CenterPoint Says Houston's Data Center Boom Will Save Ratepayers $5 Billion. It Just Raised Its Own Spending Plan by $4 Billion to Chase It.

Houston's electric utility has spent the last two years on the losing end of the math, at least from a ratepayer's chair. Storm-hardening costs climbed after Hurricane Beryl knocked out power across the region in 2024, and CenterPoint Energy has filed roughly ten separate rate requests since 2025 seeking more than $8 billion, most of them unrelated to data centers at all. The city of Houston has fought a number of those filings, denying one distribution-cost application outright and delaying another by 45 days pending further review. So it was a genuinely unusual thing for this utility to say out loud in August, when it told investors that the flood of new industrial customers crowding onto its grid, overwhelmingly data centers, could save its 2.9 million existing Houston-area customers more than $5 billion over the next decade.

The logic behind that number is ordinary utility economics, even if the scale isn't. A grid's fixed costs, the wires, the substations, the transmission lines, get divided among everyone who uses them. Add a large new customer that pays its own way rather than leaning on infrastructure built for it, and the average cost per kilowatt-hour for existing customers goes down instead of up. CenterPoint's August announcement, which it branded its Customer Savings Initiative, rests on exactly that mechanism: as much as 14 gigawatts of proposed large-load projects, almost all of them data centers, connecting to its Houston-area system and spreading the grid's fixed costs across a far bigger base of paying customers.

The bet behind the number

Two months later, on October 27th, the company put real money behind that bet. CenterPoint raised its 10-year capital plan by $4.0 billion, to $70.7 billion, and pulled $1.5 billion of that forward into the next five years, lifting the 2026-2030 portion of the plan from $34.2 billion to $38.7 billion. Chairman and CEO Jason Wells told investors the growth driving it was "unprecedented," and that even this larger number "represents only a portion of the opportunities in front of us" — the company says it has already identified roughly $10 billion more in potential projects it hasn't formally added to the plan. CenterPoint also raised its 2026 earnings guidance to $1.90 to $1.92 a share and initiated 2027 guidance at $2.07 to $2.09, a projected 9% increase.

What's actually filling that pipeline is specific enough to check. CenterPoint now expects to energize 8 gigawatts of data center load in Greater Houston by 2029, with 3.5 gigawatts already under construction, and executives have told investors they expect roughly 14 gigawatts of new large-load demand online by 2030. The company has submitted about 17 gigawatts of large-load projects into ERCOT's "Batch Zero" process, the first run of Texas's new framework for studying big loads of 75 megawatts and above together instead of one application at a time. The Public Utility Commission of Texas approved that framework on June 18th, after Governor Abbott ordered a statewide audit of how data centers were jumping the interconnection line in the first place. Statewide, ERCOT has reported more than 438 gigawatts of large-load interconnection requests in its queue, with nearly 90% of that total coming from data centers — which is the scale of the problem Batch Zero was built to sort through. ERCOT isn't expected to produce a final transmission plan for the batch until sometime in 2027.

What the $5 billion pledge leaves out

CenterPoint hasn't published the math behind its own number. There's no year-by-year table, no per-customer estimate, and no breakdown of how much of the $5 billion depends on which gigawatt of demand actually arrives. That's not necessarily spin; utilities rarely publish granular cost models this early in a buildout. But it does mean the pledge is a projection built on a pipeline of submitted applications, not a result built on signed, paying load, and those are two very different things to put in front of a ratepayer.

The gap between a submitted interconnection request and an actual, paying customer is exactly where large-load tariffs and minimum-bill clauses earn their keep, and it's the part of this story most coverage skips over. If a data center signs a contract for 300 megawatts with a minimum-bill provision requiring it to pay for most of that capacity whether it uses it or not, existing ratepayers really are protected if the project shrinks or slips years into construction. If the tariff language ends up softer than that, or a customer walks away before the capital it triggered gets paid down, the savings math runs the other way and lands back on the same 2.9 million customers CenterPoint is promising to help. ERCOT's 17 gigawatts of submitted Houston-area projects is a queue, not a commitment. The minimum-bill terms the Public Utility Commission eventually locks in for Batch Zero loads are what decide whether that queue becomes a real $5 billion win or another cost-allocation fight a few years from now.

I think the right way to read this quarter's numbers is to treat the $70.7 billion and the $5 billion as two separate claims, not one. The capital increase is real and already committed, backed by a specific equity financing plan and a specific five-year acceleration. The savings figure is a forecast that depends on a queue of proposed projects converting into signed, protected contracts at close to the rate CenterPoint is assuming today. Those two things don't have to succeed or fail together. A utility can spend its way into a bigger, more expensive grid and still deliver real savings on the other side, but only if the large customers driving the spending actually pay their own freight once the meter is running, not just when the press release goes out.

It's worth remembering the backdrop here, because it cuts against treating any of this as a company turning over a new leaf. CenterPoint's roughly $8 billion in rate requests since 2025 had nothing to do with data centers, and Houston's own regulatory affairs office has kept pushing back on them anyway. None of that is a reason to doubt the economics behind spreading fixed costs across a bigger customer base; the mechanism is sound and well understood. It's a reason a Houston ratepayer has earned the right to be skeptical when the same company that keeps filing for more money turns around and promises the opposite, even when, this time, the arithmetic actually points the other way.

What actually proves the pledge

The real test isn't anything said on an earnings call. It's whether the large-load tariff language the Public Utility Commission eventually locks in for Batch Zero projects makes the minimum bill real and enforceable, whether the 14 gigawatts CenterPoint is counting on shows up as signed service agreements rather than withdrawn applications, and whether the next rate case Houston sees actually reflects a lower cost per customer instead of a higher one. CenterPoint has given itself roughly three to four years to make that case. Houston ratepayers, who have heard plenty of big numbers from this utility before, are right to wait for the one that actually shows up on a bill.