Research desk synthesis · Inaugural source-register review, October 30, 2026. This feature is an authored interpretation of previously published, linked material. It is not a report of a newly announced October 30 transaction. Original source dates and unresolved project stages remain part of the evidence.
Begin with the business, not the building. Capital One’s corporate page identifies McLean headquarters, while its AI page describes research and applications in customer service, personalization, and software security. Appian’s official site likewise connects McLean headquarters with AI-enabled process automation. These records show that Virginia’s AI coverage needs an enterprise-software and applied-research lens as well as an infrastructure lens. [Capital One; Capital One; Appian]
Government technology adds another layer. Booz Allen’s corporate fact sheet identifies McLean headquarters and federal AI services. Leidos’s January 22, 2026 announcement describes a partnership with OpenAI and identifies Reston headquarters. The latter is a dated company announcement about planned integration; it does not independently verify the results of every customer deployment. [Booz Allen Hamilton; Leidos]
Our interpretation: a headquarters address, an AI capability, and a local economic impact are three separate claims. The first two can be supported by these company records. The third needs additional evidence about where work occurs, contracts, suppliers, hiring, and measured results. A credible Virginia research desk should connect the questions without turning that connection into an unsupported impact estimate. [Capital One; Capital One; Booz Allen Hamilton; Leidos]
University activity is equally broad. UVA’s computer science page identifies multiple AI research areas. Virginia Tech’s Sanghani Center describes interdisciplinary AI and data analytics work. George Mason documents research, education, and responsible adoption. These institutional descriptions identify capabilities and places to investigate; they are not interchangeable with a technology license, a startup financing, or an operating data center. [University of Virginia; Virginia Tech; George Mason University]
The geographic lens should extend beyond Northern Virginia. VCU maintains an AI and Data Science research cluster in Richmond. Old Dominion identifies AI and machine learning as a domain supporting its strategic research areas in Hampton Roads. Our interpretation is that regional coverage should follow evidenced institutional activity rather than assuming that every part of the AI economy follows the same campus-development pattern. [Virginia Commonwealth University; Old Dominion University]
Infrastructure announcements need their own ledger. Google’s August 27, 2025 statement describes a Virginia investment program through 2026; AWS’s live commitment page describes a separate expansion plan through 2040. The periods, scope, and underlying activities differ. We therefore do not sum those announcements into a verified amount spent, a campus valuation, or a measure of current Virginia IT capacity. [Google; Amazon Web Services]
An operating reference is narrower but more concrete. Microsoft explicitly identifies Boydton operations. Google identifies its Loudoun and Prince William data centers and discusses ongoing Northern Virginia operations. Even so, those sources do not establish the commissioning status, load, employment, or capital cost of every building or expansion associated with an operator. [Microsoft; Google]
A campus specification has a different evidentiary role. QTS’s Richmond page supplies a critical campus capacity figure and a master-plan description. Vantage’s VA3 page supplies critical IT buildout specifications. Such owner-published material helps describe the product being offered; it cannot by itself establish the electrical capacity already delivered by a utility or the completion of every planned phase. [QTS Data Centers; Vantage Data Centers]
Watch the dates inside the source. The live QTS Ashburn page still combines an under-construction description with a Q2 2026 target for DC2. Our October review did not find completion evidence in that source. Treating the target as an accomplished milestone would quietly convert a schedule into a fact, so the register keeps the stage unresolved. [QTS Data Centers]
For property diligence, power is a separate workstream. Dominion’s intake guidance describes the load letter, ramp schedule, and subsequent infrastructure steps. Our interpretation is that a credible underwriting file should tie its assumptions to the project-specific service plan, required infrastructure, timing, and commercial obligations. Proximity to a line, a developer specification, or a request alone cannot do that work. [Dominion Energy]
Regulated service also has an allocation question. The SCC’s fact page describes large-load customer safeguards and cost-allocation initiatives. Our interpretation is that developers and property investors should examine the controlling tariff and order for their own project rather than applying a broad policy summary as a universal rate quote or guarantee of service. [Virginia State Corporation Commission]
Land and environmental evidence should travel together. Loudoun’s guidance documents its March 18, 2025 land-use amendments, while DEQ publishes a distinct air-permit trail. Our interpretation is to maintain separate approval states for entitlement, construction, generation, utility service, and operation. A successful step in one track should not silently complete the others. [Loudoun County; Virginia Department of Environmental Quality]
Tax assumptions deserve the same discipline. VEDP’s live exemption guidance describes qualifying equipment, agreements, performance requirements, and reporting. Our interpretation is that the relevant benefit belongs in a project model only with its actual qualification and timing assumptions made explicit. A marketed location or operator announcement is not proof that a particular buyer, tenant, purchase, or project qualifies. [Virginia Economic Development Partnership]
The human and institutional pathways complete the picture. NOVA and Microsoft publish data center training routes; VITA maintains agency AI governance resources, and JLARC’s 2026 review recommends improvements. Our interpretation is that Virginia’s opportunity should be tracked through accessible training, responsible adoption, company activity, and delivered infrastructure together. Keeping those records distinct makes their connections more useful to readers and real-estate decisions. [Northern Virginia Community College; Microsoft; Virginia Information Technologies Agency; Virginia Joint Legislative Audit and Review Commission]