top of page

Soluna Plans Willacy County AI Data Center

3 hours ago
12 min read
Illustrative photograph. Photo: panumas nikhomkhai / Pexels.

Soluna Holdings is advancing a planned AI facility in Willacy County, Texas, while major commercial commitments remain unfinished. The Soluna AI data center proposal pairs a large local investment estimate with unresolved questions about financing, capacity, and delivery. A November 16 company update adds an important development: a prospective tenant has signed a nonbinding agreement, while definitive leases remain a priority.

The proposal puts county officials, prospective customers, and Soluna’s development team in different positions. Officials must assess public benefits and resource commitments. Customers need dependable computing capacity on an enforceable schedule. Soluna must connect those expectations through contracts, engineering, and financing.

The central tension is the distance between development progress and an operating AI facility. That distinction matters because existing Bitcoin hosting at Kati 1 provides a local reference, while Kati 2 introduces different customer requirements. Soluna’s separate collaboration with Applied Digital at Dorothy 3 offers another comparison, but it does not establish commitments for Willacy County.

What the Soluna AI data center proposal includes

Kati 2 is a proposed expansion with substantial ambitions, rather than an operating facility or a completed financing announcement.

In a Kati 2 interview published November 13, CEO John Belizaire described an investment approaching $4 billion, including buildings and computers. The Rio Grande Guardian reported approximately 40 permanent positions and roughly 1,200 construction workers during an estimated 14-month build.

Those figures describe management’s expectations. They do not establish capital already raised, equipment purchased, or employment already created. The investment estimate also includes computers, which matters when assessing who would supply capital and own the resulting assets.

According to that interview, the project would draw power from the neighboring Las Majadas wind development. Belizaire described liquid cooling using an offsite-supplied fluid mixture. The reported schedule targeted early 2027 construction and a second-quarter 2028 opening, conditional on approvals.

These details give residents a concrete proposal to evaluate. However, an investment estimate cannot answer every question about the eventual facility. Its composition matters alongside its headline size.

A building, electrical infrastructure, and computing equipment serve different purposes within the same development. Their ownership, financing, and replacement arrangements can differ. Public discussion should therefore distinguish the physical campus from the equipment installed by its eventual users.

That distinction also affects how readers interpret the announcement’s scale. A proposed investment containing computers is broader than a construction budget for buildings alone. Comparing those figures directly would obscure what each project estimate actually measures.

Employment requires similar care. Construction work and permanent operations represent different economic benefits. A temporary building program can create demand for local accommodation, transport, and contractors without producing an equivalent permanent workforce.

For residents, neither category should disappear into the other. Construction opportunities deserve attention on their own terms. Permanent positions deserve separate scrutiny around hiring requirements, training, and whether local applicants can realistically qualify.

The proposal’s immediate significance is therefore specific. Soluna has presented a larger prospective role for its Willacy County campus, with consequences for local planning and infrastructure discussions. That does not settle the eventual ownership structure, operating footprint, or public return.

Readers should also separate the publication date from the development’s physical timeline. An interview published in November can describe a future construction target while later corporate disclosures report earlier site preparation. Those statements require reconciliation rather than automatic acceptance as an identical schedule.

Power access gives Soluna a starting position

Existing power infrastructure strengthens the development case, but its capacity figures need careful interpretation.

Soluna’s Kati campus update dated September 8 said Kati 1 had reached 83 megawatts of energized capacity. A megawatt measures electrical power, rather than computing performance or annual electricity consumption.

The same disclosure described a conditional classification covering the full 166-megawatt Kati campus. ERCOT, the Electric Reliability Council of Texas, administers the relevant Texas grid process. Soluna said verification remained outstanding.

This creates an important distinction from the local interview’s presentation of 166 megawatts for Kati 2. The corporate release attached that figure to the wider campus, including its existing first phase. Readers should not treat those descriptions as interchangeable.

Capacity terminology can change the apparent size of a project without any physical change occurring. A campus total, an individual building’s electrical demand, and power allocated to computing equipment answer different questions. A headline number needs its boundary.

The distinction becomes more important when future expansion enters the discussion. An ultimate development plan can exceed the capacity already covered by an existing interconnection position. That larger ambition does not automatically inherit every status attached to the earlier footprint.

For county officials, a clearly defined electrical boundary supports better public discussion. Residents need to know which facilities a capacity figure covers. Prospective customers need to know how much capacity belongs to the phase they would occupy.

Soluna’s location strategy starts with an understandable proposition: develop computing close to electricity generation. That can provide a different starting point from identifying land first and then seeking a power arrangement. It does not remove the need to establish operating conditions.

Curtailment means reducing electricity generation when the system cannot accept or economically use the available output. Using otherwise curtailed generation offers a potential business opportunity. However, the amount available at a particular hour remains relevant to a customer’s operating requirements.

AI workloads make that distinction commercially important. A customer reserving infrastructure needs a service commitment, rather than a general description of nearby generation. The contract must connect the power arrangement to the computing service being purchased.

Consider an enterprise planning a scheduled model-training run. Its procurement team needs to understand whether interruptions affect completion dates, reservation terms, or access to replacement capacity. A campus location alone cannot answer those questions.

The same applies to an application serving users throughout the day. Available electricity and dependable application service are related, but they are different commitments. Buyers need evidence connecting the facility design to their specific workload.

This is where the Soluna AI infrastructure impact becomes tangible. The project’s relevance depends on converting a location advantage into a defined, usable service. Preserved interconnection capacity helps establish the starting position, while customer obligations determine what must follow.

Tenant interest still has to become a lease

The clearest commercial development is a prospective tenant agreement whose nonbinding status remains central to the story.

Soluna’s November business update reported a new letter of intent covering 118 megawatts of IT capacity at Kati 2. IT capacity means electricity allocated to computing equipment, rather than the facility’s entire electrical demand.

The company described a planned campus exceeding 350 megawatts. It also said the previous prospective tenant’s exclusivity period had ended before the site returned to marketing. The new agreement therefore represents renewed commercial interest, rather than continuation of an already completed lease.

A letter of intent records proposed terms or negotiating intentions. Its nonbinding status means readers should not equate it with an enforceable commitment to occupy the facility. The remaining negotiations can determine whether the development advances on its proposed terms.

For Soluna, this places commercial execution at the center of the next phase. Engineering work can proceed while customer requirements evolve. However, unresolved tenant decisions can affect the final design, delivery obligations, and financing structure.

For the prospective tenant, the negotiation works in the other direction. The customer must assess whether the site can meet its requirements before accepting a long-term obligation. Interest in a location does not eliminate due diligence.

A definitive lease would help connect these interests. It could establish the contracted footprint, delivery milestones, and allocation of responsibilities. Its actual terms would matter more than the announcement’s use of a familiar commercial label.

The latest disclosure also reported construction beginning in October. That differs from the early 2027 target presented in the local interview. Early site work and the main building program can represent different milestones, but the available wording does not fully resolve their relationship.

Readers should therefore avoid choosing whichever date makes the project appear furthest advanced. A useful next disclosure would identify completed work, remaining approvals, and the schedule for customer-ready delivery. Those milestones would make progress easier to assess.

Applied Digital provides a relevant industry comparison through Soluna’s separate Dorothy 3 project. The November update described a planned joint venture there, with a prospective customer arrangement also subject to conditions. That collaboration belongs to another site.

Its relevance is structural rather than transferable. A development partner can contribute capabilities or commitments within a particular agreement. It does not automatically become the partner, financier, or tenant for every project in the developer’s portfolio.

This distinction protects readers from a common mistake in infrastructure coverage. Several announcements can share a company name while involving different land, power arrangements, counterparties, and closing requirements. Combining them produces a project that no individual disclosure actually describes.

With Soluna Kati 2 explained through its contracts, the project becomes easier to follow. The key question is whether prospective demand becomes an obligation that supports construction and financing. Announced interest is meaningful evidence, but its legal status determines how much weight it carries.

The commercial pressure is therefore concrete. Soluna needs a customer agreement that supports delivery. The prospective customer needs sufficient confidence in delivery to sign. Progress depends on resolving that relationship.

Financing must follow the project’s actual scope

A development estimate describes intended spending; a financing structure establishes how that spending becomes possible.

In a project financing discussion, Soluna said Kati 2 funding would combine project-level and parent-company financing. Management described an expected debt-to-equity mix near 80/20, while explicitly making the final structure dependent on market conditions and customer requirements.

That discussion also tied the construction schedule to tenant requirements and a signed lease. It provides useful context for the current proposal. However, an earlier financing expectation should not become a claim that financing has closed.

Project-level financing means funding arranged around a particular development and its expected obligations or cash flows. Parent-company financing places the funding at the corporate level. The distinction matters because the risks and claims on assets can differ.

For readers assessing the Soluna AI data center proposal, the important point is sequence. A proposed customer agreement can inform the facility design. A defined design and contracted revenue can then support discussions with capital providers.

That sequence is not necessarily a rigid timetable. Developers can negotiate several workstreams together. Still, progress in one workstream should not be reported as completion of every other dependency.

A prospective lender, for example, would need to understand what the facility is expected to deliver. A customer would need confidence that funding supports those obligations. Equity investors would need to understand the capital required and the risks they retain.

The near-term reporting task is consequently more demanding than repeating an investment estimate. Useful disclosures would identify the financed phase, the parties providing funds, and any remaining conditions. They would also distinguish the developer’s spending from customer-supplied equipment.

That last distinction matters for the local economic story. Equipment installed by a tenant can contribute to a development’s stated scale. It does not necessarily mean the developer must finance every component through its own balance sheet.

The reverse assumption would also be premature. Readers should not assume a tenant will purchase all computing equipment unless a disclosed agreement establishes that responsibility. Ownership and funding remain factual questions requiring specific answers.

Financial structure can also affect the timing of construction. A project can advance design or preliminary work before securing the complete funding package. Such activity demonstrates development effort, while leaving the full delivery obligation unresolved.

This is an analytical distinction rather than evidence that the project will fail. Infrastructure development normally contains conditional steps. The concern arises when coverage removes those conditions and presents the final outcome as already secured.

The strongest assessment therefore stays close to the documents. A signed financing commitment would strengthen confidence in the funded scope. An announced funding strategy establishes management’s intended approach, with execution still ahead.

For enterprise buyers, this financial detail has an operational consequence. A reservation at an unfinished facility depends on the developer’s ability to complete it. Procurement teams should connect contractual delivery promises with evidence that the underlying development can be financed.

Water and local benefits need defined boundaries

The environmental and economic case requires measurable commitments that distinguish facility operations from broader project effects.

Belizaire’s local interview described a closed liquid loop containing glycol-enhanced water supplied from outside the community. He said the planned cooling arrangement would avoid groundwater use for daily computer cooling. Those statements remain company descriptions of the proposed design.

Direct liquid cooling moves heat from computing components through circulating fluid. The presence of liquid inside a server system does not, by itself, establish the facility’s total water consumption. Heat still has to leave the wider cooling system.

The Department of Energy’s cooling water guidance explains that water consumption depends on heat loads and the efficiency of heat removal. Its guidance concerns systems using cooling towers, rather than confirming Kati 2’s specific design.

That distinction is essential. General cooling guidance can identify questions to ask, but it cannot establish which equipment Soluna will install. The final facility design and operating disclosures must supply those answers.

A useful resource statement would identify its boundary. Daily drinking water, construction water, cooling-system fill, and replacement fluid represent different uses. Combining them into a single reassuring phrase would make verification harder.

Offsite supply also needs precise interpretation. Transporting fluid to the facility changes where it enters the project. It does not mean no water was required to produce or supply that fluid.

Similarly, a claim about local groundwater does not establish the entire energy supply’s water footprint. Each claim should remain attached to the resource and activity it actually describes. Wider conclusions require wider evidence.

There is no need to dismiss a low-water design because those questions remain. A design that reduces local demand can offer a meaningful benefit. The reporting standard is to identify the benefit accurately and avoid expanding it into an unsupported universal claim.

Power claims deserve the same treatment. Locating beside wind generation does not independently establish every hour’s electricity source or the facility’s effect on the grid. Those outcomes depend on operating arrangements and system conditions.

Local benefits present another set of boundaries. The interview described proposed county tax arrangements still under negotiation. A payment in lieu of taxes is a negotiated payment associated with a tax arrangement, rather than ordinary full taxation.

The public interest depends on the actual agreement. Officials and residents need to understand the covered assets, duration, payment obligations, and consequences if promised development does not occur. A proposed arrangement cannot substitute for its eventual terms.

The appropriate skeptical perspective is therefore specific rather than oppositional. Public benefits should be compared with the commitments that make them enforceable. Neither a promotional projection nor an assumed worst case supplies that comparison.

The Soluna AI infrastructure impact will ultimately depend on execution at this local level. Hiring opportunities, contractor participation, resource reporting, and negotiated payments determine how the project connects with its host community. An investment headline cannot measure all of them.

Residents should also have a clear way to distinguish company commitments from expected spillover benefits. An obligation to make a payment differs from an expectation that nearby businesses will gain customers. Both matter, but only one necessarily appears in a contract.

Three developments will clarify Kati 2’s trajectory

The next useful evidence is a definitive tenant commitment, a clearly funded construction phase, and documented local approval terms.

First, watch whether the prospective tenant signs a definitive lease. The most informative announcement would identify the contracted capacity and relevant delivery conditions. That would strengthen the case that customer interest supports an executable development.

A lease announcement would still require careful reading. Conditions, phased delivery, and responsibilities can affect what the commitment means. Readers should assess the obligation described rather than assuming the word “lease” resolves every uncertainty.

Second, watch for financing attached to a defined construction scope. A funding disclosure becomes more useful when it identifies which phase can proceed and what remains conditional. That would connect the investment ambition with a practical delivery program.

Construction updates should support the same assessment. Completed earthworks, building construction, equipment installation, and customer-ready operation are different milestones. Reporting them separately would help reconcile the project’s earlier and later timeline descriptions.

Third, watch for Willacy County’s documented decisions and negotiated terms. Public records should clarify the status of any proposed abatement and associated payments. That would allow residents to evaluate the agreement actually offered to their community.

A decision at Dorothy 3 would not answer this final question. County-specific approvals belong to the project and jurisdiction named in the documents. Shared ownership does not make those permissions interchangeable.

These signals matter because the wider electricity challenge is measurable. The Energy Information Administration’s server electricity outlook projects increasing server consumption across commercial buildings, with greater growth in standalone data centers. That outlook explains the stakes without validating any individual development.

For developers and enterprise buyers, the practical implication is to track usable capacity rather than announced scale alone. A future campus can be relevant to a procurement strategy while remaining unsuitable for an immediate deployment commitment.

A pharmaceutical research team considering a future computing reservation illustrates the distinction. It would need an agreed service date, workload requirements, and remedies for delay. The local interview identifies pharmaceutical research as a prospective application, rather than naming an existing customer.

That example also explains why nontechnical readers should care. The infrastructure behind an AI service influences when capacity becomes available and what obligations suppliers can accept. Development uncertainty can travel through the supply chain to the organizations buying computing services.

With Soluna Kati 2 explained as a sequence of verifiable commitments, future announcements become easier to judge. A larger capacity figure changes the ambition. A signed lease changes the commercial position. Closed financing changes the resources available for delivery.

Keep those categories separate when following the Soluna AI data center proposal. Read the next tenant announcement alongside the construction scope and county documents. Check whether each new commitment covers the same phase, capacity boundary, and delivery date. Ask which conditions have disappeared and which remain outstanding. For residents, focus on enforceable local benefits and resource commitments. For buyers, focus on service readiness and contractual protection. For investors, focus on who supplies capital and carries completion risk. The most useful next question is concrete: what has become binding since the previous update?

Give every agent the context to do better work

Connect your agents to the knowledge, decisions, and history already organized in remio.

remio currently supports Windows 10+ (x64) and Macs with Apple silicon.

Your AI Partner at Work
Get more done with remio

Plan. Create. Deliver.
All in one place.

bottom of page