AI Is Making Power and Connectivity a Design Requirement

AI demand is changing the building systems conversation. Owners need repeatable data & digital infrastructure standards for power-aware planning, network resilience, OT segmentation, vendor access, and operational continuity before AI-heavy systems arrive.

June 5, 2026 · By Drew Hall

Let's demystify this.

AI demand is not only a data center story. It is becoming a building systems story.

When hyperscale AI campuses start moving dirt, when large data center construction financings close, when Google is reported to be putting billions into data center capacity, and when EV charging providers start moving off the grid, the signal is hard to miss: power, connectivity, and resilience are no longer background utilities.

They are design requirements.

For commercial real estate owners, that does not mean every building becomes a data center. It means the operating assumptions behind buildings are changing. More devices. More data flows. More automation. More vendor access. More tenant expectations. More systems that cannot simply fail quietly.

The owner question is not, "Should we add AI?"

The better question is, "Is the building ready for the systems load AI will create?"

If the answer is unclear, the risk is not abstract. It shows up as operational interruption, vendor lock-in, cyber exposure, stranded CapEx, tenant dissatisfaction, and decision systems built on weak data.

If you don't own your data & digital infrastructure, your vendors do.

That sentence matters more as AI demand rises. AI-heavy systems do not remove the need for strong building design. They raise the cost of weak design.

The Signal Is Coming From Outside Traditional CRE

The strongest signal is not a dashboard demo. It is capital moving toward power-intensive digital capacity.

Connect CRE reported that Oracle and OpenAI have begun construction on a Michigan data center campus, with financing that includes equity from Related Digital funds affiliated with Blackstone. That matters because AI capacity is not being treated as a lab experiment. It is being treated as real estate, finance, energy planning, and operating capacity at scale.

In another market note, Connect CRE flagged Google's reported $15B data center investment. Again, the point for an office, retail, industrial, or mixed-use owner is not that your property must copy a hyperscale design. The point is that the market is repricing the importance of power and connectivity.

At the same time, Connect CRE reported on EV chargers moving off the grid. That is a different use case, but it points to the same design issue. Power availability, continuity, and control are becoming constraints that shape deployment decisions.

Here's what most integrators won't tell you: AI is not one system. It is a load pattern across many systems.

It touches building networks, access control, video, HVAC, elevators, lighting, metering, tenant applications, vendor portals, mobile connectivity, occupancy data, energy systems, and decision workflows. If the base layer is fragmented, AI does not fix it. AI amplifies the fragmentation.

Buildings Were Not Designed for This Level of Systems Density

Most existing commercial buildings were designed around a simpler assumption: systems could operate in their own lanes.

The security contractor had a network. The building automation vendor had another. The access control provider had its own remote access path. The tenant Wi-Fi provider did something separate. The energy platform pulled data through an integration. A carrier or ISP handled connectivity. Somewhere, someone had a spreadsheet with circuit information.

That model can work when expectations are low and systems are mostly isolated.

It breaks down when owners need reliable data, real-time decision support, cyber governance, remote vendor access, and portfolio-wide operating standards.

AI-heavy systems add pressure in five places:

Power planning. More devices and edge systems require better load awareness, UPS design, cabinet planning, device placement, and continuity planning.
Network resilience. AI-enabled operations depend on stable connectivity. Single-path designs, undocumented circuits, and unmanaged switches become operating risk.
OT segmentation. Operational technology is not the same as office IT. Building systems need identity, access, monitoring, and segmentation rules that match operational reality.
Vendor access. More vendors want remote access to more systems. Without owner-controlled access policy, every vendor connection becomes a trust decision made property by property.
Data governance. AI systems need data that is timely, usable, and governed. If data sits inside vendor platforms with unclear ownership, the owner cannot build durable intelligence.

This is why the architectural conversation matters. You cannot bolt reliable intelligence onto an unreliable base layer.

The asset manager should care because this is not a technology preference. It is a risk and value issue.

Weak building systems design can create avoidable OpEx variance. It can slow leasing and tenant improvement delivery. It can complicate lender and buyer diligence. It can create cyber exposure that was never priced into the business plan. It can make every new platform more expensive to install and harder to replace.

Good design does the opposite. It creates repeatability.

Repeatability is where portfolio value starts.

Power-Aware Device Planning Is Now a CRE Discipline

Power used to be treated as a utility question until something failed. That is no longer enough.

Owners need power-aware device planning at the building level. That means knowing what is connected, where it is located, how it is powered, what it depends on, and what happens when a component fails.

This sounds basic. In many buildings, it is not.

A practical owner standard should answer questions like these:

Which systems require backup power?
Which devices are mission-critical for life safety, access, security, tenant operations, or revenue operations?
Which closets or cabinets have UPS coverage?
Which network devices support critical OT systems?
Which vendors depend on which circuits?
Which systems can operate locally if cloud connectivity is interrupted?
Which failures create tenant-facing events?

The point is not to overbuild every property. The point is to design intentionally.

An office tower, outpatient medical building, industrial facility, and mixed-use property will not have the same requirements. But each should have a standard method for mapping power dependency, network dependency, and operational continuity.

That standard should be portable. If every property is mapped differently, the portfolio cannot learn.

This is where many owners get stuck. They buy capable systems, but the building has no common design language underneath them. A camera platform may be modern. A controls platform may be modern. A tenant app may be modern. But the property itself remains a collection of one-off decisions.

AI does not reward one-off decisions. It rewards governed, repeatable patterns.

Connectivity Is an Operating Control Point

Connectivity is often treated as procurement: pick a provider, install circuits, negotiate cost, move on.

That view is too narrow.

Connectivity is an operating control point. It determines whether systems can communicate, whether vendors can be governed, whether data can be collected, whether outages can be isolated, and whether tenants experience the property as reliable.

In the OpticWise model, this starts in Layer 1: managed data & digital infrastructure. That layer includes the network, connectivity, access, segmentation, monitoring, and operating standards that make the property controllable.

Owner control does not mean the owner runs every technical function internally. It means the owner is not trapped by undocumented dependencies, vendor-owned data paths, or property-specific workarounds.

A resilient building design needs more than bandwidth. It needs:

Clear network ownership
Documented circuits and failover paths
Segmented OT networks
Secure remote access rules
Device identity and inventory
Monitoring and alerting
Repeatable naming and data standards
Change control
Vendor accountability

That is not glamorous. It is what keeps the building governable.

The 5S® user experience promise helps translate this into owner outcomes: Seamless Mobility, Security, Stability, Speed, and Service. Those are not cosmetic goals. They are operating requirements. If tenants, operators, vendors, and systems cannot connect reliably and securely, the building's intelligence layer will be weak.

OT Segmentation Is Not Optional

There is a difference between IT and OT.

IT usually supports business systems: email, files, enterprise applications, user devices. OT supports the physical building: controls, access, cameras, sensors, elevators, lighting, metering, and related systems.

They overlap, but they should not be treated as the same environment.

When OT networks are flat, undocumented, or vendor-managed without owner rules, risk compounds quietly. A remote access path installed for convenience can become a long-term exposure. A device added for a small project can sit unmanaged for years. A vendor password can outlive the contract. A system integration can move data without clear lineage.

AI-heavy operations make this more important, not less.

If a future decision engine is going to recommend operational actions, it must sit on top of trusted systems. It must know which data is authoritative. It must know who can access what. It must know which systems are allowed to act and under what permissions.

That is governance. And governance starts below the application layer.

A building owner should not wait until an AI project begins to ask these questions. By then, the project team is often forced to work around the building's weaknesses instead of designing from a sound base.

The better path is to establish the data & digital infrastructure standard first.

The OpticWise Read: Build the Base Layer Before the Intelligence Layer

OpticWise is a data & digital infrastructure company for commercial real estate owners. Our view is simple: the intelligence layer is only as strong as the owner-controlled systems layer beneath it.

Peak Property Performance® organizes this through the PPP 5C™ plan: Clarify, Connect, Collect, Coordinate, and Control.

For this trend, the mapping is direct.

Clarify: Start with a current-state review of the building's data & digital infrastructure. Document circuits, networks, devices, vendors, remote access paths, power dependencies, data flows, and owner control gaps. The goal is not a binder. The goal is decision clarity.

Connect: Establish the owner-controlled network and connectivity layer. Through SIC® and ElasticISP®, owners can create a more repeatable standard for connectivity, segmentation, monitoring, and operating control across properties.

Collect: Bring usable data out of isolated systems. BoT® and Building of Things® are OpticWise's way of framing the connected building environment as an owner-governed operating asset, not a random device collection.

Coordinate: Set the rules for access, data use, vendor permissions, identity, lineage, retention, and workflows. This is where the building stops being a collection of systems and starts becoming a governed operating model.

Control: Place decision systems on top of the governed base layer. Property Brain™ helps turn one property into an owner-controlled intelligence asset. Portfolio Brain™ extends the same pattern across assets so performance knowledge can compound.

This is the two-layer model.

Layer 1 is managed data & digital infrastructure: the owned foundation.

Layer 2 is the owner-controlled intelligence layer: the governed plane where AI, analytics, workflows, and decision systems can operate under owner rules.

The order matters.

If you skip Layer 1, AI becomes another vendor layer sitting on top of weak data, fragile connectivity, unclear permissions, and inconsistent operating standards. That may create dashboards. It does not create control.

A Practical Owner Checklist

Before adding AI-heavy systems to a property, ask a few plain questions.

Do we know which building systems depend on which networks and circuits?

Do we have an owner-controlled method for vendor remote access?

Are OT systems segmented from business IT and tenant networks?

Do we know which devices and systems require backup power?

Can critical systems continue operating through a connectivity interruption?

Do we own the operational data needed for future decision systems?

Can this design be repeated across the next property without starting over?

If the answer is no, the next step is not another point solution. The next step is a data & digital infrastructure review that gives ownership a clear map of risk, dependency, and priority.

That is the quiet work that makes AI useful later.

The market is already telling us where things are headed. Hyperscale AI campuses, reported multibillion-dollar data center investment, large construction financings, and off-grid charging models all point to the same systems reality: power and connectivity are strategic design inputs.

Commercial buildings do not need to become data centers.

But they do need to be designed like digital operating assets.

Find a better way.

Own your data & digital infrastructure. Build for the long game.

References Cited

Your Next Step

Complimentary CRE Data & Digital Review Session

One building. Map who owns what, where data lives, who has permission to act on it, and where operational burden stacks up vs your KPIs.