Vodafone Egypt has begun commercial deployment of Ericsson’s RAN Processor 6672, putting AI processing directly at its radio sites and making Egypt one of the first markets to adopt the hardware following its commercial launch in Japan.
The processor is built for 5G Advanced and AI workloads, allowing the network to run AI at the site itself rather than routing it to a central location — so capacity can respond to traffic as users move and usage patterns shift.
Ericsson says the unit delivers up to four times the processing capacity of previous hardware while cutting energy consumption by more than 50%, and allows a single processor to replace three basebands, simplifying site infrastructure.
Two layers at once
The deployment arrives days after Vodafone Egypt and Cassava Technologies launched the country’s first AI factory, offering government and business customers GPU-as-a-Service on NVIDIA hardware with data hosted inside Egypt.
Read together, the two show Vodafone Egypt building AI capability at both ends of the network simultaneously: centralised compute for enterprise and government workloads, and distributed processing at the radio edge for network operations. Both form part of the operator’s plan to invest more than EGP 20 billion ($300 million) this fiscal year.
“We are investing for a future-ready network,” said Catalin Buliga, Technology Director at Vodafone Egypt. The processor, he said, increases efficiency “in terms of processing power and energy consumption for our Ericsson Radio Sites, while enabling AI to run directly on the Radio Site for better customer experience.”
The edge bet, across the continent
Vodafone Egypt is the third major African operator in recent months to commit to AI at the radio layer, each with a different vendor.
MTN Group’s chief technology and information officer Charles Molapisi set out a plan in May to convert the group’s African tower estate into a distributed AI compute grid, installing GPU configurations at base stations so the same hardware runs both the radio network and edge inference. MTN Digital Infrastructure then invested in ODC’s Series A to develop AI-native RAN platforms.
MTN South Africa separately expanded ZTE’s AI-driven EnergyTurbo system across the Western Cape this year, after a Cape Town trial lifted per-site energy savings from 2.8% under baseline features to 13.5% with AI.
The logic is consistent. Operators hold thousands of powered, connected, secured sites. Running AI there means the network adapts locally without the latency of central processing — and it turns existing infrastructure into compute capacity.
Energy is the argument
The efficiency claims matter more in Africa than in most markets.
Energy is among the largest operating costs for African mobile operators and the binding constraint on network expansion. A site that does more with less power, and needs one processor rather than three, changes the economics of rollout — particularly as grids remain unreliable and diesel backup expensive.
Dalia Merheb, Country Manager and Head of Sales at Ericsson Egypt, said the deployment addresses growing capacity requirements while simplifying sites and improving energy efficiency, creating “a stronger foundation for continued network evolution.”
Ericsson is Vodafone Egypt’s main radio access network supplier across Greater Cairo and the Delta. Neither company disclosed the number of sites, the rollout timeline or the value of the deployment.
