Dell’Oro Group Forecasts $1 Trillion AI Back-End Switch Spending by 2030

Dell’Oro Group forecasts data center switch spending for AI back-end networks to approach $1 trillion from 2026 to 2030, with scale-up accounting for more than half of the market by 2030. The projection affects procurement and network-capex planning for operators supporting accelerated compute.

Market outlook

The report projects cumulative spending on data center switches deployed in AI back-end networks to approach $1 trillion over the next five years from 2026 through 2030. It also forecasts scale-up networking to account for more than half of the market by 2030.

“We have substantially raised our revenue and port-shipment forecasts since our January 2026 outlook to reflect both stronger demand for accelerated compute and higher infrastructure costs,” said Sameh Boujelbene, Vice President at Dell’Oro Group. “Despite this significant upward revision, we believe the market will remain supply-constrained rather than demand-constrained over our five-year forecast horizon.”

Scale-up technology mix and timing

According to the report, scale-up networking will account for more than half of the AI back-end switch market by 2030. It also states scale-up will more than double the total addressable market for switches deployed in AI back-end networks.

“Scale-up networking will account for more than half of the AI back-end switch market by 2030 and will more than double the total addressable market for switches deployed in AI back-end networks,” said Boujelbene. The report says NVLink currently dominates scale-up deployments, while Ethernet and UALink are gaining traction as operators deploy custom and merchant accelerators beyond NVIDIA’s GPUs.

Ethernet outlook across AI back-end networks

The report expects Ethernet to become the leading scale-up technology by the end of the decade. It also expects Ethernet to remain the long-term leader across the broader AI back-end networking market covering scale-up, scale-out, and scale-across.

Reported pacing and architecture notes

The report expects the ramp of 1.6 Tbps to be at least five times faster than 800 Gbps. It links the forecast pace to increased AI infrastructure scale and adoption across scale-up, scale-out, and scale-across networks, with further acceleration expected during the transition to 3.2 Tbps.

It also says advanced pluggables, XPO, NPO, and CPO will coexist, with XPO and NPO gaining earlier momentum and CPO ramping up over time as the technology and supply chain mature. Despite Ethernet dominance, the report says multiple networking technologies and architectures will coexist, creating multibillion-dollar opportunities as operators balance accelerator platforms, performance requirements, scale, and ecosystem preferences.

Included report scope

Dell’Oro Group’s Data Center Switch – AI Back-end Networks Forecast Report covers switches deployed in AI back-end networks to connect accelerated servers. It includes market information on scale-up, scale-out, and scale-across technologies and use cases, and it covers revenue split by regions and customer segments, including Top 4 US Cloud, Top 3 China Cloud, Neocloud, AI Model Builders, and Large Enterprises.

Overall, the note forecasts rapid growth in AI back-end switch spending through 2030 and assigns the largest share to scale-up networking, while projecting Ethernet as the leading scale-up technology by the end of the decade. This Analyst Signals brief reflects a neutral, fact-based summary of the original research note.