Ericsson Defends Custom Chips Against Nvidia Pact
CEO Per Narvinger maintains custom silicon advantage as rival Nokia pivots to GPU-based radio access network development.
Updated on Oct. 21, 2026 in Semiconductors
Live Poll
Do you trust specialized, custom-built hardware over widely available, standardized technology products?
Ericsson CEO Per Narvinger has defended the company's reliance on custom ASICs for RAN software following its latest quarterly results. Meanwhile, competitor Nokia has entered a $1 billion agreement with Nvidia to integrate GPUs into its mobile infrastructure products.
Why it matters
The split strategy highlights a fundamental debate in telecommunications over whether custom silicon or general-purpose GPUs are better suited for managing AI-native workloads in mobile networks.
Ericsson reports $3.9 billion in mobile network sales with a 49.8% gross margin, while deploying AI-native link adaptation across 60,000 basestations. Nokia aims to leverage its $1 billion Nvidia partnership to double spectral efficiency by 2028.
The players
Ericsson
A multinational networking and telecommunications company that provides mobile infrastructure and software.
Per Narvinger
The current CEO of Ericsson who is overseeing the firm's strategic focus on custom silicon hardware.
Nokia
A global telecommunications equipment manufacturer currently shifting its strategy toward GPU-based network processing.
Nvidia
A major technology company known for designing graphics processing units that are increasingly being applied to telecommunications infrastructure.
The details
Ericsson maintains that RAN algorithms are limited by transmission intervals rather than raw processing power, justifying its custom silicon approach. Conversely, Nokia argues that GPU integration allows for more efficient AI implementation, while Ericsson continues to utilize Intel CPUs for Layer 1 functions in its cloud RAN software.
Timeline
Q2 2026: Nokia reported its mobile infrastructure operating margin.
October 2026: Ericsson CEO defended custom ASICs after Q3 results.
2028: Nokia target for doubling spectral efficiency.
The Tech Race
The transition toward GPU-based processing in mobile networks represents a significant pivot from the industry's long-standing reliance on the use of Application-Specific Integrated Circuits (ASICs). This shift places Nokia in direct competition with Ericsson's entrenched proprietary hardware strategy as both companies race to optimize network AI.
The divergence in hardware strategies will dictate how effectively mobile carriers can deploy AI-driven network optimizations to improve consumer data speeds and reliability. While consumers may not see immediate changes, the hardware race determines the future cost and scalability of global 5G and future network deployments.
The takeaway
The telecommunications industry is currently split between vendors betting on specialized custom silicon and those embracing general-purpose GPU architectures for AI. Investors and operators should watch whether these differing hardware paths translate into measurable gains in network efficiency over the next three years.
Further reading
Explore deeper insights into the sector by visiting our Semiconductors page.
Live Poll
Do you trust specialized, custom-built hardware over widely available, standardized technology products?
