What happened in Santa Clara? The precedent everyone's quoting.

Santa Clara, California is a city of about 130,000 on the south edge of San Francisco Bay. Over roughly three decades, it became the single most data-center-dense municipality in the United States. By 2024, more than fifty data centers operated inside its 18 square miles, consuming roughly 60 percent of all electricity sold by the city utility. The build-out did not happen by accident, and it did not happen quickly. Understanding how Santa Clara got there — and what residents saw along the way — is useful for any Texas county now at the beginning of the same conversation.

A short history

Santa Clara is unusual among California cities in that it runs its own electric utility, Silicon Valley Power (SVP). This is a publicly-owned, not-for-profit municipal utility — not PG&E. That single fact shapes the rest of the story. Because the city owns the electrons, it also owns the infrastructure and the cost base. Every major decision about rates, transmission, and capacity connects back to the city council, not a state commission.

Data-center development in Santa Clara began in the early 2000s on the old industrial parcels that had been home to chip-fabrication plants in earlier decades. When fabs moved offshore, their buildings and electrical service remained. Hyperscale operators recognized that the combination of existing heavy power feeds and low municipal rates was unusually attractive.^[1]

Between 2005 and 2020, the city's data-center footprint grew from a handful of early facilities to more than fifty. By 2024, the cumulative electric load of data centers in Santa Clara was estimated at around 800 megawatts.^[2] The city's residential load, by comparison, is only a fraction of that.

What was promised

At each step of the build-out, the cases made to the council and the public were consistent. They are worth noting because the same arguments appear almost verbatim in Texas hearings today.

• Tax base. Industrial real estate generates property tax revenue, utility users' tax, and utility gross receipts without the school-enrollment and road-wear costs of residential development.
• Jobs. Construction employment during the build phase; smaller but high-wage operations and engineering jobs after.
• Load diversification. Big commercial customers were said to stabilize the utility's rate base and help subsidize residential service.
• Reliability investment. Data centers would pay for substation upgrades and redundant feeds that would benefit everyone else on the grid.

Each of these arguments was, in its narrow form, partly true. The more interesting question was which ones held up at scale.

What actually happened

Electricity rates

Santa Clara residents still pay electric rates that are among the lowest in California — in 2024, roughly half of what PG&E charged a nearby homeowner for the same kilowatt-hour.^[3] That is a genuine and measurable benefit of the municipal model, not unique to the data-center era.

But the margin has tightened. As data centers moved from roughly 20 percent of sales in the 2010s to roughly 60 percent by the mid-2020s, the city undertook a series of rate studies and concluded that large-customer allocation needed adjustment.^[4] Specifically, transmission-upgrade costs were being shared across all ratepayers, even though some of those upgrades were driven by new large loads. Beginning in 2023, the city began negotiating higher demand charges for new data-center customers and, in some cases, direct-paid transmission recovery. That debate is ongoing.

Property values

Residential property values in Santa Clara rose substantially over the build-out period, but that rise tracks the broader Silicon Valley housing market. It is not clear how much was attributable to data-center activity specifically; most of it reflects regional tech-industry demand. Homes immediately adjacent to data-center parcels have shown more variable price behavior — sometimes trading at a small discount to comparable homes farther away, reflecting concerns about industrial noise, traffic, and view corridors.^[5]

Water

Santa Clara receives most of its water from the Santa Clara Valley Water District, which draws from surface reservoirs, imported water, and groundwater. In 2021, the district declared a drought emergency as groundwater levels fell to historic lows. Commercial water use, including evaporative cooling at data centers, was not the largest category of demand — but it was a visible and growing one.^[6]

The city and the water district responded by investing in recycled-water infrastructure. New data-center campuses beginning in the early 2020s were required to connect to recycled-water lines for cooling, rather than using potable municipal supply. This did not reduce total water demand, but it did separate cooling use from residential drinking supply.

Tax revenue

Direct property tax collections from data-center parcels have been substantial and largely as advertised. Utility users' tax and utility gross-receipts tax are more complex. Because data centers are large electricity customers, the city's utility revenue rose with the build-out; some of that revenue funds the general fund, which in turn pays for parks, police, and general services. A 2022 grand-jury report on Santa Clara's utility finances flagged that the scale of data-center revenue made the city unusually dependent on a single industry category.^[7]

The 2022 settlement structure

In 2022, after several years of public debate, the Santa Clara city council adopted a set of policy changes that have since been widely cited as a template for other jurisdictions.^[8] The main elements:

• Higher demand charges on new data-center customers, and a formal cost-of-service study to inform future rate-setting.
• Recycled-water connection requirements for new campuses, funded in part by the operators.
• Standardized disclosure of facility-level water and power use as part of the development-approval process, making public records more usable.
• Transmission cost recovery tied more directly to the customers driving new infrastructure.

The settlement did not ban new data centers, and it did not solve every dispute. Some residents argued it was too lenient. Some in industry argued it chilled investment. Both reactions are in the public record.^[9]

Santa Clara is not a warning and not a success story. It is a case study in what happens when a community lets a single industry become the majority of its electric load before it decides how to allocate the costs.

Why Texas is structurally different

A handful of Texas cities — Austin, San Antonio (CPS Energy), New Braunfels, Denton, Garland — own their electric utilities, and for those residents the Santa Clara template is directly relevant. The rest of Texas is different. Most counties are served by investor-owned utilities under ERCOT, which is overseen by the Public Utility Commission of Texas (PUCT). The feedback loop between a resident complaint and a rate adjustment is longer and more formalized in Texas than it was in Santa Clara.^[10]

That matters in two ways. First, in Texas the cost-allocation conversation happens primarily at the PUCT and at the Legislature — most visibly in 2025's Senate Bill 6, which addressed large-load interconnection rules. Second, residents concerned about their power bills have fewer direct levers than Santa Clara residents had. County commission decisions about zoning and abatements remain local; rate decisions do not.

What a Texas county can actually learn from Santa Clara

Setting aside the Texas-specific structure, five observations from Santa Clara have been repeatedly cited by municipal finance staff and utility analysts, and they hold up as generally useful:

1. Concentration risk is real. A jurisdiction where a single industry becomes a majority of its tax or utility base faces unusual sensitivity to that industry's decisions.
2. Cost-allocation policy is easier to set before the build-out than after. Once the infrastructure is in place, the bargaining position shifts.
3. Water separation matters. Mandating recycled or non-potable supply for cooling is mechanically straightforward and politically low-cost before the first connection is permitted.
4. Disclosure is cheap and cumulatively useful. Facility-level water and power figures filed as part of land-use approvals accumulate into data residents can work with over time.
5. Property values near specific parcels are more variable than city-wide averages suggest. Residents in the immediate vicinity should be consulted separately from the broader community.