NAVWAR AI Data Center Plan Tests San Diego’s Housing Vision

The NAVWAR AI data center concept introduces a 400-megawatt computing facility where San Diego expected thousands of homes and a new mixed-use district.

Manchester Financial Group is exploring the facility as part of the 70.3-acre Navy Old Town Campus redevelopment. The developer previously emphasized housing, including possible homes for military families, alongside hotels, restaurants, retail, offices, and public spaces.

No data center has been approved. Manchester says the idea remains one of several possibilities, and its conversations with energy providers were preliminary. The distinction matters because early reporting sometimes describes the concept as a settled replacement for housing.

Yet the scale makes the discussion consequential long before a final plan exists. A 400-megawatt facility would create an extraordinary electricity requirement on land beside downtown, Interstate 5, and one of the region’s major transit centers.

That possibility turns the redevelopment into a contest between two uses for scarce urban land. One would create a neighborhood around housing and transit. The other would reserve substantial capacity for the physical infrastructure behind artificial intelligence.

Local officials are already pushing back. Their objections concern housing, electricity, transparency, and the limited authority San Diego holds over federally owned property.

The proposal therefore raises a larger question facing cities across the United States. When AI infrastructure reaches metropolitan neighborhoods, who decides whether its economic value outweighs the homes, services, and public life that could occupy the same land?

The NAVWAR AI Data Center Is Still a Concept, Not an Approved Project

The reported change is not a final land-use decision, but the disclosure of a computing project large enough to reshape every other part of the redevelopment.

NAVWAR, the Naval Information Warfare Systems Command, occupies aging industrial buildings near San Diego’s Old Town and Midway communities. The facilities trace their history to a World War II aircraft plant and no longer match the command’s modern operational needs.

The Navy wants new offices and technical facilities without relying entirely on a conventional congressional appropriation. Its preferred structure involves leasing part of the property to a private development team in exchange for a replacement NAVWAR complex.

The Navy selected Manchester/Edgemoor LLC as its potential master developer in January 2024. The selection began exclusive negotiations rather than granting final approval for a specific design.

A redevelopment overview published by the project team describes the site as approximately 70.3 acres. It identifies modern NAVWAR facilities as the project’s central purpose while allowing private development to support that exchange.

Public attention initially centered on a dense mixed-use neighborhood. A 2024 project summary described thousands of apartments, potentially including military housing, plus hospitality, retail, restaurant, and commercial uses.

The public narrative changed after records revealed that Manchester had asked local energy organizations about powering a 400-megawatt data center. According to the public records, one inquiry reached San Diego Community Power on November 10, 2025.

In that email, Manchester President and CEO Ted Eldredge asked for information before meeting potential data center investors in Doha. The message identified the NAVWAR property and the 400-megawatt requirement.

Those inquiries indicate serious feasibility work, but they do not establish a construction commitment. Manchester says no negotiations or energy agreements are underway.

“The potential plans for this site continue to be refined, and no final project has been approved,” the development team said in a statement reported by local media.

Eldredge has also described other potential uses, including a hospital, stadium, university campus, defense industry facilities, and life-science space. He said any final design would remain mixed-use and would include homes.

That leaves several basic questions unanswered. The developer has not disclosed the data center’s footprint, height, cooling system, customer, construction phases, or expected operating date.

It is also unclear whether 400 megawatts represents an initial load, a fully built campus, or a preliminary upper limit used for utility analysis.

Those details determine whether the data center can coexist with housing or would displace much of it. Until the development team releases a site plan, claims that housing has already been eliminated go beyond the available evidence.

The concept is still significant because infrastructure requests often influence a project before the public sees architectural drawings. A facility cannot reach 400 megawatts without transmission capacity, substations, interconnection studies, water planning, and firm customers.

The electricity question therefore becomes an early test of whether the idea is commercially realistic. It also exposes the scale of the choice facing the Navy and San Diego.

A 400-Megawatt Campus Changes the Meaning of Mixed Use

Calling the redevelopment mixed-use does not resolve the conflict when one proposed use carries an industrial-scale electricity demand.

A megawatt measures electrical power at a particular moment. Data centers require continuous power for servers, networking equipment, cooling systems, storage, lighting, and supporting infrastructure.

The proposed 400-megawatt load would place the NAVWAR concept among very large data center projects. Industry coverage has noted that it could become one of California’s largest facilities if fully developed.

That comparison still needs caution. Public reports often mix requested utility capacity, planned capacity, and actual operating load. A project can reserve a large connection while building only part of its announced campus.

Manchester has not identified an AI company, cloud provider, or colocation operator committed to the site. A colocation operator rents secured computing space and electrical capacity to multiple customers.

Manchester also lacks an established data center development record. Eldredge acknowledged that his company has historically developed hotels rather than computing campuses.

“We’re not in the data center space,” he told Voice of San Diego. He characterized the exploration as an attempt to find a financially attractive way to deliver the Navy’s building.

That statement clarifies the project’s economic mechanism. Private development must generate enough value to finance new federal facilities and make the wider transaction worthwhile.

AI demand may appear capable of producing that value. Developers and infrastructure funds are pursuing sites with access to land, power, fiber, financing, and supportive public policy.

However, demand for potential connections is not the same as demand for completed buildings. Energy researcher Jonathan Koomey described a speculative rush in which many proposed data centers will never reach construction.

His warning applies directly to the NAVWAR proposal. A 400-megawatt inquiry can help a developer test power availability, attract investors, or secure a place in an interconnection process.

It does not prove that an end customer will sign a long-term agreement. It also does not show that financing, equipment, cooling, or grid upgrades are available on an acceptable schedule.

The physical site creates additional complications. NAVWAR sits close to downtown, the airport, Interstate 5, existing communities, and the Old Town Transit Center.

That location is valuable because it connects housing and workplaces to rail, trolley, bus, and regional transportation. The same urban setting can make a large industrial computing campus harder to integrate.

Large data centers tend to create fewer daily trips than office towers, apartment districts, hospitals, or universities. That can reduce some traffic pressures, but it also weakens the case for reserving transit-adjacent land for them.

Their employment profile is similarly uneven. Construction can support substantial temporary work, while long-term operations often require fewer workers than a comparably sized mixed-use district.

A data center might still anchor defense, cybersecurity, or research activity related to NAVWAR. No public plan yet demonstrates that relationship, however.

The development team has not said whether the servers would handle Navy workloads, commercial AI training, cloud services, or several customer types. Without that information, references to an AI data center describe a market category rather than a defined mission.

The uncertainty matters for security as well as economics. A commercial campus beside a major military technology command would require careful separation of networks, physical access, utilities, and operational responsibilities.

It is possible to manage those boundaries. Yet the developer and Navy would need to explain how the arrangement supports NAVWAR without creating avoidable operational dependencies.

The data center concept therefore cannot be evaluated through capacity alone. Its value depends on the customer, connection plan, construction sequence, and contribution to the Navy’s financing structure.

Housing and Compute Capacity Are Competing for the Same Strategic Land

The primary conflict is not technology against progress; it is compute capacity against a rare opportunity for transit-oriented housing.

San Diego faces persistent pressure to add homes near jobs and transportation. The NAVWAR campus offers an unusually large, publicly controlled redevelopment site within the city’s urban core.

The Old Town Transit Center connects trolley lines, buses, commuter rail, and intercity rail. Interstate 5 runs beside the property, while downtown and San Diego International Airport sit nearby.

That geography made the site attractive for a dense neighborhood. Housing, shops, services, and offices could place more residents within reach of transportation that already exists.

Representative Scott Peters has argued that the property should become a “mini downtown” with homes and neighborhood-serving amenities. His congressional district includes the NAVWAR campus.

Peters said a data center belongs in a remote industrial location rather than on this site. His objection focuses on opportunity cost, meaning the value of the best alternative that San Diego would surrender.

Councilmember Marni von Wilpert raised a related concern about the development process. She said earlier briefings showed housing, parks, commercial space, and jobs but did not mention a data center.

“I found out in a newspaper article that they may be changing their mind,” she told FOX 5 and KUSI.

Manchester disputes the suggestion that it abandoned housing. Eldredge told local public media that the final project would include homes alongside retail, commercial, and industrial uses.

Both positions can be technically true. A final plan might retain housing while giving a significant portion of the site, electrical capacity, and development economics to a data center.

The crucial metric is not whether any homes remain. It is how many homes survive compared with earlier concepts, where they sit, and whether supporting amenities remain financially viable.

A data center can affect housing without occupying every available acre. Substations, cooling equipment, security setbacks, backup systems, and transmission infrastructure can shape adjacent development.

Noise also requires evaluation. Server buildings may appear quiet from a distance, but cooling equipment and backup generation can produce persistent sound during operation or testing.

A mixed-use plan would need credible buffers between those systems and residential buildings. It would also need emergency planning that accounts for both military and commercial infrastructure.

The Navy’s real-estate structure limits the city’s direct control. Because the federal government owns the land, the project does not follow the same local permitting path as an ordinary private development.

That does not eliminate environmental review or public participation. It does reduce San Diego’s ability to enforce its preferences through familiar zoning and approval processes.

The Navy’s project materials say it is updating environmental analysis with site-specific information from Manchester/Edgemoor. A revised draft environmental impact statement is expected in 2027.

That document should reveal whether a large data center enters the formal range of alternatives. It should also examine energy, water, air quality, transportation, noise, cultural resources, and cumulative effects.

The housing comparison must remain central throughout that review. Treating the data center as an isolated technical facility would conceal the land-use decision embedded inside it.

The relevant question is not whether San Diego needs AI infrastructure. It is whether this specific site delivers more public value as a major computing campus than as a transit-oriented neighborhood.

The Power Question Will Decide Whether the Proposal Is Real

The NAVWAR AI data center becomes credible only when its developers show where 400 megawatts will come from and who will pay for the required infrastructure.

Electricity is the first practical constraint. A facility of this scale cannot simply connect to the distribution lines serving nearby homes and businesses.

It would likely require major utility studies, new or expanded substations, transmission work, and carefully planned redundancy. Developers typically seek multiple power paths so a single failure does not interrupt customer operations.

Manchester contacted both San Diego Gas & Electric and San Diego Community Power while exploring the concept. These organizations perform different roles within the regional electricity system.

San Diego Gas & Electric owns and operates much of the local transmission and distribution infrastructure. San Diego Community Power procures electricity for participating customers through a community-choice model.

A data center may therefore depend on one organization for physical delivery and another for electricity procurement. That division complicates decisions about cost, capacity, clean-energy claims, and risk.

Officials have questioned whether the region can support the proposed load without affecting other customers. No public interconnection study currently answers that question.

There is also no disclosed agreement explaining who would fund new infrastructure. That issue matters because data center growth has triggered national debates over cost allocation.

If a utility expands its system for one unusually large customer, regulators must decide how much that customer pays. Poorly structured agreements can shift infrastructure risk toward other ratepayers.

The developer could address that concern through direct payments, dedicated assets, long-term contracts, or other protections. None has been announced for NAVWAR.

Clean energy introduces another layer. Purchasing renewable electricity over a year does not guarantee carbon-free power during every hour the servers operate.

A credible energy plan would identify the facility’s hourly load, renewable supply, storage, backup generation, and periods of grid dependence. It should also disclose whether operations can reduce demand during emergencies.

Water use remains uncertain because Manchester has not identified a cooling design. Some facilities use evaporative systems that consume water, while others rely more heavily on air cooling or closed-loop equipment.

Each option involves tradeoffs among electricity, water, cost, climate, and local conditions. San Diego cannot assess those tradeoffs until the developer specifies the proposed system.

Backup power also deserves scrutiny. Traditional data centers often use diesel generators to maintain operations during grid failures.

A large campus can require numerous generators, fuel storage, and regular testing. Those systems may create localized emissions and noise even when grid electricity supplies normal operations.

Alternative backup technologies are available, including batteries, fuel cells, and different generator fuels. The project team has not committed to any particular configuration.

The grid question extends beyond local engineering. Data center companies have announced enormous power needs across the country, but not every request represents a project ready to build.

Utilities increasingly need evidence that customers can finance construction and use the capacity they request. Otherwise, speculative projects can crowd interconnection studies and distort long-term planning.

That is why the NAVWAR inquiry should be treated as a feasibility signal rather than proof of an inevitable facility. The strongest evidence will come from formal utility applications, deposits, system studies, and signed customer agreements.

Until those appear, 400 megawatts describes ambition more clearly than deliverability.

Federal Control Makes Transparency the Immediate Test

The greatest near-term risk is a major land-use choice advancing through private and federal negotiations before residents can evaluate a complete plan.

The Navy’s objectives are understandable. NAVWAR operates from facilities that the project team describes as obsolete, and replacing them through ordinary federal funding could be slow or uncertain.

A public-private redevelopment can unlock land value while delivering modern government offices. The Navy used a related structure for its downtown Broadway complex.

That financing model also creates tension. The private uses that generate the highest return do not automatically provide the greatest value to nearby communities.

Housing can deliver residents, customers, public spaces, and support for transit. It also creates schools, traffic, infrastructure needs, and long-term service obligations.

A data center may generate lease value and investment with fewer residents. Yet it can consume enormous electrical capacity while contributing limited street activity.

Choosing between those outcomes requires public information about both benefits and costs. Residents currently lack enough detail to make that comparison.

The original 2021 environmental review evaluated several broad redevelopment alternatives. Those included a NAVWAR-only option and denser public-private development scenarios.

The Navy says it is revising that work using the selected developer’s site-specific information. The revised analysis is expected to receive public review and comment before a final environmental statement and decision.

That process offers an important checkpoint, but environmental documents do not automatically settle questions of public trust. The development team can improve confidence before 2027 by releasing its assumptions.

Useful disclosures would include a current land-use program, anticipated housing range, data center footprint, energy request, utility upgrade concept, cooling method, and development timetable.

The team should also clarify what changed between earlier public presentations and the November 2025 power inquiry. That timeline would show whether the data center became a serious option or remained a preliminary financial exercise.

Manchester says its discussions with energy providers were informational and produced no negotiations or agreements. That statement narrows the claim but does not explain why elected officials were unaware of the inquiry.

Councilmember Sean Elo-Rivera has said the concept was not disclosed when he asked San Diego Community Power about proposed data centers. As a member of the agency’s board, he has sought an explanation.

Those concerns do not prove improper conduct. They reveal a governance gap created by overlapping federal, private, utility, and local responsibilities.

The Navy controls the land. Manchester/Edgemoor is developing the plan. Utilities control essential infrastructure. San Diego officials represent affected communities but possess limited permitting power.

When authority is divided this way, disclosure becomes more important. No single local hearing can guarantee that all decisions reach the public at the same time.

The federal setting may also change the political calculation. Local opposition cannot directly veto a project in the same way that a city council might reject a private rezoning request.

However, the Navy still needs a viable development partner, environmental compliance, utility cooperation, financing, and a defensible public record. Sustained controversy can influence each of those requirements.

The development team therefore has an incentive to present evidence early. A project described mainly through leaked emails will face greater resistance than one supported by specific, testable commitments.

Three Signals Will Show Which NAVWAR Future Is Advancing

The next phase should be judged through formal documents and commitments, not renderings, exploratory emails, or broad assurances.

The first signal is the revised draft environmental impact statement expected in 2027. Its range of alternatives will show whether the Navy formally treats a large data center as part of the redevelopment.

Readers should look for the facility’s stated capacity, footprint, construction phases, cooling design, backup systems, and effects on housing. A vague industrial category would leave the central conflict unresolved.

If the review includes detailed analysis of a 400-megawatt campus, the proposal has advanced beyond informal exploration. If it omits the concept, the reported inquiry likely remains outside the Navy’s active plan.

The second signal is a formal electricity pathway. That could include an interconnection request, utility study, infrastructure agreement, customer deposit, or procurement arrangement.

Such records would show that the developer is moving from investor discussions toward engineering and commercial commitments. They would also reveal whether the full 400-megawatt figure remains intact.

Any power plan should state who pays for transmission and substation upgrades. It should explain protections for existing customers and identify the expected energy supply.

A credible filing would also distinguish total reserved capacity from expected operating demand. Without that distinction, the headline number remains difficult to interpret.

The third signal is a site plan that gives housing a measurable place. Manchester has said the eventual redevelopment will still include homes.

The public should therefore expect a housing range, building locations, phasing schedule, affordability approach, and relationship to the Old Town Transit Center. A promise of mixed use is too broad.

If housing remains central and the data center occupies a defined supporting role, the redevelopment may preserve its neighborhood vision. If housing falls sharply, the proposal represents a clear land-use reversal.

These signals will also show who is actually behind the computing demand. A named operator or committed customer would make the project more credible than an unidentified investor group.

Until then, the NAVWAR AI data center should be understood as a high-impact proposal with a substantial verification gap. It is neither an approved replacement for housing nor an idea that officials can dismiss as harmless brainstorming.

San Diego’s choice reaches beyond one redevelopment. Cities need housing, military modernization, grid reliability, and the infrastructure supporting AI services.

They cannot assume those goals fit together automatically on every parcel. Land beside major transit carries a different public value from remote industrial property.

The Navy and Manchester/Edgemoor now have an opportunity to make the comparison concrete. They can disclose what the computing campus requires, what it finances, and what the city gives up in return.

Residents, electricity customers, and technology buyers should watch the environmental review, power filings, and housing plan in that order. Those records will reveal whether the NAVWAR concept is an investable project or another speculative entry in the AI infrastructure rush.

The decisive question is straightforward: can the developers demonstrate that this 400-megawatt proposal serves NAVWAR and San Diego better than the housing-centered community residents were shown?