Nokia CEO says supply limits could slow AI data-center buildout

Nokia data center buildout demand would support a faster pace of construction if sector-wide supply constraints were removed, according to CEO Justin Hotard. In remarks reported by CNBC on Oct. 5, Hotard said the point reflects customer appetite for AI infrastructure, while stopping short of a forecast that construction will actually double.

Hotard told CNBC’s The Tech Download that customers would probably build facilities at twice the pace if they were not constrained. “If we could build 2x faster, our customers could build 2x faster, they probably would,” he said.

CNBC identified shortages of key memory chips and energy among the constraints facing the AI data-center sector. Hotard cited those limits in arguing that the current buildout remains at an early stage. That is his assessment of demand, rather than a measured industry-wide construction forecast or evidence that the constraints will ease.

What is limiting AI data-center construction, according to Nokia’s CEO?

Hotard pointed to supplies of key memory chips and energy. CNBC reported that these are among the bottlenecks facing the sector. The available reporting does not rank them against other potential limits or establish a timetable for projects to move faster.

For investors following Nokia, the company’s connection to the buildout is its networking equipment. CNBC reported that Nokia sells technology used to link chip racks inside data centers and to connect data centers in different locations. That places the company in the infrastructure chain around AI computing rather than in the market for AI models themselves.

Hotard also said demand does not depend solely on frontier AI labs such as Anthropic and OpenAI releasing new models. In his view, companies could make substantial progress by deploying technology already available even if no new frontier model arrived over the next three years, CNBC reported.

Is AI data-center spending expected to hold up?

That remains unresolved. CNBC reported an ongoing debate over whether the industry is adding data-center capacity too quickly and whether demand for AI services can support the investment over time.

As context, CNBC cited research presented at the Brookings Papers on Economic Activity estimating $10.3 trillion of investment in AI infrastructure, including data centers, chips and related equipment, from 2025 through 2032. The estimate is research cited by CNBC, not a result established by Hotard’s comments.

His remarks offer a view from a company that sells networking gear into the sector: constrained customers would build more if they could. They do not settle the wider question of whether long-term demand will justify the scale of planned AI infrastructure spending.

This story draws on original reporting from CNBC.