This audio is automatically generated. Please let us know if you have any comments.
The AI building boom has more money to spend and more problems to solve.
Annual spending on AI infrastructure could reach approx $1.5 trillion by 2031according to Bain, a Boston-based global management consulting firm. Until 2030, the company projects a cumulative $5 trillion to $6.5 trillion in data center spending. To justify those expenses, the AI market would need to approach $6 trillion in annual revenue by 2031, according to the company.
The mind-boggling price tag is “conceivable” given how fast the world’s biggest tech companies have done it. increased data center spendingsaid Peter Hanbury, a partner at the consultancy. However, the question is whether these AI companies will produce enough revenue to sustain construction at this rate.
For contractors, immediate headwinds still revolve around energy constraints and electrical labor, Hanbury said. According to Bain, lead times for key materials are long enough that decisions made today determine capability several years from now.
Increased community opposition is also affecting timelines. Local opposition blocked or delayed at least 75 projects worth $130 billion in the first quarter of 2026, nearly equaling the impacts of all of 2025, according to the firm.
Here, Hanbury talks to Construction Dive about the duration of the AI investment wave, construction constraints and construction feasibility.
Editor’s Note: This interview has been edited for brevity and clarity.
CONSTRUCTION DIVER: How realistic is the projection of $5 trillion to $6.5 trillion in data center spending needed by 2030?
PETER HANBURY: I think it’s realistic based on the spending trajectory we’re already seeing. The biggest hyperscalers, Microsoft, Google, Amazon, Meta, and Oracle, could spend roughly $780 billion in capital expenditures in 2026 alone, nearly five times what they were spending three years earlier.
Peter Hanbury
Courtesy of Bain
Not all of this is data center capital spending, but it shows the scale of capital that is already being deployed. Against this backdrop, a cumulative investment of $5 trillion to $6.5 trillion in data centers by 2030 is certainly conceivable.
The bigger question is whether the economy can sustain this level of investment over time. Our AI paper estimates that annual spending on AI infrastructure could reach about $1.5 trillion by 2031. If capital spending represents roughly 25% of industry revenue, supporting that spending would require an AI market approaching $6 trillion in annual revenue.
So the short-term question is not necessarily whether there is enough capital.
What would need to go right over the next four years to justify this level of investment?
A few things have to happen at the same time.
First, AI must move beyond use cases focused on efficiency and productivity and begin to create entirely new sources of revenue and economic value. We will need substantial growth from new sources of value, such as autonomous systems, physical AI, new consumer experiences, and entirely new AI-enabled products and industries.
Second, physical bottlenecks must be eased. Capital doesn’t help if a project can’t get power, chips, skilled labor or permission to build. Energy, in particular, needs system-level solutions, more generation, faster interconnection, behind-the-meter capacity, storage, and greater coordination between utilities and technology.
Third, the capital model will have to be more creative. Hyperscale balance sheets can carry a huge amount, but at this scale we are likely to see more risk sharing between developments, infrastructure investors, utilities, sovereigns and governments, as well as approaches that include co-financed energy infrastructure.
And finally, the industry will have to be much more disciplined about which projects are built. Not all advertised gigawatts will make sense.
If these things happen, the level of investment is achievable.
From a construction standpoint, what does the industry need to do to deliver this amount of capacity on time?
It will require industry to move from managing individual projects to running programs on an industrial scale. This means integrated energy and site planning, early procurement and more prefabrication, modular designs and coordinated contractor and supplier portfolios.
At this scale, the slowest constrained entry sets the schedule for the entire program.
Which construction trade do you see becoming the biggest bottleneck?
Electrical manpower is likely to be the strongest constraint, particularly high voltage, substation and mission critical electrical talent. Mechanical and piping trades will also come under pressure as liquid cooling scales, along with smaller but highly specialized groups of controls and commissioning talent.
The problem is exacerbated as campuses grow because these projects require large numbers of skilled workers at once. This will drive more work towards off-site manufacturing, prefabricated assemblies, regional labor strategies and long-term partnerships with key trades.
What should contractors consider when evaluating their data center pipelines?
Contractors should increasingly underwrite the pipeline.
I would ask four questions: is the power real? Is the commitment to the customer and the financing real? Is the building permit real? And is the design stable enough to build?
Permission increasingly means more than permission. It includes community support and social license on energy use, water, noise, emissions and other local impacts.
Power is becoming the door element. A data center can be built in a few years, but adding large network capacity can take four years or more. On larger campuses, this means that electrical infrastructure is now firmly in the critical path of data center projects.
Contractors are increasingly working around substations, transmission, interconnection and, in some cases, on-site generation and storage as part of the same delivery program. A gigawatt-scale data center is more like a power project with a very large computing load.
Are there other trends contractors should be paying attention to?
One of the biggest changes is the chip-to-network code design. Historically, the facility and IT stack could be designed with a fair amount of separation. The building didn’t have to change drastically with each new generation of chips. AI is collapsing these boundaries.
Changes to graphics processing units and custom silicon frame density. Rack density drives networking and cooling. Cooling and compute density change the electrical architecture. And all of these choices affect the building and ultimately the power source.
For contractors, the implication is that the server roadmap increasingly helps design the building. And that silicon roadmap is getting more varied and faster as Nvidia continues to accelerate the cadence of its platform while hyperscalers simultaneously drive custom silicon.
Players who can coordinate compute, power, cooldown, and build will have an advantage over those who optimize each layer independently.
