Why Louisiana Power Is Harder to Get Than the Headlines Suggest
By Erick Watson (吴子文), SproutVest
Louisiana just crossed $250 billion in announced investment. A $100 billion spaceport in Vermilion Parish, a $50 billion AI campus in Richland Parish, an $18 billion data center program in the northwest corner, and a string of LNG and steel projects behind them. The natural conclusion, the one every press release invites, is that Louisiana is where the power is. If the hyperscalers found gigawatts there, surely a 15 megawatt project can find its share.
That conclusion is wrong, and the error will cost the people who act on it years and millions. The announced gigawatts are real. They are also spoken for, financed by mechanisms that do not scale down, and delivered through a grid whose structure the headlines never mention. Anyone buying, building, or underwriting compute in the 1 to 50 MW range needs to read the power math underneath the announcements, because that math points in a different direction than the press releases do.
The Announced Gigawatts Belong to the Announcers
The largest project in the state will be served by roughly ten new gas-fired power plants. The company behind it is paying for them: the generation, the transmission, the interconnection, all of it, under supply agreements measured in decades. That is the template every megaproject in Louisiana now follows. Bring your own balance sheet, fund your own generation, commit for fifteen years, and the state will move at remarkable speed to accommodate you.
Notice what that template requires. It requires the ability to finance a power plant. It requires the creditworthiness to sign a supply commitment longer than most funds’ entire lifespan. The state’s new fast-track approval pathway for large loads was written around exactly these commitments, which means the fast lane is structurally reserved for companies that can make them. A 15 MW tenant cannot finance a 750 MW plant, and no one will build one for it.
The gigawatts in the headlines are therefore not a pool that smaller buyers draw from. They are bilateral arrangements between a utility and a single customer, and the customer already has a name. The second wave arrives to find that the announced abundance was never inventory.
The Grid Under the Press Releases
Here is the structural fact that determines the second wave’s actual options: across most of Louisiana’s territory outside the investor-owned utilities, electricity is delivered by distribution cooperatives that own no generation at all. The Acadiana co-op serving the Lafayette region buys every kilowatt wholesale. The largest co-op in the state, serving the parishes around Baton Rouge, buys its full requirements from an investor-owned utility under contract.
This matters enormously for a mid-scale project. When you ask a distribution co-op for 20 MW, the co-op cannot simply say yes. Your request travels upstream into a wholesale supply contract, and from there into a regional transmission queue, where it waits behind everyone else’s requests on someone else’s timeline. The co-op’s enthusiasm, which is often genuine, is not the same thing as the co-op’s authority.
The exceptions are worth knowing precisely because they are exceptions. Lafayette’s municipal utility owns its own generation, its own transmission, and its own fiber network, all under one civic roof, which makes it one of the few load-serving entities in the state that can actually decide to serve a mid-scale load and then do it. Territory structure, not announcement volume, is what determines your energization date. Most buyers never check the structure until the timeline has already slipped.
Scrutiny Is Now Part of the Interconnection Timeline
There is a second development the headlines undersell. Louisiana’s ratepayers have started asking who pays for all of this, and the question is getting sharper. A consultant to the state’s utility regulator estimated that one proposed plant purchase, driven principally by data center load, would add $8 to $13 per month to the average residential bill. The governor has publicly warned against anyone taking advantage of power markets at ratepayers’ expense, and his administration now conditions the state’s data center tax exemption on companies fully funding their own generation and transmission and delivering community benefits.
None of this stops the megaprojects, which can absorb the conditions. It does change the environment for everyone else. Every large-load interconnection request in the state now lands on desks that have learned caution, and a mid-scale project without a hyperscaler guarantee behind it inherits that caution without the political capital to offset it. Cost allocation, who pays for the substation, who backstops the contract if the tenant fails, has become the central question regulators and utility boards ask. If your project cannot answer it crisply, your project waits.
What a Second-Wave Buyer Should Actually Do
The pattern to avoid is the one playing out right now in every boom market: a buyer falls in love with a site, signs a letter of intent on land, and only then discovers that the power that made the site attractive was a press release, a substation with no spare capacity, or a queue position measured in years. The land was never the asset. The energization date was the asset, and nobody diligenced it.
The discipline is unglamorous. Verify the power math independently before committing to anything: the utility’s actual posture, the wholesale supply headroom behind it, the realistic interconnection timeline, and the cost allocation you will be asked to carry. Choose territories by structure, favoring load-serving entities that control their own generation or have demonstrated large-load appetite, rather than by proximity to a famous neighbor. And test the whole deal against the alternative of simply renting the compute, because a site that energizes in 2030 is a fundamentally different asset than one that energizes in 2027, and rented capacity is available today. We published our full framework for that comparison in Build, Buy, or Rent: An Operator’s Guide to AI Compute Decisions.
Louisiana’s boom is real, and the state has earned it. But the second wave will be won by buyers who treat the power math as the deal rather than a detail, and who get an independent read before the letter of intent rather than after. That independent read is exactly what our compute infrastructure advisory practice exists to provide. If you have a site, a deal, or a negotiation in front of you, bring the timeline and the power question, and we will tell you whether the announcement you are standing on can hold your weight.
Erick Watson is Managing Director of SproutVest. He led the buildout of a one megawatt, Tier 3 datacenter in Washington State, led product for an AI valuation platform that ran inference against all 105 million US households, and through Chainlodge, an AI datacenter infrastructure venture, sees today’s compute demand from the supply side. Reach him at hello@sproutvest.com or cal.com/sproutvest.
Where is your leadership effective, and where is it costing the company?
Most of the problems this blog covers trace back to how the founder runs the company. The Trellis Leadership Diagnostic maps that in 24 behavior-anchored items across six dimensions: about 12 minutes, instant results, free to take self-serve.
Take the Leadership Diagnostic →Exploring a fractional or advisory engagement instead? Book a discovery call →