Solar canals, green hydrogen, and co-located demand. Sharpening the infrastructure case for the people who decide whether it lives or dies.
Sources: CAISO, EIA, PG&E, DOE, ARCHES, UC Merced / CSCI, USDA, ICCT
Build solar on land the state already owns, canals and aqueducts and parking lots, funded by revenue bonds. A self-financing corridor that makes power, saves water, and absorbs curtailed energy through co-located demand.
Project Nexus in Turlock is fully operational at 1.6 MW across narrow- and wide-span canal sections. UC Merced confirms the expected power output, reduced evaporation, and suppressed weed growth. A seven-university consortium led by USC is scaling it statewide.
No longer theoreticalCAISO curtailed 3.4 TWh in 2024, up 29%. Solar is 93% of it, and over 70% is transmission congestion, not oversupply. Path 15 through the Central Valley faces projected congestion in 84% of hours by 2039. The energy is stranded without local demand.
A problem that compoundsCyrusOne and Ameresco are building a 100 MW data center at NAS Lemoore, the first major Central Valley facility. PG&E's pipeline is 10 GW. Every 1 GW of data center load saves existing ratepayers 1 to 2% on monthly bills.
Market ahead of policySolar canals save 63 billion gallons of water a year, enough for 2 million people. They suppress canal weeds for $69M in annual savings, preserve 80,000 acres of farmland, and build jobs in the counties with the highest poverty rates.
No rival project does all of thisFive weak spots that needed answers before this reached a committee hearing. Each one is now closed.
| Cost category | Canal-top solar | Ground-mount + T&D | Edge |
|---|---|---|---|
| Panel + structure | $2.00–4.00/W | $0.94–1.01/W | Ground |
| Land acquisition | $0, state-owned | $150K–2.5M/acre | Canal |
| Water savings | +$69M/yr | $0 | Canal |
| Transmission | $0, local demand | $700M–1.1B | Canal |
| Canal maintenance | +$40K/mile/yr saved | $0 | Canal |
| Permitting / NIMBY | State right-of-way | 75 projects vetoed | Canal |
The argument is not "cheaper solar." It is cheaper delivered energy, once you count the costs ground-mount quietly shifts onto ratepayers.
100 MW array plus storage, Fresno County. $250–400M capital via tax-exempt revenue bonds at a 3.5–4.5% coupon, 30-year term. Electricity, grid services, and water savings cover debt at 1.1 to 1.3x. "No taxpayer dollars" stops being a slogan and becomes a number.
ClosedHydrogen economics were aspirational, so lead with data centers and treat hydrogen as Phase 2. Literal blanks in the source documents are all filled. And the missing political map, who sponsors and who fights, is now drawn.
Closed"This puts construction jobs and data center tax revenue in your district while cleaning the air your constituents breathe." Target Assembly Water, Parks & Wildlife and Senate Energy, Utilities & Communications.
"Your canals become revenue-generating assets. The bonds service themselves from products your own growers buy." Start with Turlock Irrigation District, already hosting the pilot.
"13 GW of solar with built-in demand. No new transmission, no rate increase. It solves curtailment while fixing water." Lever: the market is already building at Lemoore.
"Solar canals are 25% cheaper." Unsubstantiated, and it invites the fight you lose.
"The total delivered cost is lower." No land, no transmission, water savings, local demand.
"No taxpayer dollars." It needs a pro forma to be credible.
"Self-financing through iBank bonds, at 1.1 to 1.3x coverage." Cite the vehicle, show the math.
Commission a system-cost comparison. Canal-top solar plus storage against ground-mount plus new transmission. Land, water, maintenance, avoided T&D. The single most important missing piece.
Build the revenue bond pro forma with iBank. Bond size, coupon, the PPA price needed to service debt. Without it, "no tax dollars" is a slogan.
Map the political coalition. A legislative champion, water-district co-signers, and which utilities to neutralize versus fight.
Lead with data centers, not hydrogen. The 10 GW pipeline is real demand today. Combining both dilutes the strongest argument.
How solar canal surplus, canal water, and $60B in agricultural demand make a hydrogen plant pencil. Even without the federal funding that just disappeared.
Sources: DOE Hydrogen Program, ARCHES H2, CAISO, CA Energy Commission, USDA, ICCT
A hydrogen plant on a solar canal does not need one buyer. It has three.
Class 8 long-haul cannot go battery-electric without killing payload. Fuel cells refuel in 15 minutes with 300+ mile range. The I-5/99 corridor is the state's heaviest freight route, with $102M federal infrastructure already awarded.
Frame it as the corridor nicheCalifornia agriculture runs on imported ammonia at gas-linked prices that spiked to $942/tonne by March 2026. A local green ammonia plant beside a canal electrolyzer sells fertilizer directly to the irrigation districts. This is what makes the bonds bankable.
Replacing an import with a local productElectrolyzers are flexible load: ramp up when solar peaks and curtailment hits. An electrolyzer running on curtailed power gets paid twice, near-zero electricity plus curtailment-absorption credits. The hydrogen itself is long-duration storage.
Paid twice for the same hourA solar canal site beats every other location on the four inputs that decide hydrogen economics.
Near-zero marginal cost during curtailment, which hits 54% of Fresno hours. Power is 60 to 80% of production cost. This is the single biggest lever.
Canal water is right there. No aquifer pumping, no new water rights. Electrolysis needs 9 to 10 L/kg, and proximity removes the sourcing problem that kills other sites.
State-owned right-of-way beside the canal. No acquisition, no ag conversion, no NIMBY fight.
None needed. The plant consumes power locally, so it bypasses Path 15 congestion entirely. It is the demand.
"The federal government walked away from California hydrogen. Revenue bonds let the state build it anyway."
20 MW PEM electrolyzer beside the aqueduct in Fresno County. Roughly 2,900 tonnes of hydrogen and 15,000 tonnes of green ammonia a year. The pitch is not "hydrogen is the future." It is "this plant pays for itself."
| Line | Item | Figure |
|---|---|---|
| CAPEX | Electrolyzer, ammonia unit, purification, site | $45–71M |
| OPEX | Electricity, water, maintenance, 15–20 FTE | $4.6–8.6M/yr |
| Revenue | H2 fuel (~2,900 t @ $4–5/kg) | $11.6–14.5M/yr |
| Revenue | Green ammonia (~15K t @ $400–600/t) | $6–9M/yr |
| Revenue | Grid services, curtailment credits | $1–3M/yr |
| Total | Annual revenue | $18.6–26.5M/yr |
Standalone coverage of 2.0 to 3.4x. Illustrative, and a municipal finance advisor should validate it, but it proves the model holds without a federal dollar.
How batteries and microgrids are replacing power lines from the bottom up, and why the current $45B transmission plan overbuilds the wrong thing.
Sources: Brattle Group, PG&E, CAISO, FERC, CPUC, SDG&E, Blue Lake Rancheria, EIA
December 2025: a fire at one PG&E substation knocked out 30% of San Francisco's power for three days. One chokepoint, one city dark. A microgrid there would have islanded and kept running.
A $45.8 to $63.2B CAISO plan over 10 to 15 years. Ratepayers absorb all cost and risk while utilities earn 9.98% on every dollar of capital. Rates are already the highest in the country.
14.2 GW of batteries already installed. Community microgrids deploy in 1 to 2 years. No single point of failure. The argument is not "zero transmission," it is "stop overbuilding it while starving the faster, cheaper option."
July 2025: PG&E and the Brattle Group ran the largest virtual power plant test in US history. Home batteries, coordinated by software, matched a gas peaker.
"The VPP delivered sustained, flat output comparable to a peaker plant, at a fraction of the cost, with zero emissions, and no new transmission required."
NEM 3.0 cut rooftop export credits 75% to slow residential solar. Instead it forced homeowners to pair solar with batteries, and attachment rates jumped from 11% to over 50%. The Court of Appeal upheld it in March 2026, making the battery pivot permanent. The utility that tried to kill solar built the distributed grid instead.
Solar plus Tesla Megapack. Ran through the 2019 shutoffs as an emergency hub for 10,000+ residents while neighbors went dark. Now anchoring a $177M tribal microgrid expansion.
2,800 customers at the end of a single line. SDG&E's automated microgrid islands the community when the main line fails. The only utility-owned community microgrid of its kind in the US.
A 5 MW community microgrid replacing the city's last fossil plant with local solar and storage, in one of San Francisco's most pollution-burdened neighborhoods.
Nine community microgrids across Northern California, a $200M program covering nearly 9,000 customers, 3,600 of them with access and functional needs.
We do not need more power lines. 14.2 GW installed, 535 MW dispatched from home batteries in one test. The distributed grid exists; policy just has to catch up.
Towns used to own their grids. Blue Lake stayed lit while PG&E customers went dark. The technology to bring local control back is operational today.
The utilities killed solar and built a battery market. Export credits cut 75%, attachment rates up to 50%, and the courts just made it permanent.
Renters deserve solar too. Balcony solar passed committee 12 to 0. No permits, no contractors. 44% of Californians rent.