The telecommunications edge compute market continues as a long-term potential than a near-term reality, which will see investment start to wane over the next several years.

Technology Business Research (TBR) predicts telecommunication service providers will continue to increase yearly spending on extending the reach and capabilities of their networks to the edge, growing from $25 billion spent in 2023 to $46.5 billion spent in 2028. However, the pace of that growth will top out in 2025, which will see a near-15% year-over-year increase, before slowing to a 12% increase in 2028.

The analyst firm noted most of this edge compute spend will be based around network transformation efforts to update legacy radio access network (RAN) systems, and “deployments by hyperscalers eager to extract economic value from distributed computing use cases,” including an increased focus around artificial intelligence (AI) opportunities.

“The telecom edge compute market continues to grow, though at a slower rate than originally forecasted, as hyperscalers increase their focus on central cloud development for AI and as telcos more gradually deploy open [virtualized] RAN due to technical challenges.” TBR senior analyst Michael Soper, wrote.

This positive sentiment continues to be echoed by top industry executives.

Verizon CEO Hans Vestberg, who has been a long-time proponent of the operator being in a strong position to power edge computing services, continues to pump up that potential despite a lack so far of a significant return.

“Our portfolio of high-performance spectrum, the capacity of our fiber, and our ability to deploy and support mobile edge compute make us as the backbone of the AI economy and the partner of choice for players in the space,” Vestberg said during the carrier’s most recent earnings call. “We will power the best AI services for our customers. What set us apart with AI is our network's mobile edge computing capabilities and deep fiber footprint. By processing data closer to the source we enable real-time AI application that requires security, ultra-low latency, and high bandwidth. This is where our network shines, opening up possibilities that simply weren't feasible before.”

Will AI benefit the edge? Verizon’s management has attempted to link those possibilities to the growing hype around AI and generative AI (genAI) opportunities.

“We also in the past couple of years built storage and computing into our network as part of our mobile edge compute,” Debika Bhattacharya, chief technology solutions officer for Verizon Business, told SDxCentral in a previous interview. “The conversations we’re having with many of our customers now about generative AI and the massive amounts of data that they don’t expect to transport between the hyperscalers with their training models might be located with the edge applications or inferencing models that are at the edge. They need a high-performance network to enable those use cases, and this is this high-performance network that connects locations, connects devices, connects users, connects manufacturing locations, headquarters, all those to the hyperscaler data centers, along with the edge compute that is going to be a game changer.”

TBR noted in its report that the current shift by hyperscalers toward boosting their central data centers to deal with the initial AI surge will need to shift over time, though the firm expects this shift could extend outside of its current market projections.

“Ultimately, hyperscalers must pivot their spend from central to edge build outs to achieve the latency and quality of service that new network use cases will require,” the firm explained. “TBR believes hyperscalers will extend their cloud footprints closer to endpoints through this decade and expects hyperscaler capex will shift significantly from central cloud to edge cloud beyond the forecast period.”