Enterprise AI Networking Stalls as Infrastructure Complexity Surges
IDC research reveals a growing gap between AI ambitions and execution as edge bandwidth demands spike.
Enterprise adoption of AI for network operations has hit a plateau despite high organizational expectations. While companies intend to deploy advanced AI to manage their systems, many remain trapped in selective use due to security concerns, integration hurdles, and a critical shortage of skilled talent.
According to IDC research reported by Network World, this stall comes at a precarious time. Distributed AI is driving a massive surge in infrastructure demands, with edge bandwidth projected to grow by an average of 51% over the next year. This growth strains NetOps teams who must keep pace with the physical requirements of AI workloads while simultaneously trying to implement AI to manage that very complexity.
The Execution Gap
Organizations face a dual challenge: they must build the infrastructure necessary to support AI while attempting to use AI to automate the resulting network sprawl. This has created a significant gap between intent and execution, where the speed of network growth is outpacing the ability of human teams to manage it manually.
To bridge this gap, 81% of organizations are increasing their spending on managed service providers (MSPs) to support their AI initiatives. The desired outcomes are specific: 31% of respondents aim to boost IT service levels and capabilities, while 30% target improved operational efficiency. Furthermore, procurement preferences are shifting; organizations show less confidence in platform-centric approaches and are increasingly opting for best-of-breed solutions to solve specific networking challenges.
The Shift Toward Autonomy
The industry is moving toward a model where AI does more than suggest fixes. IDC data shows that 46% of respondents prefer AI systems capable of both determining and executing network actions autonomously. This represents a fundamental shift in the professional landscape, moving the network engineer from a manual configurator to an orchestrator of AI systems.
However, this transition introduces new risks. Brandon Butler, a senior research manager at IDC, notes that from a security perspective, "you have to fight AI with AI." The complexity of distributed AI environments means that failure to bridge the planning-execution gap could lead to widespread operational instability.
What to Watch
As organizations move out of the pilot phase, the next two years will be decisive. Mark Leary, a research director at IDC, suggests that 2026 will be the tipping point when organizations discover if AI in networking delivers a real operational impact or remains stuck in a cycle of limited trials. The primary indicators of success will be whether autonomous systems can stabilize the projected 51% growth in edge bandwidth without compromising security.