Ahead of Consensus.
Intelligence across tech and capital markets, for investors, policymakers, and business leaders.
9 minute read
Every earnings report carries some burden of proof, but few this year have carried as much as Nokia’s. The stock had spent the better part of eighteen months rewriting its own valuation, climbing on the promise that a century-old telecom equipment maker could reinvent itself as critical infrastructure for the artificial intelligence economy. That promise had outrun the arithmetic. Shares fell nearly 40 percent from their June peak of EUR 14.97 as investors, spooked by a profit warning from Swedish rival Ericsson over rising memory chip costs, began asking whether Nokia’s AI narrative was substance or momentum.
Thursday’s second-quarter results did not settle that question so much as complicate it usefully. Comparable operating profit rose 18 percent year over year to EUR 434 million, ahead of the roughly EUR 382 million analysts had expected, while comparable net sales reached EUR 4.82 billion. The numbers arrived with enough texture, and enough contradiction, to reward close reading rather than headline scanning. Nokia is neither the legacy hardware vendor its critics describe nor the pure AI infrastructure play its bulls have priced in. It is something more interesting: a company mid-transformation, with the financial statements to prove both the promise and the strain.
Here is where discipline is required of any serious reader. Nokia’s reported operating margin turned negative, falling to negative 1.0 percent from a positive 3.3 percent a year earlier, producing an operating loss of EUR 50 million against a EUR 147 million profit in the prior year. Reported earnings per share came in at zero, down from EUR 0.02. None of this reflects deteriorating demand. It reflects an acceleration of restructuring charges that management chose to absorb now rather than later, a decision that says as much about confidence as it does about cost.
Nokia also reclassified two smaller businesses, Fixed Wireless Access CPE and Enterprise Campus Edge, as discontinued operations, agreeing to sell the former to Inseego and calling a sale of the latter highly probable. That reclassification, not operational outperformance, is what allowed full-year comparable operating profit guidance to rise technically to a range of EUR 2.1 billion to EUR 2.6 billion. Investors who mistake this for an upgrade are reading the footnote wrong. Nokia was careful, and correct, to describe it as a technical revision. The distinction between comparable and reported results is not accounting theater here; it is the clearest lens available into a company simultaneously investing, restructuring, and growing, three activities that rarely show up as one clean number.
The most consequential news this quarter may not have been in the earnings table at all. Nokia has agreed to acquire NXP’s semiconductor fabrication campus in Chandler, Arizona, initially leasing capacity there from early 2027 before converting the site to indium phosphide production for optical components, with full ownership expected by early 2029. Paired with a San Jose facility ramping production later this year and a tenfold expansion of packaging capacity in Pennsylvania, Nokia is assembling something rare among network equipment makers: domestic American manufacturing depth for the components that AI data centers cannot function without.
This build-out sits alongside Nokia’s July launch of what it calls the industry’s first commercial AI-native radio access network platform, developed with Nvidia and promising more than a doubling of spectral efficiency by 2028. The Nvidia relationship, anchored by a billion-dollar equity investment completed last November that made the chipmaker a roughly 2.9 percent Nokia shareholder, has done more than lend credibility. It has given Nokia a partner whose own capital is now tied to the thesis that radio networks, not just data centers, will need to be rebuilt for artificial intelligence workloads. Few telecom equipment companies can say the same.
Markets are forward-looking instruments, and Nokia’s stock behaved like one in the days before this report, rallying more than 5 percent across two sessions as analysts at SEB Equities and Danske Bank raised targets and JPMorgan lifted its price target on the New York-listed shares to $21 from $14. That rally reflected anticipation, not confirmation. What Thursday’s results confirmed is narrower but more durable: that AI-driven demand is real, that it is showing up in orders and revenue rather than merely in investor decks, and that converting that demand into reported profit will require Nokia to absorb real costs along the way.
The second half of the year now carries the weight of proving this was not a single strong quarter but the shape of things to come. Nokia reports again on October 22. Between now and then, the company’s task is unglamorous but essential: turn EUR 2.8 billion in orders into revenue, finish the restructuring it has chosen to accelerate, and let the reported numbers, not just the comparable ones, begin telling the same story the stock has been telling since last year.