trump officials ax job training, hydrogen plant projects in San Joaquin County" (October 9, 2025). stocktonia.org
23. Office of U.S. Senator Alex Padilla, "Padilla, Schiff Slam Unlawful Elimination of Federal Funding for ARCHES Hydrogen Hub". padilla.senate.gov
24. DSIRE Insight, "U.S. Data Center Gold Rush Drives Surge in New Utility Tariffs" (April 2026); Pennsylvania PUC model large-load tariff order (2026); California Senate Bill 57. dsireinsight.com
25. American Affairs Journal, "How Will Data Centers Pay for Power?" (May 2026), on the 2026 Ratepayer Protection Pledge. americanaffairsjournal.org
26. PG&E Corporation, Q1 2026 results (Diablo Canyon license renewal). sec.gov
27. City of Lodi, "Northern San Joaquin 230 kV Transmission Project" (project overview and approximately $30 million Lodi portion). lodi.gov
28. California Public Utilities Commission, Final Environmental Impact Report, Northern San Joaquin 230 kV Transmission Project (June 2025), including the CAISO 2012–13 and 2017–18 Transmission Planning Process identification of overloads, voltage deviations, and NERC compliance issues, and the LEU Guild Substation design. ia.cpuc.ca.gov
29. Pacific Gas and Electric Company, "Northern San Joaquin 230 kV Transmission Project" (CPCN application filed September 2023; project scope and Thurman Switching Station). pge.com
30. Lodi411, "Lodi's Electrical Capacity for Summer Heat and Growth Plans" (construction targeted for 2026–2027, operation by 2029). lodi411.com