AI at the Edge of 5G
How network architecture is redefining competitiveness in Asia
A short but pointed intervention by Rafael A. Junquera, Co-Founder and Editorial Director of TeleSemana.com, is drawing attention inside the global telecom industry. Writing in Spanish, Junquera frames the current evolution of 5G not as a technology upgrade, but as a strategic fork in the road that could shape power relations in telecommunications for decades.
Junquera’s remarks accompany the release of a new TeleSemana Position Paper, which is available as a free download via TeleSemana.com. The document analyses how Nokia, Ericsson and Huawei are pursuing fundamentally different approaches to artificial intelligence in the Radio Access Network (RAN) — the core layer that connects mobile devices to the network.
According to Junquera, operators now face a “brutal dilemma”: the three strategies are not interoperable paths, but competing visions of how intelligence, control and value creation should be embedded in future networks.
Three Competing Architectures
Junquera identifies three distinct models.
Nokia, in close partnership with NVIDIA, is pushing AI toward the edge of the network. By deploying GPU-accelerated intelligence closer to base stations, Nokia argues that operators can enable low-latency services, real-time optimisation and new enterprise use cases. NVIDIA’s large-scale investments in telecom-grade AI infrastructure underline that this is a long-term strategic commitment rather than a pilot experiment.
Ericsson is taking a more cautious route. AI functions are largely centralised in cloud environments, prioritising security, governance and operational stability. This model reduces risk and vendor lock-in, but sacrifices immediacy at the network edge.
Huawei follows the most vertically integrated approach. AI is deeply embedded into its own proprietary hardware at the access node, maximising performance and responsiveness. The trade-off, Junquera notes, is increased technological and geopolitical dependency.
The result, he argues, is not a technical debate but a structural choice with long-term consequences for operators’ autonomy and business models.
Industry Reaction: Performance Versus Control
Junquera’s analysis quickly triggered responses from industry professionals. Daniel Estrada, 4G/5G Senior RF Engineer, publicly commented that Huawei’s approach may deliver a decisive advantage:
“By integrating AI directly into access nodes, operators can analyse traffic in real time, optimise performance immediately and enable value-added services — the long-promised monetisation of 5G.”
Estrada’s question — whether Ericsson risks falling behind — highlights the tension Junquera describes: centralised control versus edge-level performance.
Why This Matters Now
TeleSemana’s decision to publish the analysis as a Position Paper is significant. Unlike daily reporting, such documents are designed to frame strategic debates among executives, regulators and policymakers. TeleSemana, which has served the telecom sector in Latin America and Spain for over two decades, positions itself not as a vendor but as an industry agenda-setter.
While the report is offered free of charge, it functions as a strategic signal: the choices operators make today about AI placement in RAN infrastructure will shape competition, sovereignty and innovation capacity well beyond the 5G cycle.
As Junquera makes clear, the AI war inside telecom networks is largely invisible to consumers — but its outcomes will define who controls connectivity in the decades ahead.
Source (explicit)
- Author: Rafael A. Junquera
- Role: Co-Founder & Editorial Director
- Organisation: TeleSemana.com
- Document: TeleSemana Position Paper on AI and the Future of RAN
- Availability: Free download via TeleSemana.com
