On October 15, 2026, Harris County Attorney Abbie Kamin stood with Commissioners Lesley Briones and Adrian Garcia and did something unusual for a county official: she read a utility’s investor presentation back at it, as evidence. The numbers were CenterPoint Energy’s own — about 8 gigawatts of new data-center load expected to be energized in the Houston area by 2029, with another 4 gigawatts or so to follow, a trajectory the company has characterized as roughly 50 percent load growth by 2029. Kamin’s point was not that the forecast is wrong. Her point was that a county with no power to shape it wants some.
The Commissioners Court that day voted to direct the county engineer, the flood control district, and the fire marshal to flag major data-center permit applications to commissioners, and to draft statutory asks for the Texas legislative session that convenes in January 2027. County officials also want an attorney general’s opinion on what counties can do at all. The ask, in Kamin’s framing: give counties authority over what connects, and where.
This is the siting fight becoming a governance fight. The desk’s Texas coverage has traced the physical and procedural story — the 474-gigawatt tracked-load tally, the statewide audit, Senate Bill 6’s financial-security rules. Harris County is asking who decides: a regulated utility’s ten-year capital plan assumes the load arrives, while the county where much of it would land wants a hand on the throttle.
The forecast, read aloud
Start with what Kamin quoted, because the arithmetic matters and the attribution matters more. The 8-gigawatt figure traces to CenterPoint’s first-quarter 2026 earnings presentation in April, as reported by Utility Dive: 12.2 gigawatts of “firmly committed” new load in the Houston area, 3.5 gigawatts already under construction, and 8 gigawatts of data center load expected to be energized by 2029 — with another 4.2 gigawatts to follow after. The company’s presentation characterized the trajectory as 50 percent load growth by the end of 2029, arriving two years faster than initially forecast.
That is company guidance, not built load, and the distinction is doing real work here. A forecast is a planning input for a utility; quoted on the courthouse steps, it becomes an argument. Kamin’s move was to treat CenterPoint’s investor-grade confidence as a governance problem: if the company is this certain about the load, why does the county have this little say over it?
The local count, cited by the county, put the existing footprint at 46 operating data centers and 55 in all stages combined — a figure reported by KPRC and picked up in the county’s materials. Treat it as the county’s tally rather than an audited census. And the scale on the utility’s own books has kept growing since April: with its second-quarter results on July 28, CenterPoint said it had filed more than 17 gigawatts of projects into ERCOT’s Batch Zero grouped study, and expects roughly 14 gigawatts to qualify as base or studied load. Against a Houston Electric peak of about 21 gigawatts, that is more than 65 percent of the system’s current peak — filed, not approved, but the number management is putting in front of investors.
In the same July release, CenterPoint raised its 2026–2035 capital plan by $1.2 billion to $66.7 billion, citing accelerating large-load demand — the second quarterly raise this year, each priced off the same queue the county wants the power to slow.
A regulated utility earns its return on what it builds
Here is the mechanism that turns a governance dispute into a business story. CenterPoint is the only investor-owned electric and gas utility based in Texas, an energy delivery company serving more than seven million metered customers across Texas, Indiana, Minnesota, and Ohio. It does not sell electricity to Houston households — retail providers do that in ERCOT’s deregulated market. It collects regulated delivery charges for operating and investing in the wires. Its earnings grow when its rate base grows: more approved capital spending, more return on that spending.
That structure makes $66.7 billion of planned investment over a decade the most important number in the story. The plan has been raised twice this year, each raise priced off the same queue the county wants the power to slow — and whether those projects proceed depends on the load actually arriving, which is why the Batch Zero tally and the 8-gigawatt energization target sit at the center of the investment case.
The Q3 print lands before the market opens on Tuesday, October 27, with the earnings call at 7:00 a.m. Central, per the company’s September 29 notice. Analysts have a “Moderate Buy” consensus and an average price target near $44.38, per MarketBeat. What October 27 will price is the plan as filed: the queue, the capex, the growth narrative. What happens after depends on forces the earnings release cannot control — the attorney general’s opinion, the December audit findings, and the January legislature.
The county’s case is about jurisdiction, not megawatts
Kamin’s argument is deliberately procedural. Texas counties have no zoning authority, and she has said state law leaves Harris County “little to no authority” over data centers. The legal backdrop is real: State Senator Paul Bettencourt’s February letter to Attorney General Ken Paxton — attached to a pending opinion request about county moratoria — asserted that counties were excluded from the 2025 moratorium statute, HB 2559, because they lack authority to impose development moratoria.
The practical gap is what makes the court vote matter more than it looks. On October 15, the court directed departments to notify commissioners of major data-center filings, asked the county attorney for a formal legal-authority report, and tasked intergovernmental affairs with pursuing legislative changes ahead of the January session. None of that is a moratorium. It is a notification and advocacy package — the maximum the court believes it can do under current law.
The outstanding legal question has its own history. In February, Hood County commissioners asked the attorney general whether counties have authority to temporarily halt data center development; half a year later, that request was still pending on the AG’s website. Harris County now wants the same question answered for its own purposes. Until it is, the county’s leverage is publicity and paperwork — flagging filings, building a legislative record, and quoting the utility’s numbers in public.
Note the asymmetry this creates. CenterPoint’s Batch Zero filings are studied at the ERCOT level under rules written by the PUCT. TCEQ permits are halted by the governor. Water is audited by the state water board. Every layer of real permitting authority over a Harris County data center sits above the county — at the state or the grid operator. The county holds the roads, the floodplain maps, and the fire marshal. Kamin’s October 15 move was, in effect, a request to be added to a permitting stack that was designed without her.
The $5 billion arithmetic — and what it assumes
CenterPoint’s answer to the politics is also arithmetic. With its second-quarter results, management estimated that the new large-load connections would reduce Houston Electric residential and commercial delivery charges cumulatively by at least $5 billion over the next decade. The logic is the standard spreading-of-fixed-costs argument: more load shares the same wires, so the per-unit delivery charge falls. The April figure was roughly $4 billion attached to 10 gigawatts of industrial load; the July update was at least $5 billion attached to the bigger tally.
Treat that as what it is: a forward-looking management forecast, not a measured saving. The company provided no annual breakdown, no per-customer figure, and no detailed calculation. It depends on how many proposed projects are actually built, when they begin operating, how much power they use, and how regulators allocate infrastructure costs. And it depends on the same megawatts the county wants the power to slow. If a governance dispute delays or diverts even a fraction of the 14 gigawatts CenterPoint expects to qualify for Batch Zero, the $5 billion shrinks with it.
The framing also deserves care. The savings relate to delivery charges — the regulated wires portion — not to the competitively priced electricity from retail providers, as the energy-trade press noted when the estimate first appeared. A Houston household reading about $5 billion in savings should know it is a projection about one slice of the bill, spread over ten years, priced off a load forecast that is itself under political attack.
This is the collision the whole story turns on. The utility’s investment thesis — $66.7 billion of capex earning a regulated return — and the county’s political complaint both start from the same 8 gigawatts. To CenterPoint, it is a growth pipeline to build for. To Harris County, it is a land-use future it cannot govern. The $5 billion savings forecast sits in the middle, a promise to ratepayers whose delivery depends on load the county is trying to slow.
The state is already moving underneath this
The desk’s earlier coverage is the connective tissue. Governor Greg Abbott’s August 3 directive halted approvals for new data center interconnections pending a full ERCOT audit of the roughly 474 gigawatts of tracked large-load requests; on September 21, he directed TCEQ to halt all data-center permits until the ERCOT and Texas Water Development Board audits finish. TCEQ must report compliance by October 19 — this past weekend — and ERCOT expects to publish its review by December 10. The governor has also pledged January-session legislation that could codify cost obligations, mandate water-efficient cooling, require annual usage reporting, and repeal data-center sales-tax exemptions.
Attorney General Paxton has opened an investigation into hundreds of data center developments across more than 30 Texas counties — Harris among them — demanding accurate water-use reporting. The county’s ask and the pending AG opinion both feed the January session, where the audit findings and the governor’s package will set the terms.
The practical upshot for anyone watching the queue: three clocks are running on the same load — the audit clock (ERCOT’s report by December 10), the legal clock (the AG’s opinion on county authority), and the legislative clock (January 2027). The October 27 earnings print prices the plan before any of them stop.
What it means for developers
For developers, Harris County just added a political risk layer to a permitting stack that already runs through ERCOT, the PUCT, TCEQ, and the water board. None of the county’s October 15 actions stop a project. But notification protocols are how opposition organizes: a filing flagged to a commissioner becomes a constituent meeting, a press hit, a legislative exhibit. In unincorporated Harris County — where most of the 55 projects in all stages sit — there is no city council to negotiate with and no zoning to comply with; there is now a commissioners court that has decided to watch, and to complain to Austin.
The broader lesson extends this desk’s recurring theme: a queue position is no longer an asset in itself. Texas developers already face SB 6’s $50,000-per-megawatt financial security, the statewide audit’s notarized attestations, and a governor who has halted permits. Harris County’s move adds county-level political scrutiny to the diligence file — site control and water sourcing on paper, and now a community-relations posture that anticipates elected officials reading your utility’s forecast aloud at a press conference.
There is also a timing effect worth pricing. Projects that can demonstrate maturity stand to gain from every thinning of the queue, as this desk has argued before. But maturity now has to be legible to a new audience: not just ERCOT’s engineers and the PUCT’s accountants, but a county attorney preparing a legal-authority report and an intergovernmental affairs team drafting statutory asks.
What to watch
Four milestones will decide whether October 15 was theater or a turning point. The first is October 27, CenterPoint’s third-quarter print: watch whether management reiterates the 8-gigawatt-by-2029 target and the 14-gigawatt Batch Zero expectation, or hedges either — and whether any analyst asks about county opposition on the call. The second is the attorney general’s pending opinion on county authority to halt development, requested in February by Hood County and now directly relevant to Harris County’s strategy. The third is December 10, when ERCOT’s audit report will show how much of the statewide queue survived verification — the document that sets the factual baseline for January. The fourth is the legislative session itself, where Harris County’s statutory asks meet the governor’s data-center package and a legislature historically sympathetic to development.
Underneath them is the question Kamin actually asked: in a state where the grid is planned at ERCOT, permits are gated by the governor, and counties have no zoning, who gets to say no — and where? Texas built the fastest-growing data center market in the country on a system designed to approve, not to govern. Harris County just asked to be written into the approval. The $66.7 billion plan assumes the answer comes fast enough not to matter. The county is betting it doesn’t have to.