LaCross AI Institute Series: Data Center Development-Opportunities and Constraints
Published 09 March 2026. AI's surge since 2022 has fueled an unprecedented buildout of data centers-now topping 11,000 worldwide.
On 17 February, the LaCross Institute for Ethical AI in Business at the UVA Darden School of Business gathered industry experts, faculty, students, staff, and alumni to map where development is headed and what it will take to deliver. Sessions zeroed in on the work: infrastructure, capital, energy, water, technology, and talent.
A value-chain lens for ethical AI
Institute director Marc Ruggiano outlined a practical frame for decisions across AI's lifecycle so teams can place ethics and outcomes in context and find wider solution sets. The value chain spans five areas: infrastructure; measurement and data; algorithms, models, and training; new business applications; and managing AI and monitoring outcomes-a focus of a forthcoming white paper.
- Infrastructure
- Measurement and data
- Algorithms, models, and training
- New business applications
- Managing AI and monitoring outcomes
The building boom is real
Dan Ephraim (PointOne Data Centers) and Ed Socia (DatacenterHawk) broke down what it takes to stand up capacity: the site and entitlements, chips, storage, fiber, optics, networking, and the literal build-drywall, generators, switchgear, concrete, and mechanical-electrical-plumbing. "We are the plumbers of the internet," Ephraim said, noting the sector's need for people who can sell, develop, operate, and manage these assets.
Capacity is measured in megawatts to gigawatt-scale, not square feet. New builds are larger from day one, intensifying the primary constraint: available electricity. As Socia put it, "We used to say bring the power to the data center. Now we say bring the data center to the power."
The power challenge and water questions
Energy finance consultant Patrick Worrall (MBA '02) cited U.S. peak load at roughly 530 gigawatts. Adding even tens of gigawatts for a single site stresses a system that's already tight. It's industrial growth at a scale the grid wasn't planned to absorb quickly.
Andy Stewart (Iconic Digital) called it a step change in demand, misaligned with utility planning cycles of 10-20 years versus 10 weeks to two years for data centers. Developers are blending sources-wind, solar, and natural gas-based on use case and reliability needs, Worrall said. Water-saving approaches are advancing too: closed-loop cooling, retention ponds, and pump station upgrades help local systems keep pace. For broader grid context, see the U.S. Energy Information Administration.
Financing at enterprise scale
Joe Harar (UVA '05), CFO of EdgeConneX, projected about $7 trillion in global data center spend by 2030, with 40% in the U.S. The seven largest companies account for roughly 90% of demand, often on 10-15 year leases.
Traditional real estate and infrastructure financing can't shoulder projects this large. With higher capital costs and longer debt service, developers are leaning into private equity-now the sector's largest investment category. As Irtiaz Ahmad (Sobina Advisors) noted, operating outside public markets helps when you must think in 5-15 year horizons, not quarters.
The technology stack: density up, timelines tight
GPU racks keep pushing density and facility design limits. For large training clusters, "we need all of it together: GPUs, storage, CPUs, and networking," said Emerson Whitney (MBA '24), Microsoft Azure supercomputing group.
Shortages of organic data are boosting CPU demand for synthetic data generation. Silicon cycles move faster than construction, so many sites go live a generation behind the latest chip-but that rarely delays go-live for customers ready to train.
Talent: the constraint you feel every day
Supply chain, power, and talent are now co-equal constraints. Recruiters emphasized candidates who thrive in matrix environments, think strategically, and bring hands-on experience-internships, projects, or operations.
Open roles are everywhere: construction, real estate, site development, project management, facility engineering, plus finance and insurance. "The supply chain is very constrained right now. That's our biggest risk," said Trey Dean (Vantage Data Centers).
For larger tech firms, hiring often starts internally, noted Abbey Kang (EMBA '21, ServiceNow). Get in, then move toward your target role. "Put your hand up," added Kendall Jennings Hayden (MBA '12, Accenture). Curiosity and initiative travel well in teams that re-form every year.
What this means for IT and development
- Engineer for higher rack density and thermal limits; plan for liquid cooling and evolving footprints.
- Design east-west bandwidth for training workloads; expect fiber and optics lead times in schedules.
- Model capacity around mixed GPU + CPU needs, including synthetic data pipelines.
- Instrument energy, water, and reliability telemetry; feed compliance and cost reporting automatically.
What this means for real estate and construction
- Sequence sites around substations, interconnects, and queue positions before land closes-bring the data center to the power.
- Secure long-lead gear early: generators, switchgear, transformers, chillers, and MEP components.
- Co-create plans with utilities and municipalities for grid upgrades, water supply, and wastewater; show closed-loop cooling and on-site retention where feasible.
- Match capital to lease terms; favor longer tenors and private vehicles built for multi-year build-operate programs.
Next steps and learning
For practical workflows and skills you can apply now, explore AI for IT & Development and AI for Real Estate & Construction. Both focus on high-utility tactics for teams building and running data-centric infrastructure.
About the University of Virginia Darden School of Business
The University of Virginia Darden School of Business prepares responsible global leaders through transformational learning. Darden's graduate degree programs (Full-Time MBA, Part-Time MBA, Executive MBA, MSBA and Ph.D.) and Executive Education & Lifelong Learning support long-term career impact. Darden has Grounds in Charlottesville and the Washington, D.C., area and a global community of 20,000 alumni in 90 countries; it was established in 1955 at the University of Virginia, founded by Thomas Jefferson in 1819 in Charlottesville, Virginia.
Press Contact
Molly Mitchell
Senior Associate Director, Editorial and Media Relations
Darden School of Business, University of Virginia
MitchellM@darden.virginia.edu
Your membership also unlocks:
