Skip to main content
    News & Trends

    Why Is Erin Brockovich Targeting AI Data Center Secrecy?

    Mark Debson

    Mark Debson

    Author

    Why Is Erin Brockovich Targeting AI Data Center Secrecy?Save

    Quick Answer

    Erin Brockovich launched a public data center transparency campaign in late April 2026, publishing an interactive crowdsourced map at brockovichdatacenter.com and an essay on her Substack, The Brockovich Report, on April 28.

    Within the first week, residents filed more than 1,600 community reports about local hyperscale data center projects. By late May the campaign had hardened into a specific demand: end the municipal nondisclosure agreements and post hoc permit announcements that local governments are using to greenlight AI infrastructure.

    Brockovich is explicit that the campaign is not anti AI. It is a procedural fight about public notice, drinking water, electricity rates, and night time noise from industrial chillers.

    How the Campaign Started

    Brockovich is best known for the Pacific Gas and Electric chromium-6 case in Hinkley, California, dramatized in the 2000 Steven Soderbergh film. For most of the past decade she has run a paid Substack newsletter, The Brockovich Report, co-written with public health journalist Suzanne Boothby.

    In April 2026 the newsletter pivoted to data centers after a flood of reader email. On April 28 she published "The New Pollution Is Data, And It is Coming to a Town Near You," announcing a self reporting map at brockovichdatacenter.com and asking residents to log nearby data center projects. The follow up post on May 27, "If Data Centers Are So Great, Why Are They Being Built in Secret?", reframed the campaign around transparency rather than opposition.

    What the Crowdsourced Map Actually Tracks

    The map at brockovichdatacenter.com is a self reporting tool, not a regulatory database. Residents fill in a form with the site location, the developer if known, and what they have observed locally. The site then plots reports on a US map and shares aggregate findings.

    Within seven days of launch Brockovich posted preliminary results on Substack: more than 1,600 community reports filed, with one striking finding that only 8 percent of respondents said local officials had publicly disclosed a project before construction permits were already signed.

    • Self reported data, not engineering audits. Useful for spotting patterns of secrecy, not for measuring grid load.
    • Open submissions. Anyone can file; the team manually reviews entries before publishing them.
    • Aggregate context. Reports are paired with public data on water rights, utility tariffs, and zoning changes where available.

    Why Procedural Secrecy Is the Real Target

    Brockovich has been careful to separate her argument from a blanket attack on artificial intelligence. The May 27 essay frames the problem as a governance failure rather than a technology question. Three patterns keep showing up in the submissions:

    1. Project secured before notice. Residents only learn about a hyperscale campus after rezoning votes, water permits, and tax abatements are already approved.
    2. Nondisclosure agreements at the municipal level. City councils and county economic development boards sign NDAs with hyperscalers under code names, blocking public records requests until permits are final.
    3. Unreturned phone calls. Developers route community questions through PR firms after construction starts, with no real avenue for site specific concerns.

    The political ask is therefore narrow: drop the NDAs, push notice earlier in the permitting process, and require utility impact statements before zoning is finalized.

    The Underlying Material Concerns

    The complaints aggregated by the map cluster around three measurable impacts that local governments are not currently required to publish before approval.

    Electricity demand. A modern hyperscale AI campus can draw 100 to 500 megawatts continuously, comparable to a midsize city. The US Department of Energy projects data center electricity use will roughly double between 2023 and 2028, reaching 6.7 to 12 percent of total US electricity. Residential ratepayers in PJM, the grid serving much of the mid Atlantic, have already seen capacity charges spike as new data center load is added.

    Water for cooling. Evaporative cooling systems used at many AI training campuses can consume hundreds of thousands of gallons of potable water per day per site. In water stressed counties this lands directly on aquifers that also supply farms and homes.

    Noise from industrial chillers. Roof mounted cooling fans run 24 hours a day. Neighbors within a half mile routinely report a steady low frequency hum that does not appear on standard zoning noise charts because zoning typically measures peak decibel events, not continuous broadband sound.

    How Developers and Local Officials Are Responding

    Several hyperscalers have publicly committed to expanded water disclosure and to local benefits agreements in 2024 and 2025. The Brockovich campaign is pushing those voluntary commitments into the permitting process itself.

    A handful of states have already moved in that direction. Virginia, the largest data center market in the world, passed legislation in 2024 requiring localities to publish project applications earlier and to study cumulative grid impact. Georgia and Texas have introduced similar bills in 2026. The Brockovich map is being cited in state level testimony as evidence that the problem is national rather than confined to a few northern Virginia counties.

    What to Watch Next

    Three signals will tell you whether the campaign translates into policy:

    • State preemption fights. Watch whether state legislatures move to preempt local NDAs entirely, or whether hyperscalers push back through industry groups.
    • FERC and grid operator filings. Capacity auctions at PJM and ERCOT will reveal how much of the residential rate spike traces directly to data center load.
    • The 2028 sunset of federal AI incentives. Tax credits tied to domestic compute infrastructure expire in 2028; that deadline is driving the current rush of secret site selection.

    The Takeaway

    Erin Brockovich is not running an anti AI campaign. She is running a public records campaign aimed at the contractual machinery that lets hyperscalers and local officials approve multi hundred megawatt facilities before residents know a project exists. The crowdsourced map at brockovichdatacenter.com is the lever, and the 1,600 reports filed in the first week show the demand for that visibility was already there. Whether the campaign reshapes permitting will depend on state level laws on early notice, NDA bans, and mandatory utility impact studies over the next 12 months.

    Mark Debson

    Written by

    Mark Debson

    I'm Mark Debson, the writer behind dmbio. I spend my days digging into the science behind everyday products, brands and habits, then translating what I find into clear answers you can read in about five minutes.

    Drafted with AI assistance, fully reviewed and edited before publishing. See our editorial & AI policy.

    Related reads