AI is growing. But are our buildings growing with it?

We talk about ChatGPT, Copilot, AI agents, automation, cloud, data, GPUs, data centres and cybersecurity. Companies are investing billions in the digital economy. Yet one physical link stands out for receiving remarkably little attention in almost all of these discussions: the building we work in.

If our companies are digitalising at record speed, are our buildings growing along with them digitally? That is not a theoretical technology question. For IT-Label, it is a new real estate question. Where the sector now measures energy performance down to the smallest detail and certifies sustainability and health, the digital layer of a building largely remains undescribed.

This article does not aim to spread fear about AI, internet capacity or so-called data congestion. The core issue runs deeper: AI is accelerating the digitalisation of almost every business process, making organisations increasingly dependent on reliable digital infrastructure. The question is whether we know to what extent buildings are prepared for that.

First the numbers: how fast is AI really growing?

FACT. Eurostat reported that in 2025 about 20 percent of EU companies with at least ten employees used AI technology, compared to roughly 13.5 percent in 2024. That is a substantial step in one year. The full figures can be found in the Eurostat publication on AI adoption among companies.

For the Netherlands, Statistics Netherlands (CBS) publishes its own measurements on the use of artificial intelligence by companies. Anyone wanting to follow the development is best served by looking directly at the CBS figures on ICT use and AI among companies, since these measured values are updated regularly.

The distinction between what has been measured and what is expected matters here. The figures for 2023, 2024 and 2025 are measured values. Where a scenario for the Netherlands towards 2029 mentions, for example, 50 to 70 percent AI use, that is a SCENARIO based on the current adoption trend, and explicitly not an official forecast from CBS, Eurostat or IT-Label. The direction is clear; the exact point on the curve is not.

From AI use to AI dependency

Perhaps the most interesting aspect is not how many companies use an AI tool, but how many business processes become dependent on AI and digital infrastructure. An employee who occasionally consults ChatGPT is something entirely different from an organisation where hundreds of people work daily with business-critical digital applications.

• Microsoft Copilot and ChatGPT Enterprise;
• AI agents and cloud software;
• video conferencing and real-time collaboration;
• data analysis and CRM automation;
• cybersecurity software and connected devices;
• Smart Building applications.

FORECAST. Research from McKinsey and Gartner shows that AI is shifting from experimentation to structural business infrastructure. Gartner publishes a recurring AI spending forecast on this; the current figures and amounts can be found in the Gartner AI spending forecast. The message is consistent: AI spending is growing substantially and is moving towards the core of business operations.

The conclusion that follows is clear. The next phase of AI is not only about more companies using AI. It is mainly about organisations deploying AI more intensively and more critically for their business. That changes the requirements for the digital chain underlying it, a theme we also explore in our article on AI ambition and real estate that can keep up.

The question is no longer whether companies use AI. The question is how dependent their daily work becomes on every link beneath it.

AI is not an internet problem. It is a dependency issue.

It is tempting to say: because of AI, our internet connections will soon become overloaded. That is too simplistic. Much heavy AI compute takes place in cloud environments and data centres, far outside the building. The problem is more subtle.

Because of AI, cloud software, real-time communication, SaaS, IoT and connected devices, organisations are becoming increasingly dependent on a continuously available digital chain: from the workplace to the internal infrastructure, via connectivity to the cloud and the data centre, and from there to the applications, the AI and ultimately the business process itself. Every link counts.

Two sentences summarise this. AI does not need to run locally in the building to make the building more digitally important. And: the more digital the organisation, the more important the reliability of every link in the digital chain becomes. This shifts the attention from computing power to continuity, and that is precisely where the building comes back into view. We explain why one connection is sometimes not enough in our article on network redundancy.

Cushman & Wakefield: AI is changing demand for real estate

AI does not automatically mean that organisations need less space. Research by Cushman & Wakefield into the office market points instead to a shift in the quality of real estate that users ask for. Attention is moving towards technology, energy infrastructure, flexibility, sustainability, user experience and overall building quality.

This partly explains the qualitative scarcity visible in various markets: plenty of supply, but limited supply that meets modern requirements. The current analyses can be found via Cushman & Wakefield insights, which also cover the Dutch office market.

The translation to IT-Label is straightforward. AI does not only change how much space companies need. It can also change which buildings they find attractive. And digital quality is an increasingly weighty criterion in that, as we describe in the IT-Label as a new distinguishing factor in real estate value.

Knight Frank: digital infrastructure as a hidden quality

In its research into workplace strategy and corporate real estate, Knight Frank describes digital infrastructure as an increasingly important quality of modern workplaces. Connectivity, workplace technology, power supply, backup provisions and future-proof infrastructure are becoming part of the baseline, not the luxury.

In the research A defining phase, Knight Frank sees the period towards 2029 as a phase in which AI, technology and data will drive the real estate strategy of organisations. In Workplace Management and in AI in Corporate Real Estate: Adoption Gains Momentum, that same picture is further elaborated: the digital layer of a building becomes decisive for how usable the space is.

This leads to an uncomfortable observation. If digital infrastructure becomes a baseline requirement, why do we still so often treat it as a hidden extra in real estate communication? The topic deserves the same self-evident position as floor area or energy label, something we also raise in IT infrastructure is missing from every rental brochure.

RE-SEARCH: what can a building handle digitally?

Dutch real estate practice asks the same question. In the article Digital Real Estate: what can a building handle digitally?, RE-SEARCH summarises the shift concisely: an office used to have to be accessible, today an office has to be connected.

RE-SEARCH describes how organisations are becoming increasingly dependent on cloud applications, real-time communication, AI tools, data analysis, connected devices, Smart Buildings and IoT. An important insight here is that different users have different digital usage profiles. A law firm does not have the same digital needs as an AI scale-up, and a logistics organisation sets different requirements again than a healthcare institution.

The chain can be read as follows: employees determine the work processes, the work processes determine the digital usage, the digital usage determines the digital need, and that need sets the requirements for the building infrastructure. Whoever knows the user's profile knows which digital foundation is appropriate. This line of thinking aligns with our explanation of what a building can handle digitally.

The digital delivery level as a blind spot

In the article Is your office already AI-proof?, RE-SEARCH identifies a concrete blind spot. Anyone searching for real estate almost immediately receives information about square metres, rent, service costs, energy label, parking spaces, accessibility, installations and delivery level.

Much less often is there an immediate answer to questions such as: is fibre optic present, which providers are available, what capacity is in place, is there redundancy, what internal data cabling has been installed, how is mobile coverage, which facilities are building-related, which Smart Building systems are present, who manages what, and what does the tenant receive digitally upon delivery?

This brings the central question to the surface: we know how many square metres a building has and which energy label it carries, but do we also know what those square metres can handle digitally? For tenants who want to get a grip on this, we compiled what to look for when assessing digital quality.

Are our buildings AI-ready?

Restraint is appropriate here. It would be easy to state that a certain percentage of Dutch buildings is not AI-proof, but such figures do not exist in a reliable, large-scale form. Making them up would undermine the story itself.

For virtually every commercial building we can find out the floor area, construction year, energy label, purpose, rent and location. But is there a national, standardised dataset in which we can read off digital capacity, connectivity, redundancy, data cabling, WiFi quality, mobile coverage, Smart Building functionality, digital demarcation and digital delivery level?

If the answer is no, that is where the real story lies. VIEW. Perhaps the biggest problem is not that we know buildings are digitally insufficiently prepared. Perhaps the problem is that we simply do not yet measure it. The absence of data is itself the finding.

From grid capacity to digital capacity

The real estate sector is now rightly asking pointed questions about power: how much is available, what is the contracted capacity, is expansion possible, is there grid congestion? These questions have become self-evident in leasing and acquisition.

Why do we hardly ever ask the same capacity questions about digital infrastructure? Important to note here: energy and data are technically different subjects, and a one-to-one comparison between grid congestion and internet capacity is incorrect. We use the term data congestion only in a broader sense: a situation in which digital infrastructure, connectivity or capacity does not sufficiently match the user's needs. This is not an official equivalent of electricity grid congestion, as we clarify in will we soon face digital congestion too?

From 2026 to 2029: an illustrative scenario

SCENARIO. The timeline below is an illustrative development scenario, based on the current adoption trend and on research by external parties. It is not a firm prediction that every company will go through these steps at exactly these moments.

1. 2026, AI ASSISTED. AI is widely deployed as a personal assistant and productivity tool.
2. 2027, AI INTEGRATED. AI becomes increasingly interwoven into CRM, finance, marketing, development and other workflows.
3. 2028, AI AGENTIC. AI agents carry out processes and tasks increasingly independently.
4. 2029, AI INFRASTRUCTURE. For many organisations, AI is no longer a separate application but part of the operational infrastructure.

Alongside this timeline, exactly one question fits: does the building develop at the same pace? As AI sinks deeper into operations, the reliability of the underlying digital chain becomes more important, and that chain begins in the building.

What should a building be able to handle digitally by 2029?

Not every building needs to offer maximum capacity or technology. It is about matching the user's needs. However, a digitally future-proof building should be able to provide insight into a number of fixed components.

These components are not abstract technology but concrete information that a user would want to know in advance. More background on the role of the digital chain can be found in our overview on smart building technology.

IT-Label: measure first, then talk about AI-ready

This is precisely where IT-Label comes in. IT-Label does not claim to be able to predict how much AI a company uses. A high classification also does not automatically mean that every user can run every AI application without limit. Its role is more fundamental: making visible what the building delivers digitally.

The IT-Label classification runs from PREMIUM to SHELL: IT1+ PREMIUM, IT1 HIGH PERFORMANCE, IT2 PLUG & PLAY, IT3 READY, IT4 CORE and IT5 SHELL. Each level describes what the building has to offer digitally, not whether a building is good or bad.

This creates a clear division of roles: the user determines the digital need, IT-Label makes visible what the building delivers digitally. Before calling a building AI-ready, we must first know the digital foundation.

The new real estate formula

Traditionally, the market weighed real estate on location, floor area and price. Sustainability was added to that. Moving towards the digital economy, digital quality is being added as well. Importantly: digital quality does not replace the existing criteria, it becomes an additional quality dimension.

Or, in the words that capture this shift: location determines where you work, connectivity increasingly determines how you can work. The location determines where you work, digital quality increasingly determines how you can work.

What does this mean for real estate value?

If companies are becoming increasingly dependent on digital infrastructure, will digital quality eventually also become a factor in real estate value? We ask that question in an exploratory way, not as a claim.

There are plausible hypotheses. Better digital transparency can speed up leasing processes. Digitally well-prepared buildings may become more attractive to certain users. Insufficient digital infrastructure can become a hidden CAPEX item. Digital quality can become part of technical due diligence, and appraisers may weigh it more explicitly in the future. We explore what this means for investment decisions in digital infrastructure, ESG and the value of real estate.

What we do not state without evidence is that a higher IT-Label automatically leads to a higher rent or valuation. What does hold is that what looks like a technical provision today may become a commercial real estate factor tomorrow.

Frequently asked questions

How fast is AI use among companies growing?

According to Eurostat, in 2025 about 20 percent of EU companies with at least ten employees used AI, compared to roughly 13.5 percent in 2024. That is a measured figure, not a forecast.

Does AI mean internet connections will become overloaded?

That is too simply stated. Much heavy AI compute runs in the cloud and in data centres. The point is that organisations are becoming more dependent on the reliability of every link in the digital chain, from workplace to cloud.

Is 70 percent of buildings not AI-ready?

That figure does not exist in reliable form and we will not make one up. What is striking is precisely that there is still hardly any standardised data on the digital quality of buildings. That lack of measurement is the core of this article.

What does IT-Label make transparent?

IT-Label maps out what a building delivers digitally, from connectivity and redundancy to Smart Building facilities, expressed in a classification ranging from IT1+ PREMIUM to IT5 SHELL.

Does every building need the highest classification?

No. It is about matching the user's needs. A floor deliberately delivered as shell and core can make perfect sense, as we explain in a building with a high classification and a floor at IT4.

2029 is closer than we think

This discussion is not only about ChatGPT. It is about a structural change in how organisations function. AI is becoming part of software, processes, communication, analysis, security, customer contact, decision-making and, ultimately, buildings. As a result, reliable digital infrastructure is becoming ever more fundamental.

We measure the growth of AI down to the decimal point. We measure the energy performance of buildings. We certify sustainability and health. Yet one fundamental question remains remarkably often unanswered: what does the user actually receive digitally upon delivery? AI is changing our companies faster than ever. The question is not only whether organisations are ready for AI, but also whether the buildings they work in can grow along with them.

If you want to know what your building can handle digitally, start by making the digital foundation transparent. See how the IT-Label works and discover what a classification means for your real estate. AI is growing. The question is: are our buildings growing with it?