Ai demand fuels memory chip price surge, threatening ericsson’s profitability
The relentless surge in artificial intelligence adoption is triggering a critical bottleneck in the semiconductor industry, sending memory chip prices skyrocketing and casting a shadow over the financial outlook for networking giants like Ericsson.
Next-gen chips lagging behind smartphone and ai innovation
While consumers marvel at the increasingly sophisticated capabilities of their AI assistants – from real-time translation to advanced photography – the unseen engine driving this revolution is consuming an astonishing 70% of the world’s memory chips. Foundries like TSMC, the primary manufacturer of cutting-edge silicon, are operating at maximum capacity, prioritizing orders from industry leaders like Apple and Nvidia, leaving Ericsson and other network equipment providers facing extended lead times and significantly inflated costs for crucial components like Application-Specific Integrated Circuits (ASICs).
Ericsson forced to renegotiate contracts
Per Narvinger, Ericsson’s head of mobile networks, bluntly states, “Right now, many of the AI workloads are competing for the same wafers that we also are interested in.” The company is reportedly exploring renegotiation options with its customers, a necessary step to mitigate the impact of these rising costs on its profit margins. Ericsson’s recent workforce reductions, down from 105,500 in 2022 to 88,000, underscore the pressure felt within the organization.
2Nm process pushes network gear behind the curve
Mobile infrastructure is currently utilizing chips that are a node behind the most advanced smartphones and some AI applications, operating at the 5nm level. However, the demands of mobile networks require chips just slightly trailing these cutting-edge workloads. This year saw the release of the first smartphone powered by a 2nm application processor – Samsung’s Galaxy S26 and S26+ – utilizing the Exynos 2600. Despite this technological leap, the cost implications are far-reaching.
A race to lower lead times
Narvinger acknowledges the situation, stating that as AI firms transition to the 2nm process node in the coming months, TSMC aims to alleviate production bottlenecks and potentially lower prices. But Ericsson isn’t waiting idly; the company is proactively approaching its clients, seeking to adjust existing contracts and secure more favorable terms. Nokia CEO Hotard echoes this sentiment, noting that customers largely understand the necessity of price adjustments driven by AI’s impact on memory chip costs.
The price of progress – and a warning to consumers
Ultimately, this isn’t just a supply chain issue; it’s a direct consequence of AI’s exponential growth. As BT CEO Allison Kirkby observes, “AI data centers require a huge haul of chips, the same silicon that many smartphone manufacturers need to obtain.” The inflationary pressures on memory chips, fueled by AI, are inevitably filtering down to consumers, potentially impacting the price of smartphones and other devices. Ericsson’s future hinges on its ability to navigate this challenging landscape, demonstrating that technological advancement doesn’t always equate to straightforward prosperity.