Ai chip squeeze threatens ericsson, could hike smartphone prices
The relentless surge in demand for artificial intelligence is creating a ripple effect throughout the tech world, and now it’s impacting network infrastructure giants like Ericsson. The insatiable appetite of AI data centers for memory chips is driving up prices and causing significant delays, potentially leading to increased costs for consumers – starting with your next smartphone.
The ai-fueled semiconductor crunch
It’s no secret that AI workloads are booming. What’s less known is just how much they’re consuming of the world's semiconductor resources. A staggering 70% of global memory chip supply is currently being devoured by AI data centers, leaving other industries scrambling for what's left. Foundries like TSMC, the world’s leading chip manufacturer, are operating at full capacity, prioritizing their biggest clients—Apple and Nvidia—for access to their advanced 3nm fabrication processes.
Ericsson, which relies on application-specific integrated circuits (ASICs) for its radios and baseband equipment, finds itself further down the line, facing inflated prices and extended lead times. While the latest smartphones and AI hardware are benefiting from 5nm technology, Ericsson’s network equipment typically lags slightly behind, utilizing chips that are now a generation behind. The Samsung Galaxy S24 and S24+ were among the first to use 2nm chips, highlighting the rapid pace of advancement.
Ericsson's bold move: renegotiating contracts
Per Narvinger, head of Ericsson’s mobile networks business group, recently highlighted the competitive landscape: “Right now, many of the AI workloads are competing for the same wafers that we also are interested in.” Faced with escalating costs and uncertain timelines, Ericsson has reportedly taken a drastic step – approaching its customers to renegotiate existing contracts and secure price increases. The hope is that the anticipated shift to the 2nm process node in the coming months will alleviate some of the pressure on TSMC, reducing bottlenecks and potentially lowering prices.
But the situation is far from guaranteed. The widespread adoption of 2nm chips by AI firms could simply exacerbate the problem, keeping prices elevated. Ericsson’s predicament underscores a broader trend: the dominance of AI is reshaping the entire semiconductor supply chain, with consequences that extend far beyond the realm of artificial intelligence itself.
Nokia, Ericsson's main competitor, is facing similar challenges. Justin Hotard, Nokia’s CEO, confirmed that they, too, are experiencing longer lead times and increased costs. He noted, however, that many customers understand the underlying causes and are willing to accept the higher prices. The message is clear: AI's insatiable demand is driving up costs across the board, and consumers will ultimately feel the pinch.
Allison Kirkby, CEO of BT (formerly British Telecom), succinctly summarized the situation: AI data centers require a “huge haul of chips,” the very same silicon that smartphone manufacturers and others desperately need. The economic principle is simple: limited supply combined with high demand inevitably leads to higher prices – a lesson learned from Econ 101.
The impact on Ericsson extends beyond its bottom line. The company has already implemented workforce reductions, shrinking from 105,500 employees in 2022 to 88,000 just three months ago, partially attributed to the challenges posed by AI-driven cost pressures. And just as Ericsson is seeking customer concessions, smartphone manufacturers may soon be forced to raise prices on their devices. The era of cheap AI-powered devices may be drawing to a close, demonstrating that the benefits of artificial intelligence come at a cost.