Articles

The Speed Trap

AI-Fueled Data Center Growth Meets Building Lifecycle Reality

For the first time in U.S. history, investment in new data center construction has surpassed investment in new office buildings. At $45.1 billion in active projects - a 228% surge since the launch of ChatGPT in late 2022 - the data center sector is reshaping where capital flows within the built environment. Office construction, by contrast, has declined 38% over the same period to $43.5 billion. This is not a cyclical blip. It is a structural reallocation of the economy.

REIT markets have priced in the shift decisively. Data Center REITs delivered a +24.33% year-to-date return through February 2026, trading at a 26.9x price-to-FFO multiple - the least-shorted property type in the sector. Office REITs, at -8.17% YTD and a 7.3x multiple, sit at the opposite end of the spectrum. The 3.7x valuation gap between these two asset classes tells a story of irreversible market repricing.

But here is the question the market has not fully priced: can the industry build this fast without creating a lifecycle deficit that compounds for decades?

The Numbers at a Glance

What Is Driving the Surge

The demand thesis is not speculative - it is structural. At Nvidia’s GTC 2026 keynote, Jensen Huang declared the year an “inflection point for inference,” unveiling hardware that integrates Groq technology with traditional GPU architecture for faster, more efficient inference at scale. The implication is clear: the compute infrastructure required for AI is not a one-time buildout. It is a continuously escalating demand curve.

Alibaba’s analysis puts the scale in perspective: autonomous AI agents consume tokens at 40–60x the rate of traditional chatbots. As enterprises move from pilot programs to production-scale agent deployments, the demand floor rises accordingly. Critical power to support global data center operations is expected to nearly double between 2023 and 2026, reaching approximately 96 gigawatts - with AI operations alone consuming over 40% of that power.

Regional grid operators are already sounding alarms. PJM projects a 6-gigawatt shortfall by 2027 - equivalent to the output of six large nuclear plants. Approximately 70% of the existing U.S. power grid is approaching end-of-life, creating a collision between surging demand and aging infrastructure.

The Lifecycle Gap: What Speed Leaves Behind

Building fast is not the same as building well. Modular construction techniques have compressed data center delivery timelines by 30–50%, bringing some projects to completion in 16–20 months. That is an extraordinary engineering achievement. But speed optimizes for one phase of the building lifecycle - delivery - while potentially shortchanging the phases that follow: operations, maintenance, adaptation, and decommissioning.

Think of it like a Formula 1 pit stop. The car gets back on the track in record time, but if the mechanics skip a lug nut, the consequences emerge at 200 miles per hour. In data centers, the “lug nuts” are the lifecycle considerations that do not appear on the construction schedule but determine whether a facility remains viable, efficient, and safe over its intended service life.

Likely Lifecycle Omissions in Rapid Development Cycles

Commissioning Depth

Compressed schedules often truncate commissioning - the systematic process of verifying that every system performs as designed. In a 50+ megawatt facility running at rack densities of 30–50 kW, the margin for error in power distribution, cooling redundancy, and failover sequencing is razor-thin. Incomplete commissioning does not cause immediate failure; it creates latent risk that surfaces during the first major stress event.

Maintainability-by-Design

When delivery speed is the primary metric, design decisions that optimize long-term maintainability - adequate service corridors, accessible mechanical systems, modular component replacement paths - can be deprioritized. Facilities designed for construction speed may prove expensive to operate and difficult to service, particularly as power densities continue to climb.

Adaptive Capacity

Customer specifications shift mid-deployment as AI hardware evolves faster than buildings can rise. NVIDIA’s shift to an annual product cadence - Hopper (2022), Blackwell (2024), Rubin (2026) - means the computing hardware inside a facility may turn over every 1–3 years. A building designed for today’s thermal profile may be inadequate for next year’s GPU generation. Facilities that lack adaptive capacity become stranded assets in all but name.

Decommissioning and End-of-Life Planning

With GPU functional lifespans as short as 1–3 years and hardware refresh cycles accelerating, data centers will generate unprecedented volumes of electronic waste. Yet 12% of data centers engage in no e-waste recycling, and 43% lack an environmental policy for e-waste management. European Commission regulations now require detailed sustainability reporting for data centers with a capacity of over 500 kW, with similar frameworks emerging globally. Operators who defer decommissioning planning to “later” will find that “later” arrives on a very short timeline.

Workforce Development

Single data center campuses now require 4,000 construction workers, up from 750 a few years ago. The operational side faces parallel pressure: facilities running liquid cooling at 50 kW per rack demand specialized expertise that the traditional FM workforce was not trained to deliver. The pipeline of qualified data center operations professionals has not kept pace with the build rate.

Key Takeaways

For Commercial Real Estate

• The capital migration from office to data center is structural, not cyclical. The 228% construction surge versus a 38% office decline represents a permanent reallocation of investment in the built environment.
• Valuation premiums reward operators who deliver capacity, but lifecycle risk is not yet priced. Investors should evaluate not just delivery speed but long-term operational resilience and adaptive capacity.
• Stranded-asset risk is real. A facility designed for 2026 thermal loads may be functionally obsolete for 2028 hardware without significant retrofit investment.

For Energy Capacity and Distribution

• Data center power demand is projected to reach 75.8 GW in the U.S. alone by 2026. Over 60% of this power still comes from fossil fuels, despite renewable pledges.
• Grid infrastructure is simultaneously aging and overloaded - 70% of the U.S. grid is approaching end-of-life, and regional operators project multi-gigawatt shortfalls.
• Operators are evolving from passive energy consumers to active grid stakeholders, co-investing in infrastructure upgrades, deploying on-site generation, and enabling demand-response flexibility.

For Operations and Maintenance

• The shift from 5–8 kW to 30–50 kW rack densities transforms every aspect of facility operations - cooling strategies, power distribution, redundancy engineering, and maintenance protocols.
• Predictive maintenance powered by AI and condition-based monitoring is replacing traditional interval-based approaches, but implementation requires investment in sensors, data infrastructure, and skilled personnel.
• The FM profession faces a pivotal moment: declining office portfolios are compressing the traditional market, while data center operations are creating demand for specialized expertise that commands premium compensation.

The Bottom Line

The data center buildout underway is historic in scale and speed. The market signals are unambiguous: capital is flowing, valuations are rising, and demand has a structural floor that continues to rise. But the building lifecycle does not negotiate with construction schedules. A facility that is commissioned incompletely, designed without adaptive capacity, and operated without a decommissioning strategy is not a long-term asset - it is a depreciating liability with excellent curb appeal.

The industry has an opportunity to get this right: to match the speed of delivery with the rigor of lifecycle planning. The organizations that do - integrating commissioning depth, maintainability-by-design, adaptive infrastructure, and end-of-life stewardship into every project - will own the next decade. Those who treat lifecycle management as an afterthought will discover, at great expense, that the fastest to build is not always the best.

What’s on the Horizon - Lifecycle Interoperability In 2026–2027

When standards-defined data integration and interoperability spans the building lifecycle (2026–2027)

Commercial real estate (CRE) has spent the last decade investing in digital representations of buildings—models, asset registers, automation systems, analytics platforms—yet too many portfolios still operate like a relay race where each runner drops the baton. Information created during design and construction often arrives in operations incomplete, inconsistent, or trapped inside a tool’s preferred format. The operational team then recreates what already existed, and the cycle repeats at every renovation, tenant change, or capital project.

The next two years show credible signs of a shift. Several major standards and frameworks are evolving to enable information to be specified, validated, exchanged, and reused with less manual translation. The emerging picture is not a single “magic standard” but an interoperable system—more like a modern shipping network where containers, customs rules, tracking codes, and port infrastructure all align to move goods predictably at scale.

The central thesis is straightforward:

Standards-defined lifecycle interoperability becomes real when governance, exchange schemas, machine-checkable requirements, operational handover structures, service-life planning, semantics, and CRE business data models align—so building information can travel across the lifecycle without losing its meaning.

This article synthesizes the current status and 2026–2027 horizon for:

·       ISO 19650 (information management processes)

·       ISO 15686 (service-life planning)

·       ASHRAE 223P (semantic data model for analytics and automation)

·       buildingSMART / ISO Industry Foundation Classes (IFC) (open schema for built assets)

·       buildingSMART Information Delivery Specification (IDS) (machine-readable information requirements and validation)

·       COBie (construction-to-operations building information exchange)

·       OSCRE (real estate industry data model)

The framing is consistent with the Building Lifecycle Management Initiative (BLMI): lifecycle value scales when information stays coherent from planning and delivery into operations, renewal, and portfolio decision-making.

Executive outlook: why 2026–2027 matters

CRE is approaching a convergence window where:

·       ISO 19650 governance language is being modernized so lifecycle information management is less segmented between “project” and “operations.”

·       IFC 4.3 is stable and widely referenced as the openBIM exchange baseline, while IFC 5 development focuses on long-term architecture improvements.

·       IDS is already finalized as a buildingSMART standard and is becoming the practical method for turning requirements into automated quality checks.

·       COBie is being delivered in more modern formats (including JSON and IFC-aligned templates) that support automation beyond spreadsheets.

·       ISO 15686 service-life planning is being refreshed to better support lifecycle strategies.

·       ASHRAE 223P advances semantic interoperability for operational analytics and automation.

·       OSCRE connects building and operational truth to portfolio/business truth (leases, reporting, ESG, capital planning).

The result is a clearer pathway to a standards-defined “digital thread” that CRE organizations can scale across portfolios.

The lifecycle interoperability stack: how the standards fit together

A helpful way to visualize interdependence is to map each standard to a role in a lifecycle information supply chain:

·       ISO 19650: the governance and process rules (who requests what, when, and how it is managed)

·       IFC (ISO 16739-1): the open schema “container” for built asset information

·       IDS: the machine-readable checklist that validates the required information is actually present

·       COBie: the operational handover manifest—what operators need to run and maintain the asset

·       ISO 15686: the service-life compass for maintenance and replacement planning

·       ASHRAE 223P: the semantic layer that makes operational data consistently meaningful

·       OSCRE: the business-domain model that links asset and operations data to CRE workflows and reporting

If any one layer is missing, organizations fall back to manual interpretation, custom mapping, and costly rework.

ISO 19650: lifecycle information governance that turns “data” into a deliverable

Current status ISO 19650 is widely adopted as the international foundation for managing information across the lifecycle of built assets, including defining information requirements, exchange points, roles, responsibilities, and common data environment expectations.

2026–2027 horizon ISO 19650 Parts 1–3 are in revision, including development-stage drafts (DIS/CD) that reflect ongoing modernization.

Why it matters: For CRE, ISO 19650 is the difference between “hoping” information arrives and making information delivery enforceable. Without governance, a portfolio’s information behaves like an unlabeled set of moving boxes: items may exist somewhere, but no one can reliably find the right box, confirm what’s inside, or trust it’s complete.

With the ISO 19650 discipline, owners and operators can define Owner/Asset/Exchange Information Requirements in a way that procurement teams can contract, project teams can deliver, and acceptance teams can audit. In practical terms, it supports fewer disputes at handover, clearer accountability, and more consistent lifecycle continuity—especially when combined with IDS for machine validation.

IFC (ISO 16739-1): the open schema that keeps building data portable

Current status: IFC is the principal open, vendor-neutral schema for representing built assets and infrastructure works over their lifecycle. buildingSMART identifies IFC 4.3.2.0 (IFC 4.3) as the latest official release, also published as ISO 16739-1.

2026–2027 horizon IFC 4.3 remains the practical baseline for exchange, while IFC 5 development focuses on longer-term architecture improvements. For CRE stakeholders, the near-term story is stability and improved interoperability through consistent implementation and certification—not constant schema churn.

Why it matters: CRE portfolios rarely live within a single software ecosystem. A building’s digital life touches architects, engineers, contractors, commissioning agents, FM teams, CAFM/CMMS/EAM platforms, analytics vendors, and capital planning tools—often across decades. IFC functions like a standardized shipping container: it does not guarantee that the contents are perfect, but it makes it far more likely that the contents can be moved and reused without being rebuilt from scratch.

When IFC is used consistently, owners gain leverage. Data exchange becomes less dependent on a vendor relationship and more dependent on compliance with a shared, open structure.

IDS: turning information requirements into automated quality control

Current status Information Delivery Specification (IDS) is a buildingSMART standard that defines information requirements in a computer-interpretable form and supports automated compliance checking of IFC models.

2026–2027 horizon: IDS is expected to mature through enhanced tooling, broader procurement adoption, and iterative refinements as the ecosystem evolves.

Why it matters: IDS is the missing “inspection mechanism” that CRE has long needed. Many owner requirements exist as narrative documents and spreadsheets—useful for humans, but inconsistent for software and hard to enforce. IDS converts the requirements into a form that can be checked, much like a barcode scanner validates a shipment.

That changes the economics of handover and model quality. Instead of discovering missing asset tags, classifications, or properties at the end—when fixes are expensive—IDS supports earlier detection and correction. It also reduces the burden on FM teams, who otherwise inherit the cost of incomplete information.

COBie: the handover manifest—modernizing how operations get what they need

Current status: COBie (Construction to Building Operations information exchange) is a structured dataset intended to support O&M handover: asset registers, spaces, systems, warranties, maintenance information, and more. It is commonly delivered in spreadsheet form, but NBIMS-US v4 resources include modern templates such as JSON schema and IFC schema representations.

2026–2027 horizon: The practical direction is modernization: COBie outcomes delivered in more automation-friendly formats and better alignment with model-based exchange.

Why it matters: Operations teams do not manage “models”—they manage assets. COBie is valuable because it focuses on the minimum viable dataset needed to run the building on day one. The relatable reality for CRE is that handover is often a messy kitchen move-in: the appliances arrive, but the manuals, warranty cards, and service schedules are scattered across boxes.

Modern COBie delivery formats support a shift from “spreadsheet as final product” to “structured data as a reusable asset.” When integrated with IFC and validated via IDS, COBie becomes less of a painful one-time deliverable and more of a repeatable lifecycle baseline.

ISO 15686: service-life planning as a portfolio maintenance compass

Current status ISO 15686 provides methods and structure for service-life planning, supporting long-term performance, maintenance, replacement, and lifecycle cost strategies.

2026–2027 horizon The series is undergoing modernization, including development of ISO/DIS 15686-1.2.

Why it matters: CRE value is won or lost over decades, not at substantial completion. Service-life planning is what turns asset data into a capital strategy. Without it, portfolios behave as cars run without a maintenance schedule: things still work—until they fail at the least convenient time.

ISO 15686 becomes far more powerful when paired with interoperability for lifecycle data. If service-life assumptions can be captured at handover (COBie/IFC), validated as complete (IDS), governed over time (ISO 19650), and linked to portfolio planning systems (OSCRE), then capital planning shifts from reactive to predictive.

ASHRAE 223P: semantics that let operational analytics scale across buildings

Current status: ASHRAE Proposed Standard 223P aims to define interoperable, machine-readable semantic models representing building system information for analytics, automation, and control.

2026–2027 horizon: The work is positioned toward publication in the 223-202x timeframe and is advancing through formal standards development processes.

Why it matters: A portfolio can have excellent BIM and handover data and still struggle to scale analytics—because operational data often lacks consistent meaning. One building’s “SAT” might be another building’s “SupplyTemp,” with different units, relationships, and contexts. That forces organizations to rebuild integrations, building by building.

ASHRAE 223P matters because it targets the “meaning layer.” It helps software understand not only that a point exists, but what it represents and how it relates to equipment and system topology. The relatable outcome is that analytics become more like a reusable app store—deployable across many buildings—rather than like custom cabinetry built separately for every project.

OSCRE: connecting building and operations data to CRE business outcomes

Current status: OSCRE provides a comprehensive Industry Data Model (IDM) to support standardized data exchange across real estate use cases.

2026–2027 horizon OSCRE’s continued development and adoption trajectories are increasingly relevant as CRE teams align operational building data with portfolio reporting, ESG metrics, and enterprise systems.

Why it matters: Interoperability fails if it stops at the mechanical room. CRE leaders ultimately need to build facts to connect to business decisions: lease obligations, tenant experience, operating expenses, procurement, capital planning, risk, and sustainability reporting.

OSCRE serves as the “accounting chart of accounts” for real estate data—providing structured definitions so the organization is not constantly translating between operational and portfolio systems. When OSCRE is used alongside building data standards, the portfolio can stop treating building information as a technical side project and start treating it as a core operating asset.

Where the gears mesh: how lifecycle interoperability becomes achievable

Interoperability becomes practical when these standards are used together:

·       ISO 19650 defines the lifecycle information management process.

·       Owners express requirements that can be formalized as IDS.

·       Project teams deliver IFC-based exchanges.

·       IDS validates the delivery automatically.

·       COBie captures operations-ready baseline data.

·       ISO 15686 informs service-life planning and long-term strategy.

·       ASHRAE 223P makes operational telemetry consistently meaningful.

·       OSCRE connects technical truths to portfolio truths.

In an analogy: ISO 19650 is the “rules of shipping,” IFC is the standardized container, IDS is the scanner that confirms the shipment meets contract, COBie is the packing slip operators rely on, ISO 15686 is the long-term maintenance schedule, 223P is the shared language for sensor meaning, and OSCRE is the enterprise ledger where real estate decisions get made.

Click the image to view the interactive infographic

When CRE will feel it: four visible shifts in 2026–2027

From policy to proof: enforceable information requirements

As ISO 19650 revisions progress and IDS adoption expands, owners can increasingly specify requirements in a way that can be automatically verified. This is the transition from “requirements as interpretation” to “requirements as measurable acceptance.”

From handover scramble to operations-ready baselines

With modern COBie delivery formats and better alignment with model-based data, handover becomes less of a frantic end-of-project event and more of a repeatable data product that feeds CMMS/EAM systems cleanly.

From points to performance: semantic operations at scale

As semantic approaches mature, analytics and automation deployments can become repeatable across buildings. The integration burden drops, and portfolio-wide optimization becomes more feasible without bespoke engineering in every building.

From building data to portfolio decisions: aligning technical truth with business truth

With OSCRE and related enterprise models, the industry gains a clearer path to connecting asset/ops data to the systems that govern value: finance, leasing, ESG reporting, and portfolio planning.

Where CRE professionals should build awareness and knowledge

·       Owner-side information requirements (OIR/AIR/EIR literacy): The skill to define what data is needed, not only what documents are needed.

·       IFC + IDS paired capability: Exchange without validation is fragile; validation without exchange is isolated. CRE teams should understand how they work together in procurement and acceptance.

·       Handover modernization: Move beyond “spreadsheet compliance” toward structured, system-ready datasets.

·       Service-life planning as a data discipline: Connect ISO 15686 thinking to digital asset records so capital planning is data-driven and defensible.

·       Operational semantics readiness: Prepare for semantic standards by building internal awareness of point meaning, topology, and system relationships.

·       Enterprise integration: Align building and operational data initiatives with OSCRE-style business models to reduce translation between technical and financial realities.

Closing: the horizon is convergence—governance + schema + validation + semantics + business context

The building industry is moving toward a practical tipping point. Governance (ISO 19650), open schema (IFC), computable requirements (IDS), operational handover structure (COBie), service-life planning (ISO 15686), operational semantics (ASHRAE 223P), and CRE business-domain modeling (OSCRE) are increasingly complementary.

Together they point to a future where lifecycle interoperability is less about custom integrations and more about predictable, standards-defined connections—like the internet, where shared protocols make exchange reliable even when systems differ.

For CRE professionals, the horizon is not merely new standards documents. It is a path toward buildings and portfolios that preserve their digital memory—so every lifecycle decision starts with trusted information rather than costly rediscovery.

Sources