For most of the past decade, the case for smart buildings has been made in the future tense. Sensors, analytics and automated controls promised to cut energy use, keep occupants comfortable and help the built environment adapt to a warming climate. The premise was sound, and much of it still is. What has changed is that the climate pressure the technology was meant to answer is no longer a projection but a measured condition. The UK has now recorded its hottest May day, 35.1°C at Kew Gardens, two degrees above a record that had stood since 1922, and the five warmest UK summers in a series running back to 1884 have all fallen since 2000. The question for owners and investors is no longer whether buildings will face heat on this scale. It is whether the technology being sold to meet that challenge performs as promised.
The market’s confidence on that point runs ahead of the evidence, starting with its own numbers, which show a rather mixed picture, Grand View Research and Global Market Insights put the global market at somewhere between $103 billion and $142 billion today, and their forecasts for the early 2030s diverge far more sharply. Mordor Intelligence projects around $310 billion by 2030; Grand View, more than $550 billion by 2033; Global Market Insights, over $825 billion by 2034; MarketsandMarkets, barely $200 billion by 2032. The implied growth rates run from 9% a year to more than 30%. A four-fold spread in forecasts is not the mark of a mature category. It reflects a category that remains loosely drawn, stretching from a single connected thermostat to a fully AI-managed corporate campus.
Wide adoption, shallow integration
Retrofit projects still account for around 62% of smart building work, according to Mordor Intelligence, and new construction, though growing faster, is doing so from a small base. The supply side appears to be concentrated as well. Honeywell, Siemens, Johnson Controls and Schneider Electric collectively delivered more than 40% of multi-site smart building deployments completed in 2024. and many large multi-site deployments now run on their integrated platforms.
Investment across proptech more broadly tells a similar story of consolidation. Venture funding peaked in 2021, fell steeply through 2022 and 2023, and has recovered only partially since, with fewer, larger rounds going to more established companies rather than being spread across newcomers. None of this describes a technology in crisis. It describes one that works well in the right conditions and has struggled to scale beyond them.
The test of a hot summer
Heat is now one the sector’s strongest selling points, and perhaps the easiest of its claims to test. The proposition is straightforward: intelligent controls, sensors and automated ventilation will keep buildings habitable as summers grow longer and hotter. The performance data, however, tells a more mixed story.
A 2026 review of intelligent HVAC control in IEEE Access found that learning-based systems reduced heating energy by around 24% against conventional rule-based controls but cut cooling energy by less than 9%. The control layer is far better at trimming winter demand than at carrying a building through a summer heatwave, and the reason is physical rather than technical. In extreme heat the binding constraint is the building’s physical design and the temperature of the air outside, not the sophistication of the software. During the July 2022 heatwave, when England first passed 40°C, night-time temperatures did not fall below 25.8°C anywhere in the country. Automated night-purge ventilation, one of the strategies these systems are marketed on, depends on cool night air to draw in. When there is none, there is nothing for the system to work with.
The environmental case runs into the same limit from the opposite direction. Relying on mechanical cooling to hold a poorly designed building at a comfortable temperature raises electricity demand and emissions and adds to the urban heat that created the problem. UNSW Sydney’s Professor Mat Santamouris AM, estimates that passive design could cut cooling demand by up to 80% in hot climates, and argues that we cannot air-condition our way out of a warming climate. A control system can make a large cooling load run more efficiently. It cannot remove the need for the load in the first place. The intelligence sits on top of the physics. It does not substitute for it.
The problem of proof
Smart buildings are sold on projected savings, however, verified performance, measured once a building is occupied and running, is rarely published. The gap between the two is important. Those behind the UK Net Zero Carbon Buildings Standard point out that a building’s actual energy use is often close to twice its design prediction. Independent assessment of the technology itself tells a similar story. In a 2026 study of AI in commercial buildings, Memoori found that vendors commonly claim energy savings of 20% to 50%, while portfolio-scale independent evaluations verify only 3% to 15%. The largest of those evaluations, run by NYSERDA across 654 sites, found claimed savings running to roughly twice what could be verified. And the verified saving tracks how much the system actually does, not how advanced it is. Passive dashboards deliver around 2% to 3%, fault detection around 9%, and autonomous optimisation around 12% to 13%. The savings are real, but they are consistently smaller, and less certain, than the headline figures suggest.
Verification of that kind is still the exception. The UK Net Zero Carbon Buildings Standard, launched in March 2026, is a rare instance of the opposite, requiring twelve months of measured in-use data before a net zero claim can stand. As Memoori has also argued, the main barrier to a return on smart building investment is not the technology. It is the absence of a governance and risk model to sit around it.
The overlooked liability
Connectivity brings its own exposure, and owners have been slow to account for it. Standard cyber insurance often does not extend to building automation systems, and insurers are now declining cover or refusing renewal where owners cannot demonstrate basic controls such as the separation of operational and IT networks. The consequences of an attack, meanwhile, fall on the asset owner rather than the vendor, a division of responsibility set out in standards such as ISA/IEC 62443. This is not a remote risk. More than a third of smart buildings have already been attacked. Legacy building management systems still run on protocols such as BACnet, Modbus and LonWorks, designed decades ago for isolated industrial networks rather than anything facing the public internet, and adding a layer of internet-connected devices to that infrastructure creates vulnerabilities that were not there before. It is a real and rising cost, and one that few investment cases for smart buildings have yet to account for.
What has moved the market
Regulation appears to be one of the strongest drivers of adoption, and among the most consistent. The UK’s Minimum Energy Efficiency Standards, the Corporate Sustainability Reporting Directive, the GRESB benchmark and the Net Zero Carbon Buildings Standard have each given owners a concrete reason to act, attaching a measurable cost to standing still.
Heat is beginning to follow the same route. Part O of the Building Regulations, the overheating standard for new homes in England introduced in 2022, changed how they are designed, requiring passive measures first, shading, ventilation and reduced solar gain, before any mechanical cooling. That framework is still evolving: the government opened a full review of Part O in 2026, and developers in London already work under two overheating regimes that do not fully align. The clearest gains in adoption have come where regulation is both logical and unified.
Beneath all of it sits a workforce question that receives far less attention than the technology. Facility managers are increasingly expected to run AI-enabled platforms, interpret building analytics, manage cyber risk and meet ESG and green lease obligations, often without the training to do any of it well. What is now asked of the technology runs ahead of what it can deliver. JLL’s 2026 Future of Work survey found that 78% of business leaders expect AI to reshape their real estate strategy, while Morgan Stanley estimates that AI could automate around 37% of tasks across real estate. That gap, between what is expected of these systems and what they can actually do, is the real question facing the sector.
Digital systems can undoubtedly improve how buildings operate, reduce waste and provide insights that were previously impossible. But they do not override poor design, compensate for inadequate maintenance or repeal the laws of physics. As the climate becomes more extreme and regulators increasingly demand real-world evidence, the future of the smart building will not be decided by how intelligent it claims to be, but by whether it delivers when conditions are at their most demanding.
