Nokia CEO Says AI Infrastructure Demand Could Double if Supply Chain Bottlenecks Were Removed

AI Infrastructure Demand Remains Severely Constrained

Nokia CEO Justin Hotard argues that AI infrastructure demand is nowhere near its ceiling, with supply chain bottlenecks acting as the primary brake on expansion. Speaking to CNBC, Hotard stated bluntly: "If we could build 2x faster, our customers could build 2x faster, they probably would."1

1

2

2

The former Intel data center and AI group leader, who joined Nokia in 2025, identifies memory chips and power limitations as the critical supply constraints throttling construction.2

2

Nokia positions itself at the center of the AI data centers buildout by providing optical networking and IP networking technology that links facilities and connects racks within them. These connectivity sales doubled to EUR 446M in the second quarter against group sales of EUR 4.82B, helping drive Nokia shares up approximately 130% over the past year.2

2

Memory Shortage Already Impacting Markets

The silicon shortage Hotard references has tangible real-world effects. DDR5 memory prices in Germany surged 414% in a year, with memory now consuming as much as half the bill of materials on budget smartphones.2

2

This scarcity affects not just data center construction but cascades through consumer electronics markets.

Hotard maintains that demand remains robust independent of breakthrough AI models. "Even if we didn't have another frontier model released in the next three years, we could probably make tremendous progress just deploying the technology that's there today," he told CNBC.2

2

This perspective suggests current AI capabilities alone justify continued infrastructure expansion.

The $10.3 Trillion Question

While Hotard sees no evidence of overbuilding, financial analysts are sounding alarms about whether revenue can justify the spending. A paper presented at the Brookings Papers on Economic Activity by Stijn Van Nieuwerburgh estimates AI infrastructure investment will reach $10.3T between 2025 and 2032—averaging 3.63% of American GDP annually.2

2

Van Nieuwerburgh warns this would exceed the relative economic scale of historical infrastructure booms including canals, railways, electrification, highways and telecommunications.

Bain added another sobering projection last week: AI needs $6T in annual revenue by 2031 to pay for data centers now under construction, against the $1.2T to $1.8T today's AI products could generate.2

2

The gap between required and projected revenue raises questions about whether the current pace is sustainable.

Europe Lags in AI Infrastructure Race

The financing structure itself presents risks. Van Nieuwerburgh notes that funding is migrating off balance sheets into joint ventures, private credit and special purpose vehicles.2

2

Europe's position appears particularly precarious. While the EU opened bidding in July for seven AI gigafactories valued at EUR 30B, only about EUR 1B of public money is actually committed.2

2

Nokia, headquartered in Espoo, Finland, finds most of its customers outside Europe.

Beyond 5G: Intelligent Networks and Space

Hotard's vision extends beyond immediate data center connectivity. He describes 6G not as a simple speed upgrade but as a fundamental shift toward intelligent networks capable of sensing, planning and acting autonomously.1

1

Nokia is also pursuing opportunities in space communications and orbital data centers, positioning for physical AI applications including robotics and autonomous vehicles.1

1

The tension between Hotard's optimism about AI infrastructure demand and analysts' warnings about revenue sustainability will define the next phase of the buildout. Watch whether memory chip production and power generation can scale fast enough to test Hotard's thesis, and whether AI applications can generate the trillions needed to justify the spending already committed.