The adoption of artificial intelligence (AI) in enterprise networking is encountering significant hurdles, according to recent findings from IDC. Despite high expectations for AI's impact on network infrastructure and operations, many organizations remain stalled at early stages of implementation.

High Expectations vs. Actual Progress

IDC's 2026 AI in Networking Special Report highlights a troubling trend: while organizations have ambitious plans for AI integration, actual advancements have largely stagnated. Mark Leary, research director at IDC, notes, "The people who were at select use were still at select use. The people who were at substantial use were still at substantial use. Over 18 months, they hadn’t moved at all, really." This stagnation has led to a widening gap between organizational intent and execution.

Persistent Barriers to AI Adoption

Organizations are striving to implement AI across two primary areas: enhancing network infrastructure to support AI workloads and optimizing network operations through AI applications. However, they face several persistent challenges that impede progress.

Security concerns remain paramount, with many organizations viewing AI as both a potential threat and a solution. Brandon Butler, senior research manager at IDC, emphasizes that to combat AI-driven threats, organizations must leverage AI technology themselves, stating, "You have to fight AI with AI from a network security perspective." Furthermore, integration challenges with existing systems and a lack of skilled talent are significant barriers. Many organizations feel ill-equipped to select the right AI solutions, prompting 81% to increase spending on managed service providers (MSPs) for support.

Growing Infrastructure Demands

Despite the sluggish pace of AI adoption, the demands on network infrastructure are escalating. Butler asserts, "The pressure is already on the network." IDC reports that 89% of data centers anticipate a bandwidth increase of at least 11% over the next year, primarily driven by AI workloads. This demand extends to inter-data center connectivity, with 91% of organizations expecting similar growth, indicating strain on distributed architectures.

Cloud infrastructure is witnessing even sharper increases in demand, with organizations forecasting a 49% rise in bandwidth for cloud connectivity. Leary points out that the integration of cloud platforms is crucial, noting, "The cloud is almost always involved," as many organizations utilize multiple data centers alongside cloud services.

Edge Deployments and Future Growth

The network edge is emerging as the next significant growth area for AI workloads. Currently, 27% of organizations have deployed AI at the edge, with 54% planning to do so within the next two years. Butler observes, "Folks who are leveraging AI more extensively are already pushing workloads to the edge," indicating a market shift. This transition is expected to further strain networks, with edge bandwidth projected to grow by an average of 51% in the coming year.

Autonomous Operations and Trust in AI

Research indicates a growing preference for autonomous AI operations. Nearly half of the surveyed organizations (46%) express a desire for AI systems capable of autonomously determining and executing network actions, while 41% favor a guided approach. This trend reflects a recognition that the complexity of modern networks often exceeds the capacity for manual management, particularly in light of the ongoing talent shortage.

Changing Platform Strategies

Organizations are increasingly skeptical of platform-centric approaches, opting instead for best-of-breed solutions tailored to specific needs. Leary suggests this shift may stem from previous disappointments with platform expectations, stating, "There has to have been some disappointment. People expected simplicity, cost savings, and stronger outcomes, but many platforms didn’t fully deliver." As hyperscale cloud providers solidify their roles as strategic partners, the reliance on cloud ecosystems in future network architectures becomes clearer.

Path Forward for Network Leaders

For network leaders, the challenge lies in successfully executing AI initiatives. IDC recommends focusing on targeted, high-impact use cases, transitioning from reactive to proactive operations, and leveraging external expertise when internal resources are insufficient. Leary advises, "Avoiding a problem pays way more dividends than fixing one faster." The increasing reliance on managed services highlights a recognition among enterprises that collaboration with providers can address challenges more effectively.

As AI adoption in networking continues to evolve, organizations must adapt to rising infrastructure demands, accelerate edge deployments, and align their AI ambitions with tangible progress. The future of AI in networking hinges on how quickly enterprises can close the gap between their plans and actual deployment.

Source: Network World News