Loading...

Meta's $10 Billion Indiana Data Center Campus: The IoT Infrastructure Behind AI

Smart Buildings

Published by IOT Indiana Research & Editorial Team

Meta's $10 Billion Indiana Data Center Campus: The IoT Infrastructure Behind AI

In February 2026, Meta broke ground on a data center campus in Lebanon, Indiana, with a reported investment of more than $10 billion and a planned capacity of one gigawatt. The project sits in Boone County's LEAP Lebanon Innovation and Business District northwest of Indianapolis. It is Meta's second Indiana data center, following the one in Jeffersonville.

Most coverage asks what it means for jobs and tax revenue. This article asks a different question: what has to be sensed, measured and controlled for a facility like this to work? Meta has not published design details for the campus, so what follows is a description of how large AI data centers generally operate, framed for readers who work in controls, energy, construction or facilities in Indiana.

What is publicly confirmed

Reports from the groundbreaking identify Mortenson and Turner Construction as the contractors. The Boone County Economic Development Corporation announced about $45 million in road improvements and $75 million in public water infrastructure tied to the area. Meta said it plans habitat improvements and runoff reduction at nearby Deer Creek, and the groundbreaking launched the Boone County Career Collaborative, a workforce program linking students with local employers. Construction employment of several thousand workers has been cited in project coverage.

Details that have not been confirmed in the sources reviewed include the final cooling design, the exact power supply arrangement and a completion date. We will not guess at those. Where this article says what a data center "typically" does, it refers to industry practice, not to Meta's confirmed plans.

Why AI changes the instrumentation picture

A conventional enterprise data center might draw 5 to 10 kilowatts per rack. AI training and inference clusters built on dense accelerator servers can draw many times that, with some designs moving to liquid cooling because air alone cannot remove the heat. At one gigawatt of campus capacity, even small percentage losses in power or cooling become large absolute numbers.

That density has three consequences for sensing:

  • Thermal margins are thin. A cooling failure can push hardware toward protective throttling or shutdown within minutes, so temperature and flow data must be fast and reliable.
  • Power is the product. Operators track electricity at many levels, from utility feeds down to racks, because capacity planning and cost depend on it.
  • Everything is interlinked. A change in server load shifts cooling demand, which shifts water and electrical consumption, so systems must share data rather than operate in silos.

The layers of IoT in a hyperscale campus

Electrical systems

Power flows from utility substations through transformers, switchgear, uninterruptible power supplies, batteries and power distribution units to the racks. Electrical power monitoring systems record voltage, current, power factor, harmonics and breaker status at each stage. Thermal imaging and partial discharge sensors can warn of failing connections. Backup generators add fuel level, run-hour and start-readiness data.

Cooling systems

Cooling plants include chillers, cooling towers or dry coolers, pumps, heat exchangers and, increasingly, coolant distribution units that serve liquid-cooled racks. Typical measurements include supply and return temperatures, flow rates, differential pressures, valve positions, pump vibration and leak detection. In liquid-cooled halls, leak sensing and fluid quality monitoring become critical safety systems.

Water systems

Cooling can consume substantial water, depending on design. Operators track water use through flow meters, conductivity and treatment chemistry sensors, and reservoir levels. Public attention to water is high in Indiana communities hosting data centers, so credible metering supports transparency with local officials and residents.

Environmental and safety systems

Data halls use temperature and humidity sensors at rack inlets, hot and cold aisle containment sensors, smoke detection, early warning air sampling, and fire suppression status. Physical security adds door sensors, cameras, access control and perimeter detection.

Building management and orchestration

A building management system or data center infrastructure management platform pulls these signals together, applies alarms and trends, and in advanced setups adjusts setpoints automatically. Protocols such as BACnet, Modbus and OPC UA are common, often tied together with edge gateways that normalize data for analytics.

SystemTypical measurementsWhy it matters
Utility and switchgearVoltage, current, power quality, breaker stateProtects uptime and supports capacity planning
UPS and batteriesCharge state, temperature, internal resistanceBatteries are a leading cause of power failures if unmonitored
Chillers and pumpsTemperatures, flow, vibration, motor currentPrevents cooling interruptions
Liquid cooling loopsFlow, pressure, coolant quality, leak sensorsProtects expensive accelerator hardware
Rack environmentInlet temperature, humidity, airflowKeeps equipment within warranty and thermal limits
Water metersIntake, blowdown, treatment chemistrySupports efficiency and community reporting
Generators and fuelStart readiness, run hours, fuel levelEnsures backup performs when needed

Efficiency metrics operators watch

Several metrics come up repeatedly in the industry. Power usage effectiveness, or PUE, compares total facility energy with the energy delivered to computing equipment. A value closer to 1.0 means less overhead. Water usage effectiveness, or WUE, compares water consumed with IT energy. Operators also track carbon intensity of electricity and, for AI, the efficiency of useful compute per unit of energy.

These numbers only mean something if the underlying sensors are accurate and well placed. A PUE calculated from poorly calibrated meters can mislead a team into the wrong investments. For Indiana technicians, calibration, commissioning and ongoing verification of meters are real, recurring work.

The construction phase is an IoT story too

Before any server arrives, a campus of this scale involves enormous construction activity. Mortenson and Turner will manage thousands of workers, heavy equipment, material deliveries and strict schedules. Typical construction technology includes equipment telematics, wearable and location-based safety systems, environmental monitoring for dust and noise, concrete maturity sensors, and drone-based progress tracking. Our guide on jobsite intelligence for Indiana contractors covers that world in more detail.

Commissioning is a major data event as well. Before a data hall goes live, teams test power and cooling systems under simulated load, recording how every component responds. Those records become the baseline for the facility's life, which makes data quality during commissioning especially valuable.

What this means for Indiana businesses

Electrical and controls contractors

Demand for qualified electricians, controls technicians and commissioning agents will be high across central Indiana for years. Firms that already hold relevant certifications and safety records have a head start, and those that can staff multiple projects reliably will be sought after.

Sensor, meter and enclosure suppliers

Data center operators buy huge quantities of repeatable components. Suppliers that can document quality, deliver consistent lead times and support standard protocols can find recurring orders, though large owners often buy through national frameworks. Local firms may find the best entry as subcontractors or specialty service providers.

Facilities and maintenance firms

Over a campus lifetime, maintenance is where the long-term revenue lies. Skills in chillers, pumps, switchgear, fire protection and building automation transfer directly. Condition-based maintenance, backed by sensor data, is increasingly expected rather than optional.

Local governments and utilities

Water, roads and electrical infrastructure all need to scale. Smart metering and telemetry help utilities see demand in real time and plan upgrades. Boone County and surrounding communities will want transparent data on usage and impacts, which sensors can supply credibly.

Questions communities are asking

Large data centers draw strong opinions. Residents and advocacy groups raise questions about electricity rates, water use, land use and noise. Utilities and regulators respond with tariffs and agreements that assign grid upgrade costs to large customers, as discussed in our article on how Indiana data centers drive power, cooling and monitoring demand. Claims on each side vary, and readers should read utility filings and regulator decisions directly rather than rely on summaries from interested parties.

From a technology perspective, one observation applies regardless of position: measured data builds trust. Public dashboards for water use, publicly reported energy figures and independent verification of environmental commitments are easier to provide when instrumentation was designed in from the start.

Practical advice for companies seeking this work

  1. Learn the owner's vendor process. Large technology companies have supplier diversity, safety and security requirements. Start registration early.
  2. Connect with the general contractors. Ask Mortenson and Turner about their subcontractor outreach, attend any local supplier events, and respond promptly to bid requests.
  3. Get certifications in order. Safety records, insurance, quality documentation and relevant licenses decide eligibility.
  4. Specialize. Being excellent at one thing, such as battery monitoring commissioning or fluid system leak detection, is better than a general promise.
  5. Invest in people. Use the Boone County Career Collaborative and regional training programs to recruit apprentices.
  6. Plan for the long term. Operations and maintenance work will outlast construction.

Data that should outlive construction

One practical habit separates well-run facilities from troubled ones: they treat the construction and commissioning record as an operating asset. Equipment tags, sensor lists, calibration certificates, test results and as-built network diagrams should be organized in a form operations staff can search years later. If a pump fails in year six, the technician should be able to find its original vibration baseline in minutes. Contractors who deliver clean, structured turnover data, rather than boxes of PDFs, give owners a lasting advantage and make themselves easier to hire again.

What could change the picture

Several uncertainties affect the project and the industry around it. AI demand forecasts have shifted repeatedly, and hardware efficiency gains could change how much power each unit of compute needs. Cooling designs are evolving, with liquid cooling spreading quickly. Regulations on water, emissions and grid interconnection may tighten. Because Meta has not published the campus's technical design in the sources reviewed, outside observers should treat specifics as unknown until the company or regulators disclose them.

The bottom line

A one-gigawatt AI campus is a machine for converting electricity into computation, and sensors are how its operators know whether the machine is healthy. For Indiana, the lasting opportunity is not only the construction boom but the long tail of calibration, monitoring, maintenance and analytics work. Companies that understand power, cooling and water instrumentation, and can show disciplined documentation and safety, are well placed to take part.

Frequently asked questions

It is in Lebanon, Indiana, in Boone County's LEAP district northwest of Indianapolis. Project reports describe an investment of more than $10 billion and a capacity of one gigawatt when complete. It is Meta's second Indiana data center after Jeffersonville.

IoT sensors and meters track power, cooling, water, temperature, humidity, leaks, security and equipment health. A management system combines these signals so operators can detect problems early, plan capacity, and tune efficiency metrics such as PUE and WUE.

Not in the sources reviewed. Public information covers the investment, capacity, contractors, local infrastructure commitments and workforce program. Descriptions of typical data center systems in this article reflect industry practice, not confirmed details of Meta's design.

AI accelerator servers pack much more heat into each rack than traditional servers, and air cooling alone can struggle at those densities. Liquid cooling removes heat more efficiently, but it adds new monitoring needs such as flow, pressure, coolant quality and leak detection.

Register with the owner and general contractors, keep safety, insurance and quality documentation current, specialize in a clear service such as commissioning or maintenance, and build a trained workforce. Long-term operations and maintenance work often outlasts the construction phase.

Common concerns include electricity rates, water consumption, land use, noise and the share of grid costs paid by other customers. Utility filings and regulator decisions are the best primary sources, and measured, publicly shared data can help build trust.

Related articles

Top