Artificial intelligence infrastructure investment in the United States is projected to reach USD 1.4 trillion in 2026, a scale of capital deployment that a former senior White House technology official compared to the construction of America's railway network in the 19th century. The projection was shared at a G20 session on emerging technologies held in Chapel Hill on 1 September.

The Scale of the Build-Out

Former White House AI adviser David Sacks told G20 representatives that approximately USD 800 billion is being invested in AI-related capital expenditure in the US in the current year alone. That figure, he said, is expected to nearly double to USD 1.4 trillion in 2026, with estimates being revised upward repeatedly as companies race to expand data centres and computing capacity.

'We really haven't seen a build-out of this kind of infrastructure,' Sacks said. 'You have to go all the way back to the railroads in the 1800s to see something of equal scale.' He added that the investment already surpasses the fibre-optic network expansion that underpinned the internet era. 'It's a new industrial revolution,' he said.

Economic and Industrial Implications

Sacks argued that, if managed correctly, the AI infrastructure surge could help the US reindustrialise, modernise its electricity grid, and restore manufacturing capacity in advanced sectors. He noted that the opportunity extends beyond American borders — other nations that embrace the technology, he said, stand to benefit as well.

The G20 session, which brought together ministers to examine how emerging technologies can drive economic growth, also considered the policy conditions needed to attract investment and accelerate the commercial adoption of new technologies.

The Data Centre Controversy

The rapid rollout of data centres has, however, become politically contentious in the United States. Communities have raised concerns about rising electricity prices and the strain on local infrastructure. Sacks pushed back on that framing, arguing that data centres, if developed responsibly, could actually lower power costs.

'Data centres, if done right, they bring electricity costs down, not up, because the AI companies will generate net new power generation,' he said. He explained that companies could build power facilities behind the meter to meet their own needs and return surplus electricity to the grid — effectively subsidising broader grid upgrades that have long been delayed.

Local Benefits: Jobs and Tax Relief

Sacks pointed to tangible community benefits already materialising. AI facilities are generating hundreds of thousands of construction and skilled-trade jobs, he said. Counties hosting data centres are also receiving tax revenues that reduce the property-tax burden on residents.

He cited Loudoun County in Virginia — one of the largest data-centre markets in the country — where residents are reportedly paying an average of USD 6,000 less in annual property taxes as a result of revenue from AI companies.

Sacks was clear that federal authorities have not sought to override local decision-making on data centre siting. 'It is fundamentally a local decision, and we respect that,' he said, while encouraging communities to 'lean into' the economic opportunity.

Regulatory Stance

On the question of governance, Sacks urged governments to nurture the AI boom rather than constrain it with premature regulation. 'The AI boom is already here,' he said. 'And I think this is a very positive thing.' His remarks signal a continued preference in current US policy circles for a light-touch regulatory approach, even as other major economies — including the European Union — have moved toward binding AI frameworks.

With investment estimates being revised upward quarter by quarter, the trajectory of US AI infrastructure spending will be closely watched by policymakers and markets alike heading into 2026.