Market sentiment took a battering in 2025, and viability remains a huge challenge. Simon Rawlinson of Arcadis considers where the bright spots are and how construction can take a little more control of its destiny this year
If you look hard enough, you can see the first signs of growth and renewal in 2026. Inflation is falling, bringing the promise of lower borrowing costs. Retail footfall rose over the crucial Christmas period – suggesting improving consumer confidence. Furthermore, Building Safety Regulator reforms appear to be working – potentially bringing forward a wave of much-needed urban housing.
But you do have to look hard. Affordability remains a challenge across too many sectors and occupier demand remains elusive. The latest Arcadis market forecast was entitled “Last roll of the dice” for good reason – highlighting just how dependent consultants and contractors will be on the opening of the investment taps in 2026.
The bull case relies on an almost Panglossian level of optimism. Rather than thinking of 2025 as another year of disappointing growth, you need to think of it as a proof point of the resilience of the UK economy. Similarly, we must anticipate that 2026 will be the year when the impact of 18 months of policy work bears fruit.
Even so, it is a stretch to think that lower inflation and the end to Budget uncertainty will unleash a wave of demand-led investment – particularly as interest rates and taxation continue to weigh down on growth and investment and as affordability limits act as a viability barrier in many markets including housing.
One effect of the high cost of construction is the cannibalisation of future opportunities – reducing the market size for housing, workspace and public infrastructure
Any further slowdown in investment will have wider implications for the market. The Financial Times has, for example, recently highlighted a number of large corporate occupiers – including Accenture and Investec – choosing to extend their existing office leases in response to London’s space shortage, but also to save on fit-out costs.
One of the implications is that staying put potentially reduces the size of the fit-out pipeline. This is the opposite of growth. One effect of the high cost of construction is the cannibalisation of future opportunities – reducing the market size for housing, workspace and public infrastructure.
The situation makes me think of the chemist Ernest Rutherford’s famous saying: “We’ve run out of money; we’ll just have to think.” Rutherford’s original point was prompted by experiments that weren’t delivering the expected results, but the point can be extended to construction too.
Our existing model is not working, so a rethink is needed. Increasingly I think that the fix in 2026 won’t come from market forces alone; it will come from changing product and process too – building on some of the changes already introduced into the planning system in London to improve housing viability.
This pattern is not just restricted to the UK. In the US, for example, the housing market has responded to an even tougher squeeze on borrowing costs by pivoting to building smaller, simpler homes, with the number of completions increasing, even as the value of output remained flat.
As we know, changing designs and specifications in the UK’s highly regulated construction sector is hard to do, so I anticipate that efforts to simplify ways of working present the best opportunity to drive down cost – reversing the trend of increased complexity that has bedevilled projects for many years.
Are there grounds for optimism? I believe there are, and I am not just thinking about robot bricklaying. Digital technologies, supported by AI, are evolving real-world capabilities so fast that we are approaching a tipping point where they match the complexities of the construction site.
One deployment example is the autonomous vehicle joint venture between Mercedes and Nvidia, announced this month, built around a brand-new portfolio of AI tools, models and processors called Alpamayo. Alpamayo is specifically designed to provide the reasoning needed to deal with unexpected “long-tail” events – the kind of events that occur on construction projects all the time.
A similar, construction-specific application has been developed by Field AI, adding site, context and risk awareness to the foundation models that robot “dogs” use to learn about and adapt to the sites where they have been deployed. Buildots also supports the development of robot-ready digital twins.
It is conceivable that, as these capabilities are combined, robotics will finally be taken out of the factory and onto the site – inverting the MMC paradigm that automation is best deployed in a factory setting.
The exciting aspect of this development is that the programming load associated with the deployment of construction automation can be expected to fall as systems become more independently intelligent – addressing RIBA past president Muyiwa Oki’s point that construction’s “sloppy and tacit process” is unsuitable for digital re-engineering.
Another development that gives me confidence is the expansion of the UK’s market for robots as a service (RaaS) into construction, led by machines facilitating demolition and setting out as well as tricky installation tasks associated with cladding and services builder’s work. De-risked investment, maximised utilisation and targeted user support will play a key role in accelerating the deployment of tools like the Hilti Jaibot, used for precise drilling of holes for fixings, which have been available for a number of years.
Innovation is not limited to work on site of course. Agentic AI is another digital evolution that has tremendous potential to simplify or automate many repeat activities and processes on projects. Agentic AI is designed to work out and plan multi-step activities to meet a goal without human input. Critically, agentic AI uses resources such as the live web or a project database to do its work and is not limited to the training data that large language models rely upon.
Increasingly, digital technologies are evolving to a point where they are deployable in the complex, ambiguous world of construction
Configuring agentic AI can be complex, requiring the setting up of API data sources and sets of instructions, but it is increasingly built into industry-standard tools including Autodesk, Oracle and Procore.
Too often construction feels resistant to innovation. Unfortunately, the complexity of projects and our lack of standardisation is a source of cost as well as a barrier to innovation. We need to change in order to add more value.
There is no panacea for our cost, value and productivity challenges but, increasingly, digital technologies are evolving to a point where they are deployable in the complex, ambiguous world of construction.
The fix in 2026 is to recognise that construction is not different and that it can benefit from the productivity gains that other sectors have secured for decades. By reducing cost and by freeing up scarce human resource, we can solve the skills shortage as well as the affordability challenge.
Simon Rawlinson is a partner at Arcadis
No comments yet