12 min read

Congress Says Data Centers Should Pay Their Own Way. Louisiana Has a $12.9 Billion Reason to Read the Fine Print.

The Senate passed a bill promising data centers will pay their own transmission costs. The fine print, and a $12.9 billion Entergy proposal in Louisiana, tell a more complicated story.
Congress Says Data Centers Should Pay Their Own Way. Louisiana Has a $12.9 Billion Reason to Read the Fine Print.
Transmission towers over a rural landscape. Generic image, not a Louisiana site. Photo by Petr Ganaj via Pexels, free license.

337.NEWS | Investigative Analysis | November 21, 2026

There's an old rule in business. If you want something built, you pay for it.

You want a bigger warehouse? You buy the steel. You need a private road? You hire a contractor. You want to put a swimming pool behind your house? You don't send the invoice to everyone else on the street.

Unless, apparently, you're building a data center.

Then things get complicated.

The electricity has to come from somewhere. The transmission lines have to reach somewhere. Somebody has to build the substations, finance the generators, and maintain the equipment long after the politicians have finished cutting ribbons.

And somebody has to pay.

That last part has become a rather uncomfortable question in Louisiana, where billions of dollars in proposed electrical infrastructure are being justified, at least in part, by the extraordinary appetite of the artificial intelligence industry.

On Thursday, November 19, the United States Senate passed the Bipartisan American Affordability and Jobs Act of 2026 by a vote of 62–33.

The legislation, commonly called BAAJA, promises to address that problem by making data centers responsible for certain costs they impose on the electrical grid.

It also happens to be one of the most consequential proposed overhauls of federal infrastructure permitting in years.

And tucked inside the same legislative machinery are provisions that could make it easier to build major industrial projects, restrict certain avenues for challenging federal decisions, and give developers greater confidence that approved projects will stay approved.

The bill has not become law. It still needs approval from the House and the president's signature.

For Louisiana residents already watching data centers, industrial development, and utility costs collide, this legislation deserves a closer inspection.

Not because protecting ratepayers is a bad idea.

Because when somebody in Washington promises to make corporations pay their own bills, it's worth checking whether the invoice actually reaches the corporation.

The $12.9 billion elephant in Louisiana's room

Let's leave Washington for a moment.

In March 2026, Entergy Louisiana filed an application with the Louisiana Public Service Commission seeking authorization for a substantial new package of electricity generation resources.

According to Entergy's SEC filings, the proposal included seven combined-cycle natural gas generating units totaling approximately 5,278 megawatts, at an estimated cost of $12.9 billion.

The application also involved battery storage and additional infrastructure.

The development is associated with the enormous electrical demand anticipated from a Meta-affiliated data center project in Richland Parish, north Louisiana.

$12.9 billion. That's the estimated generation investment, subject to regulatory review and approval.

Now, $12.9 billion doesn't automatically mean $12.9 billion being dumped onto residential electricity bills. That's an important distinction.

Entergy says its agreements are designed to protect existing customers. The company also projects approximately $2 billion in additional customer savings over 20 years from its expanded Meta arrangement, on top of $650 million previously announced.

Those figures are Entergy's projections. They are not evidence that customers have already received those savings.

But this is where the conversation gets interesting.

An electric utility doesn't ordinarily finance infrastructure the way you or I finance a backyard shed.

Regulated utilities recover approved costs through rates and other financial arrangements. Exactly who bears those costs depends on contracts, regulatory decisions, cost-allocation rules, and whether the projected customers actually purchase the electricity they promised to use.

A contract might provide substantial protection against those risks.

It might also leave important questions unanswered.

And when the numbers run into the billions, even a comparatively small allocation error can become somebody else's very expensive problem.

So the question isn't whether Entergy has promised protection.

The question is what the contracts guarantee, what risks remain, and who pays if the forecasts turn out to be wrong.

Those are questions worth asking before the concrete is poured.

Not twenty years afterward.

The fine print starts with Section 2107

For all the political theater surrounding this legislation, Section 2107 contains something genuinely important.

It attempts to establish a clearer national framework for allocating the electrical transmission costs associated with data centers.

Under the Senate-passed proposal, federal regulators would revise transmission pricing policies so covered computing facilities pay both a share of existing transmission infrastructure and the additional transmission costs attributable to their demand.

It also establishes protections intended to prevent new data centers from shifting their incremental electricity supply and grid costs to other customers.

In ordinary English, the idea is straightforward.

If a developer's facility creates the need for another transmission line, the developer shouldn't be allowed to distribute that expense among thousands of households that never requested the additional electricity.

That's a reasonable proposition.

But the bill draws an important distinction between two kinds of costs.

Embedded costs are the costs associated with infrastructure already in service. Incremental costs are the additional expenses required to accommodate new demand.

That distinction matters because developers can impose costs in both categories.

Imagine a warehouse tenant moving into a commercial building. The tenant pays to install specialized equipment, but also uses the loading docks, elevators, parking lots, and other facilities that existing tenants helped finance.

Paying for the new equipment doesn't necessarily mean the tenant is paying an appropriate share of the existing building.

The electrical grid works on a considerably larger and more complicated scale, but the principle is similar.

Section 2107 attempts to address both types of transmission expense.

It is not, however, a universal federal takeover of electricity rates.

And that's where Louisiana comes back into the picture.

Washington can write the rules. Louisiana still controls important parts of the bill.

Federal Energy Regulatory Commission oversight primarily concerns interstate electricity transmission and wholesale electricity markets.

Retail electric rates, local distribution systems, and many utility resource-planning decisions remain matters of state jurisdiction.

Section 2107 itself preserves that division. It affirms state authority to implement certain policies on electricity rates for new data centers, and the bill does not grant FERC authority over retail electricity sales or local distribution.

That means the Louisiana Public Service Commission doesn't suddenly become irrelevant.

Far from it.

The LPSC still has a significant role in determining whether proposed utility investments are justified, how approved costs are recovered, and what protections customers receive under Louisiana-regulated service arrangements.

For Louisiana, that's important because the state has already been developing its own framework for large electricity customers.

Entergy's filings refer to the LPSC's Lightning Initiative and the utility's Fair Share Plus guidelines as protections associated with large data center projects.

These arrangements are supposed to prevent ordinary customers from shouldering costs properly attributable to massive new industrial loads.

So now we have two layers of protection being discussed.

One in Baton Rouge.

One in Washington.

The unresolved question is how those protections interact, especially where agreements were negotiated before the proposed federal law takes effect.

And whether the additional federal rules produce stronger protections or merely reinforce what regulators already intended to accomplish.

That isn't a technicality.

That's the difference between a new safeguard and a congratulatory press release describing an old one.

Then somebody changed the definition.

Here's where the story takes a turn.

On November 17, the sponsors proposed a manager's amendment that would have narrowed the definition of facilities subject to certain transmission-cost requirements.

Earthjustice, an environmental law organization, reviewed the proposed amendment language and identified a potentially significant problem.

Under the proposed revised definition of a covered computational load site, certain facilities primarily used for internet access, cable, information services, telecommunications, VoIP, or other wire communications services could be excluded.

Earthjustice warned that exclusion could have covered cloud services such as Microsoft Azure, Google Cloud, and Oracle Cloud, AI services including ChatGPT, Claude, and CoPilot, and Starlink (SpaceX) facilities.

Those are Earthjustice's interpretations of the proposed language, not established legal exemptions for those companies.

That amendment was not adopted. The Senate passed the bill on November 19 without it, and the sponsors say they still want to add it in the House.

But consider the implications.

Congress sets out to prevent data centers from avoiding their full transmission costs.

Then Congress considered language that could have allowed some facilities performing exactly the kinds of computing activities driving electrical demand to argue that they fall outside the relevant definition.

There's an obvious question here.

If an AI company could have avoided a cost requirement because its computing facility qualifies as an information service, what exactly was the requirement designed to accomplish?

To be fair, this is not a settled conclusion.

The precise application would depend on the enacted language, the facility's operations, and the regulators responsible for implementing the law.

Nor does an exemption from one transmission-pricing provision necessarily exempt a facility from every other ratepayer protection.

But the concern is substantial enough to deserve an answer before the legislation reaches the president's desk.

Especially in a state actively recruiting the very companies that might benefit from that distinction.

The other half of this bill has nothing to do with your electric bill

Here's something worth remembering.

BAAJA isn't exclusively a data center bill.

It also rewrites parts of the federal permitting process for energy, mining, transmission, and other infrastructure projects.

Supporters argue that the existing system has become too slow, expensive, and unpredictable.

They have a legitimate argument.

A delayed transmission project can delay the construction of renewable energy facilities. A drawn-out federal review can hold up necessary infrastructure. And uncertainty can discourage private investment even when projects are environmentally responsible.

The permitting process is not automatically better simply because it takes longer.

But neither is a permitting decision automatically better because it happens faster.

BAAJA would change the scope and timing of certain environmental reviews, modify judicial-review procedures, and provide additional protections for projects that have obtained necessary federal approvals.

Congressional researchers have documented how its project-certainty provisions could constrain subsequent agency action. The bill generally bars an agency from revoking, suspending, or materially changing a federal permit for a fully permitted project. The exceptions are narrow: a court order, a material breach or other violation of law, fraud, a request by the permit holder, or action needed to prevent a specific, urgent, substantial, and proximate threat to life, property, national security, or defense, based on new information. General environmental harm, on its own, is not listed as an exception.

This is where the legislation gets uncomfortable for people who have spent the past several years watching industrial development unfold across Louisiana.

Because there's a very practical difference between removing unnecessary paperwork and removing opportunities to discover a problem.

One makes government more efficient.

The other can make government less accountable.

And it's entirely possible for a piece of legislation to do some of both.

A permit is not supposed to be a lifetime warranty against scrutiny.

Imagine buying a piece of property.

You discover a major industrial facility is being planned nearby. Maybe its footprint has expanded. Maybe its water needs are larger than previously disclosed. Maybe updated studies reveal environmental impacts that weren't adequately considered when the original permit was issued.

You would reasonably expect the relevant agency to have some authority to revisit those concerns.

BAAJA's project-certainty provisions are intended to make it more difficult for the federal government to reverse course after a developer secures the necessary approvals.

A proposed November 17 amendment would have clarified that the protections apply only to projects possessing all necessary and final permits, rather than allowing a developer to claim the same protection merely because it obtained one preliminary authorization. That amendment was not adopted.

Earthjustice called that clarification a modest but positive change, while maintaining that the revised provisions remain too restrictive and could insulate certain permitting decisions from meaningful subsequent review.

Both positions deserve scrutiny.

Developers need predictable rules.

Residents need enforceable protections.

Neither interest should automatically cancel out the other.

What concerns me is the possibility that a government could approve an inadequately examined project, then point to a federal permitting-certainty provision when residents demand another look.

That's a potential consequence worth examining, not a claim that every approved project would become untouchable.

And it's particularly relevant in Louisiana, where industrial development routinely involves overlapping state, federal, and local approvals.

The information problem Congress quietly acknowledged

There's another section of this bill that may be more revealing than the political speeches.

Section 2114 would require more systematic federal reporting of electricity consumption by data centers.

Currently, the Energy Information Administration's principal electric utility sales survey does not separately identify data center consumption in the way the proposed legislation contemplates.

The bill would change that by requiring reporting that distinguishes electricity used by data centers, including facilities supporting artificial intelligence, cloud computing, and cryptocurrency mining.

Think about that.

We're planning enormous expansions of electrical infrastructure to accommodate an industry whose electricity consumption the federal government is still trying to measure more precisely.

That doesn't mean nobody has estimates.

Utilities certainly have demand forecasts. Developers provide projections. Regional grid operators conduct system planning.

But projections aren't the same as verified consumption.

And somebody eventually has to reconcile the difference.

What happens if the anticipated demand never materializes?

What happens if computing efficiency improves dramatically?

What happens if a facility closes, scales back, or moves before the infrastructure supporting it has been paid off?

Those aren't hypothetical concerns invented by opponents of technological progress. They are among the cost-allocation risks the legislation is intended to address.

Power plants are long-lived assets.

Artificial intelligence business models can change considerably faster.

And a natural gas generating station doesn't disappear from a utility's books simply because its intended customer decides the economics of its data center no longer work.

The part Louisiana communities could actually use

Buried deeper in the legislation is Section 2117, which would establish a Data Center Community Assistance Program through the Department of Energy.

The proposal would provide state, local, and tribal governments with technical resources for evaluating data center proposals.

That includes information about electricity demand, infrastructure costs, water requirements, backup power generation, employment projections, and financial impacts.

It would also provide an opportunity for independent analysis using information developers voluntarily supply.

The program would not independently approve or reject projects, and the provision does not itself authorize a specific funding amount.

This is an interesting admission.

Congress appears to recognize that local governments are being asked to evaluate enormously complex industrial projects without always having the technical resources necessary to challenge the assumptions presented to them.

A parish government may be perfectly capable of managing roads, drainage, zoning, and ordinary development.

That doesn't automatically qualify its members to evaluate gigawatt-scale energy forecasts or the financial structure of a multibillion-dollar utility agreement.

And there's nothing embarrassing about admitting that.

The embarrassing part would be approving a project without understanding its consequences, then insisting everything was carefully evaluated.

Independent technical assistance could be valuable.

Provided communities actually use it.

And provided the information supplied is sufficient to conduct a meaningful review.

So what does this mean for Southwest Louisiana?

Not every industrial development is a Meta-scale hyperscale facility.

Not every proposed data center requires a dedicated power plant.

And not every electrical infrastructure improvement associated with a new customer is automatically an inappropriate subsidy.

But Southwest Louisiana has a particular reason to pay attention.

Our region is already home to energy-intensive industrial facilities. We have communities accustomed to refinery expansions, LNG developments, transmission infrastructure, pipeline corridors, and increasingly ambitious computing and power-generation proposals.

The industrial footprint is substantial.

So are the consequences when infrastructure planning goes wrong.

A Louisiana resident concerned about a nearby project doesn't necessarily oppose economic development.

They might simply want to know who pays for the road.

Who pays for the power.

Who supplies the water.

Who is responsible when something breaks.

And whether the people making the promises will still be around when the bill comes due.

Those questions don't make somebody anti-business.

They make somebody reasonably interested in the terms of the transaction.

For the proposed Louisiana developments already moving through state regulatory processes, several questions now deserve direct answers from Entergy, the LPSC, and the relevant developers:

  • Which infrastructure costs are contractually assigned to the data center, and which remain eligible for recovery from other customers?
  • What financial guarantees protect ratepayers if a major computing facility reduces demand, delays construction, or closes?
  • How would the federal requirements affect existing agreements, and would any provisions apply only to newly connected facilities?
  • Could the proposed computational-load definition, if added in the House, affect projects currently under consideration in Louisiana?
  • How will residents obtain independent verification of promised customer savings?

As of publication, the reviewed records do not establish that Meta or another named Louisiana developer has used the proposed exemption, or that the $12.9 billion generation proposal would be transferred to ordinary residential customers.

Those distinctions are important.

Investigative reporting ought to identify what the evidence demonstrates, not simply what the evidence makes possible.

The bill has two promises. We should examine both.

There is nothing inherently wrong with wanting America to build things faster.

There is nothing inherently wrong with wanting technology companies to pay their own infrastructure costs.

In fact, both objectives could produce genuine benefits for Louisiana.

More reliable electricity.

Additional industrial investment.

Better infrastructure.

Potentially lower costs for existing customers.

But good legislation needs to be evaluated by what its provisions actually accomplish, not how attractive its title sounds.

The Senate has presented a bill that promises to protect ratepayers while making industrial projects easier to develop.

That is the bargain being offered.

And bargains ought to be examined before somebody signs the contract.

Because Louisiana has seen enough economic development announcements to understand that the announcement is usually the easy part.

The real work comes afterward.

When the contractors arrive.

When the transmission lines go up.

When the land changes hands.

When the machinery starts running.

And when the invoice finally shows up.

If these companies can afford billions of dollars to build the future, they should be able to afford the infrastructure required to operate it.

If Congress intends to guarantee that principle, it should write the law clearly enough that corporations cannot easily argue their way around it.

And if Congress intends to make industrial permitting faster, it should be prepared to explain exactly which public safeguards are being changed in the process.

That's not opposition to progress.

That's asking to see the contract before agreeing to pay the bill.


Documents and sources

Editor's note: This article distinguishes enacted law from pending legislation, company financial projections from realized customer savings, and proposed legal interpretations from established regulatory outcomes. The Senate-passed legislation remains subject to House consideration and possible amendment.

337.NEWS discussion

Join the conversation.

Sign up to comment on this story and follow the local reporting behind it.