Nokia CEO Justin Hotard has a simple answer for anyone wondering whether the artificial intelligence infrastructure boom has already gone too far: the industry would move considerably faster if the physical world allowed it. Hotard said data center customers could build at roughly twice today’s pace without the supply constraints limiting projects, framing scarcity rather than weak demand as the brake on AI infrastructure expansion. His argument lands at a sensitive moment for the sector, as investors weigh enormous capital commitments against questions about how quickly AI demand can turn infrastructure spending into sustainable revenue. For Nokia, the debate also reinforces its transformation from a traditional telecom equipment heavyweight into a supplier positioned deeper inside the physical machinery of the AI economy.

Nokia Sees Supply, Not Demand, Setting the Pace

Hotard pushed back against the idea that the current construction wave represents obvious overbuilding, pointing instead to customers that remain constrained by their ability to secure the infrastructure required for expansion. Memory chips and energy capacity rank among the bottlenecks affecting the sector, meaning developers cannot necessarily translate every planned megawatt or infrastructure order into operating capacity on their preferred timeline. “I don’t think you can say in any manner we’re overbuilding today because reality is that if we could build 2x faster, our customers could build 2x faster, they probably would,” Justin Hotard said. “So that gives me confidence that we’re still in the early days” of the AI buildout, Hotard added.

A slowdown caused by insufficient customers would signal that operators have begun running ahead of AI consumption, while a slowdown caused by scarce power, memory and infrastructure components suggests latent demand remains trapped behind supply constraints. However, Hotard’s “2x faster” assessment remains an executive view of customer appetite rather than an independently measured forecast that industry construction would literally double if those bottlenecks disappeared. Nokia already has commercial exposure to that scarcity dynamic, with Hotard previously saying customers were placing longer-term orders as they attempted to secure supply in a constrained market.

Nokia Is Becoming a Picks-and-Shovels AI Company

The AI boom increasingly rewards companies that sell the infrastructure surrounding accelerators rather than the accelerators alone, and Nokia has moved aggressively toward that layer of the stack. Its portfolio includes optical and IP networking technology that connects computing resources inside data centers and links facilities across larger distributed infrastructure footprints. That puts Nokia close to a fundamental challenge facing increasingly large AI systems: adding processors means little if operators cannot move enormous volumes of data between them quickly enough. The company can therefore capture AI infrastructure spending without needing to compete directly with the semiconductor companies producing the GPUs that dominate the industry’s headlines.

The shift has already started appearing in Nokia’s financial performance, giving Hotard’s demand argument more substance than rhetoric alone. Nokia reported €446 million in net sales from AI and cloud customers during the second quarter of 2026, more than double the level from a year earlier, while order intake from those customers reached €2.8 billion. The company said roughly half of that quarterly order volume should convert into revenue over the following 12 months, although Hotard cautioned that order patterns can remain lumpy and that investors should not expect the same intake every quarter. Nokia’s comparable operating profit also climbed 18% to €434 million during the quarter, while comparable net sales reached €4.82 billion.

AI Infrastructure Demand Extends Beyond Frontier Models

Hotard also challenged another assumption hanging over the infrastructure boom: that developers need an endless succession of increasingly capable frontier models to justify continued investment. Instead, he sees considerable runway in deploying models and AI capabilities that already exist, effectively separating infrastructure demand from the industry’s race to produce the next major model generation. “I think the reality is that we’re even where we are today is so early in the deployment, and the power of these models is so massive that there’s a lot of there’s a lot of demand to run in just deploying,” Hotard said. That argument shifts attention from model creation toward adoption, where enterprises, cloud providers and other organizations still need computing, networking and supporting infrastructure to put existing AI capabilities into production.

“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.” Hotard’s position implies that AI infrastructure has two potential engines of demand: increasingly powerful models that require larger computing clusters, and wider deployment of capabilities already available today. The second engine could become particularly important if frontier development slows, because inference, enterprise adoption and large-scale implementation can continue consuming infrastructure even without another dramatic leap in model capability. Meanwhile, that thesis gives infrastructure suppliers such as Nokia a broader addressable opportunity than a market dependent entirely on the release cycles of frontier AI developers.

Nokia’s AI Bet Depends on Infrastructure Staying Scarce

For Nokia, supply constraints present both an opportunity and a complication as the company expands beyond its historic telecommunications stronghold. Scarcity can encourage customers to place longer-term orders and strengthen demand visibility for networking suppliers, but the same shortages can restrict how quickly equipment makers convert that demand into deployed infrastructure and revenue. Rising memory costs have already affected telecommunications equipment vendors, showing how the AI boom can simultaneously create new customers and tighten the components required to serve them. Nokia’s challenge will be turning its position inside that constrained supply chain into durable growth as the market eventually adds more power, networking capacity and hardware.