Meta's latest Louisiana pitch is easy to summarize: more than $50 billion, as much as 5 gigawatts of computing capacity, and enough local spending to turn a rural parish into an artificial-intelligence hub. The harder part is separating what Meta announced on Monday from what Louisiana regulators have actually approved.
That gap is the story. Meta's July 13 announcement says the Richland Parish expansion will be one of the world's largest AI-infrastructure investments. But the seven new gas plants, three battery sites, nuclear uprates and transmission buildout associated with the next phase remain under regulatory review. Meta's ambition is moving faster than the public approval record.
This does not mean the project is imaginary. Construction is active, an earlier power package has been approved, and local businesses and schools are already reporting benefits. It means the headline numbers describe a possible end state, not the data center's current size or an approved 5-GW power supply.
- The claimMeta says its Louisiana investment will exceed $50 billion and the campus could reach 5 GW of compute, but those are expansion ambitions rather than today's operating load.
- The approval gapThree plants for the initial phase have commission approval; seven additional gas units totaling 5,278 MW remain under regulatory review.
- The clockLouisiana regulators have targeted December 16, 2026, to consider the expansion power package.
- The customer caseEntergy projects $2.65 billion in combined customer savings, but that figure depends on approvals, construction costs and Meta's load arriving on schedule.
What Meta announced — and what it did not
Meta published the new expansion announcement at 5:30 a.m. EDT on July 13. It said its investment in the Richland Parish region will exceed $50 billion and that the data center has the potential to scale to 5 GW. The company also said it will pay the full cost of the energy, water and related infrastructure used by the facility.
Those are material claims, but the wording matters. Meta describes more than $50 billion as a regional investment and 5 GW as the expansion's computing capacity. It does not provide a line-by-line capital budget, a construction schedule for every phase or a regulator-approved 5-GW load figure in the announcement.
The official state project page still describes the original Richland Parish commitment as a $10 billion project, with more than 500 direct jobs and more than 1,000 indirect jobs. That older figure is not evidence that Meta's new number is wrong; it shows how dramatically the stated scope has changed and why the public needs a clearer bridge between the initial project, later financing arrangements and Monday's expansion claim.
TECHi previously examined Meta's $27 billion Blue Owl financing structure and the burden it could shift onto rural infrastructure. The new catalyst is different. It is no longer mainly about who owns and finances the campus. It is about whether the surrounding utility system can build enough power, storage and transmission to support Meta's proposed scale without leaving other customers exposed.
The seven-plant plan is still pending
Entergy Louisiana disclosed the proposed expansion power package in March. Its official agreement announcement calls for seven new combined-cycle natural-gas plants totaling more than 5,200 MW, battery storage across three locations, nuclear uprates, roughly 240 miles of 500-kV transmission and Meta support for as much as 2,500 MW of new renewable resources.
That is a proposed portfolio, not a completed one. Entergy's June investor presentation lists four Richland units and three Pointe Coupee units totaling 5,278 MW, with expected service dates in 2030 and 2031. The same presentation labels the expansion as pending regulatory review.
The Louisiana Public Service Commission has set an expedited schedule for docket U-37882, but its April 15 commission transcript is not an approval of the plants. It targets a commission decision by December 16, 2026. Until the commission acts, the seven-plant package should be described as proposed.
The public record also leaves some details unresolved. Entergy says battery storage will span three locations, while its public investor material identifies two 200-MW sites and leaves another site to be determined. The exact hyperscale load is redacted in public filings. Those omissions do not invalidate the plan, but they limit how confidently anyone outside the proceeding can model the final system.
The power mix is broader than seven gas plants
The gas-buildout headline is important because seven large combined-cycle units would lock in a substantial new generation fleet. Yet describing the plan as gas-only would also be incomplete.
Entergy's proposal combines gas generation with grid-scale batteries, nuclear uprates, purchased power and potential renewable additions. The utility says the gas units would be capable of future carbon capture and hydrogen co-firing, although capability is not the same as a funded deployment. It also says Meta could help fund up to 2,500 MW of new solar resources.
That proposed buildout also connects to a broader market question TECHi has examined: whether artificial-intelligence data centers are becoming a power-stock story.
The practical question is sequencing. Entergy's investor deck places the seven gas units in 2030-31, while Meta's requested ramp begins earlier. Storage, transmission, existing generation, the initial approved plants and purchased power will therefore have to carry different parts of the load as the campus expands. That makes the schedule as consequential as the eventual nameplate total.
It also explains why a single 5-GW number can mislead. Meta uses 5 GW to describe potential compute scale. Entergy describes more than 5.2 GW of proposed gas generation. Those figures are related, but they are not interchangeable measurements of current power use. Cooling systems, electrical losses, redundancy, utilization and the timing of each construction phase all stand between computing capacity and what the utility must deliver at a given hour.
What Louisiana has already approved
The pending expansion should not be confused with an earlier package that already cleared the commission. In December, Entergy said the LPSC had approved three generation facilities and associated transmission investments for the initial phase of Meta's project.
That approved package includes two Richland Parish plants expected in late 2028 and a third facility near Waterford expected by the end of 2029. Entergy said the initial agreement runs for 15 years, requires Meta to fund the full cost of the interconnection infrastructure and is expected to produce at least $650 million in savings for other customers.
The distinction is simple but essential: three initial plants have approval; seven additional plants do not. Treating all ten as one approved build would overstate the regulatory status. Treating the project as merely conceptual would ignore construction and the approved infrastructure already under way.
For readers trying to understand whether artificial-intelligence data centers can raise household bills, this split is why broad national arguments need local contract details. TECHi's state-by-state guide to data-center electricity costs shows that the risk depends on who pays for interconnections, how long the hyperscaler remains committed and whether utility forecasts survive commission scrutiny.
Meta's customer-savings claim has a real structure behind it
Meta's strongest counterargument is that Louisiana customers are not being asked to subsidize the expansion. Entergy says the new agreement is structured so Meta pays its full cost of service and is expected to deliver roughly $2 billion in customer savings over 20 years, in addition to the $650 million associated with the initial agreement.
The utility attributes those savings to a large, long-duration customer spreading fixed system costs across more electricity sales. It also says Meta will fund project-specific infrastructure and provide credit protections. Those provisions are more substantive than a vague promise to be a good corporate citizen.
Still, an expected savings model is not the same as money already returned to customers. It depends on regulatory approval, construction costs, load arriving on schedule, financing assumptions and Meta remaining on the system under the agreed terms. The commission's job is to test those assumptions in public, not simply accept the headline benefit.
This is where the next nine months matter. The December target gives regulators a compressed window to examine an unusually large resource package. Cost allocation, exit protections, construction overruns and the value of assets if Meta's load grows more slowly than forecast deserve as much attention as the projected savings total.
Local gains are visible, even if the latest evidence comes from Meta
Meta's announcement is built around people rather than megawatts. It says more than $1.6 billion in contracts have gone to Louisiana businesses, that local infrastructure investment has exceeded $1 billion and that the company is donating $5 million to Louisiana Delta Community College for scholarships tied to data-center trades.
The company also quotes Richland Parish educators who say teacher bonuses rose sharply and local business owners who describe rapid customer growth. These accounts should not be dismissed. A project of this size can change wages, school funding and demand for local services before the first server hall reaches its ultimate capacity.
But most of Monday's specific benefit claims are supplied by Meta itself. The independent official baseline is more modest: Louisiana Economic Development has previously cited more than 5,000 construction jobs and about 500 operational jobs for the initial project. Readers should treat Meta's new contract, scholarship and business-growth figures as company-reported until agencies publish comparable audited totals.
There is also a time-horizon mismatch. Construction creates an intense, temporary surge in labor and local spending. Permanent data-center operations employ far fewer people than the construction phase. A fair evaluation should track what remains after the buildout: lasting tax revenue, reliable infrastructure, skilled jobs, utility costs and environmental obligations.
Why Meta's compute race is also a Louisiana utility story
A 5-GW campus would not be a normal corporate facility. It would be infrastructure on the scale of a regional power system, built to support the training and operation of increasingly large artificial-intelligence models.
That changes the competitive question for Meta. The company is not only competing for chips and researchers. It is competing for interconnection queues, turbine capacity, transmission corridors, construction labor and regulatory permission. The limiting factor for frontier-model development can move from silicon to steel, gas supply and public process.
It also changes the risk. Building too little power could delay Meta's model-training roadmap. Building too much too early could leave ratepayers and Entergy shareholders arguing over assets designed around a load that arrived late. The proposed contract is meant to contain that risk, but the commission must still verify that the protections work under less optimistic scenarios.
Disclosure: This article is infrastructure and market analysis, not investment advice. The proposed generation plan, regulatory schedule, customer-savings estimates and construction timing can change. Check current commission records and company disclosures before making financial decisions.
The TECHi verdict
Meta's Louisiana expansion is real enough to reshape Richland Parish and large enough to matter to the national artificial-intelligence race. The company has also made a clearer ratepayer-protection case than many hyperscale projects by agreeing to fund dedicated infrastructure and support a long-term utility contract.
The mistake would be treating Monday's announcement as the end of the process. More than $50 billion is Meta's stated regional investment, not an independently itemized public budget. Five gigawatts is a potential end state, not the current load. Seven new gas plants are in a regulatory proceeding, not under a final approval.
The most useful milestones now are not another ribbon cutting or a larger headline number. They are the LPSC's December decision, final cost-allocation terms, disclosed battery and transmission sites, and a schedule that shows how approved and proposed resources match Meta's ramp. Until then, the Hyperion story is an enormous compute ambition attached to a power plan that Louisiana still has to judge.
FAQ
Frequently asked questions
How large will Meta’s Louisiana data center be?
Meta says the Richland Parish expansion has the potential to reach 5 GW of computing capacity and represents more than $50 billion in regional investment. Those figures describe a potential end state, not the site’s current load or a fully approved power package.
Have the seven new Louisiana power plants been approved?
No. Entergy Louisiana’s seven-plant expansion is under review in Louisiana Public Service Commission docket U-37882. The commission’s procedural schedule targets a decision by December 16, 2026.
What power infrastructure is already approved for Meta’s project?
An earlier package covering three generation facilities and associated transmission investments has commission approval. Two Richland Parish plants are expected in late 2028, with a third facility near Waterford expected by the end of 2029.
Will Entergy customers pay for Meta’s data center?
Entergy says its agreements require Meta to pay the project’s full cost of service and fund dedicated infrastructure. The utility projects customer savings, but those estimates depend on regulatory approval, construction costs and Meta’s load arriving as forecast.
About the Author
Qaiser Sultan writes TECHi's Two Takes column, a dual-perspective format that argues both sides of a market debate and then picks one. He focuses on contested calls: whether a valuation is defensible, whether management guidance is credible, whether a trade setup has enough asymmetry to matter. The format demands honest engagement with the strongest counter-argument — which is why it runs here and not as another one-sided hot take.
