Giga Energy’s $1B, 60MW Texas Lease: $1.67M per MW a Year

Giga Energy, a modular AI data center developer, said on 17 November that it signed a $1 billion lease agreement with an undisclosed US-based AI cloud provider for Project Marl, a 60MW data center in Hunt County, Texas. The lease runs 10 years, which works out to about $1.67 million per MW a year, by our arithmetic.

Giga says it will finish the campus in under eight months and that its closed-loop cooling uses no municipal or groundwater to cool the computing equipment. Reuters, which put the deal at about $1 billion, said the facility is expected to enter service in spring 2027.

Business Pill · Grid Cost Allocation

An evergreen Business Engineer explainer on who pays when new grid capacity is built for large loads. It teaches the general idea only and does not discuss Giga Energy or Texas.

The key insight: By our arithmetic, Giga Energy’s $1 billion, 10-year lease for 60MW is about $1.67 million per MW a year, roughly one-ninth to one-eleventh of the annualized GPU compute revenue per MW that Sharon AI’s chief executive described.

What Giga Energy Announced

Giga broke ground on Project Marl in July and will develop, own and operate the campus for the customer, the release says. It describes a 60MW GigaBase AI data center on an approximately 30-acre campus, about 2 MW per acre by our arithmetic, providing computing capacity for hyperscale and AI cloud customers.

The company says the data center campus will be ready for service in spring 2027, and that the project will finish in under 8 months, “less than half the time of traditional builds”. Reuters said the company’s factory-built construction approach is expected to allow completion in less than eight months.

Reuters referred to the facility as Project Clay; Giga’s own release calls it Project Marl. Giga says it has factories in Long Beach and Houston and more than 258.5 MW of critical IT capacity completed and under development.

Annual dollars per megawatt: Giga Energy’s $1 billion, 10-year lease for the 60MW Project Marl (our arit
Annual dollars per megawatt: Giga Energy’s $1 billion, 10-year lease for the 60MW Project Marl (our arithmetic, assuming a flat schedule, which Giga does not disclose) against the GPU compute contract rates Sharon AI’s chief executive gave on 16 November. A data-center lease and GPU compute are different layers.

What Was Posted on X

@choblin29 posted the deal at 15:38 UTC on 17 November: a 10-year, $1 billion lease with an undisclosed US AI cloud provider for a 60MW data center in Texas, built in under eight months with no municipal or groundwater used for cooling. We checked each point against Giga Energy’s own release and Reuters.

What the Rent Says per Megawatt

The $1 billion over 10 years is $100 million a year. Spread over 60MW, that is about $1.67 million per MW a year, or about $16.7 million per MW over the full term, by our arithmetic. The release does not give the rent schedule, escalators or who pays for power.

As we read the release, this is rent for a campus Giga develops, owns and operates; the release does not mention GPUs or chips. For comparison, Sharon AI, a neocloud that rents GPU compute, said its long-term contracts rose “from ~$15 million to more than $19 million annualized revenue per MW”, as our Sharon AI piece reported. Giga’s lease rate is about one-ninth to one-eleventh of those figures, by our arithmetic.

Hunt County already hosts another AI build in our coverage. Core Scientific has roughly 110 MW leased to AMD there for a fifteen-year term, at about $11.8 million of land, power and building cost per megawatt, by our arithmetic in our Core Scientific piece. Together the two projects come to about 170 MW.

The Water and Power Pitch

Rather than open-loop evaporative cooling, Project Marl uses a fully closed-loop system that operates independently of local water supplies, the release says. The only local water will be for sinks and employee bathrooms, “on par with the usage of two homes annually”, according to Giga.

On power, Giga says the project invests in new electrical infrastructure that taps underutilized capacity on dedicated power infrastructure in the county, and is designed so it does not draw on distribution capacity serving existing homes and businesses.

The release frames both points around Governor Greg Abbott, who it says recently called for data centers to pay for their own power, use water-efficient cooling technologies and prevent infrastructure costs from being passed on to Texas residents. President and co-founder Brent Whitehead said: “Communities shouldn’t have to choose between economic growth and protecting local resources.”

Giga lists about 300 construction jobs and 24 permanent full-time positions, with estimated annual salaries of $116,000 to $180,000, plus millions in new recurring annual tax revenue.

Bar chart: leased AI data-center capacity in Hunt County, Texas, in our coverage; Core Scientific about 110 MW leased to AMD for fifteen years; Giga Energy Project Marl 60 MW on a 10-year, 1 billion dollar lease; about 170 MW together
Two leased AI builds in Hunt County, Texas, in our coverage: Core Scientific’s roughly 110 MW for AMD and Giga Energy’s 60MW Project Marl, about 170 MW together, our arithmetic. Data: Giga Energy release, 17 November 2026; Core Scientific 8-K and July presentation, as reported in our earlier piece.

Against the Local Pushback

Water and power are where local opposition has focused. Amazon says the first three buildings in West Richland, Washington, would use an estimated 10 million gallons of water a year for cooling, and the city council voted 6-0 for interim rules that prohibit data centers in light industrial zones, as our Washington piece reported.

In Spain, Aragon’s draft decree lists a water limit of 0.1 litres per kWh, according to Hoy Aragon, in our Aragon piece. In Ohio, the FT said voters backed 21 initiatives to slow data center development (our Ohio piece).

In Texas, the Texas Tribune said Governor Abbott has a “current freeze on environmental permitting and grid connections” for data centers during an audit period, and the governor-elect’s campaign page calls for “an immediate moratorium on data centers to last until the legislature can meet to put in place real protections”, as our Core Scientific piece reported.

The Structural Read

As we read the release, the rent pays for the shell, power and operations of a campus Giga develops, owns and operates; the release does not mention GPUs or chips, so the far larger compute revenue per megawatt sits with whoever fills the halls.

As we read the release, Giga pitches the campus on the questions local opponents raise elsewhere: no municipal or groundwater for cooling, dedicated power infrastructure rather than distribution capacity serving homes, and a build of under eight months.

Brent Whitehead, president and co-founder of Giga Energy, in the company’s release, 17 November 2026

“Communities shouldn’t have to choose between economic growth and protecting local resources.”

Three Implications

FOR NEOCLOUD PRICING Sharon AI’s per-megawatt contract rates are in our Sharon AI piece.

FOR TEXAS Core Scientific’s Hunt County build for AMD and the Texas freeze debate are in our Core Scientific piece.

FOR LOCAL WATER FIGHTS Amazon’s Washington water figures and the West Richland vote are in our Washington piece.

The Business Engineer Lens

This story maps onto the Business Engineer framework The State of AI Data Centers.

The analysis says: “The structural bottleneck isn’t computing, it’s electrons.”

As we read it, Project Marl shows that bottleneck in a lease: Giga says the campus taps underutilized capacity on dedicated power infrastructure in Hunt County, and the rent works out to about $1.67 million per MW a year, by our arithmetic. The analysis predates the deal and does not discuss Giga Energy.

What Is Not Established

Giga does not name the AI cloud provider. The release does not give the rent schedule, the power contract, the build cost, the number of GPUs, or the source of the dedicated power infrastructure. Reuters and the release use different project names.

The per-MW figures are our arithmetic on a flat $1 billion over 10 years and 60MW; the actual payment schedule may differ. We did not contact Giga Energy for this piece. Not investment advice; for information only.

This Week In The AI Supercycle

AI Lost Its Central Banker, the Lab Oligopoly, and the Future of Open Source

This week’s issue: the Nvidia leadership change, read as the AI build-out losing its central banker: who now finances the GPU landlords, backstops the data centres and sponsors open models, and why two labs are becoming the bank.

Read the full issue →

Business Engineer Framework

The State of AI Data Centers

The Business Engineer analysis of the AI data-centre build-out and the power behind it.

Read the premium analysis →Get 50% off →

The Bottom Line

Giga Energy’s $1 billion, 10-year lease for its 60MW Project Marl in Hunt County works out to about $1.67 million per MW a year, by our arithmetic. The company says the campus will be ready for service in spring 2027 and will use no municipal or groundwater to cool its IT equipment.

95,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.

A note on sourcing. The lease, the project, the cooling, power, jobs and timeline statements are from Giga Energy’s release of 17 November 2026, distributed by GlobeNewswire; Reuters reported the deal the same day and called the facility Project Clay. Sharon AI, Core Scientific, Amazon, Aragon, Ohio and Texas figures come from our earlier pieces, linked. Per-megawatt figures and sums are our arithmetic. Giga Energy was not contacted for this piece. Not investment advice; for information only.

Sources: Giga Energy release (GlobeNewswire), 17 Nov 2026 · Reuters, 17 Nov 2026 · @choblin29 on X, 17 Nov 2026

Scroll to Top
FourWeekMBA