AI for IT Stalls as Network Complexity Increases
Enterprise ambitions for AI integration in IT remain high, yet new research indicates a significant gap between plans and execution, compounded by growing pressures on network infrastructure. The 2026 IDC AI in Networking Special Report highlights that while organizations aim for advanced AI deployments, many are still in the early stages, with minimal progress observed.
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 reveals the widening divide between organizational intent and actual implementation.
Stalled Progress, Familiar Challenges
Organizations are actively pursuing AI across two main fronts: enhancing AI workloads within network infrastructure and applying AI to streamline network operations. However, persistent challenges hinder this progress.
Security remains the top concern, acting both as a barrier to deployment and a primary use case for AI itself. Brandon Butler, senior research manager at IDC, states, “You have to fight AI with AI from a network security perspective,” emphasizing the need for advanced security measures as malicious actors increasingly leverage AI technologies.
Integration with existing systems and a lack of skilled talent are additional obstacles. Leary observes, “Most folks don’t feel their staff can fully evaluate and select the right solutions.” Consequently, many organizations are increasing their reliance on managed service providers (MSPs), with 81% reporting heightened spending in this area to support their AI initiatives.
Infrastructure Demand is Accelerating
Despite slow adoption rates, the pressure on network infrastructure due to AI is already evident. Butler remarks, “The pressure is already on the network. The question now is whether organizations can keep up with what AI is demanding of their infrastructure.”
According to the report, 89% of data centers anticipate a bandwidth increase of at least 11% within the next year, driven by AI workloads. This demand extends to inter-data center connectivity, with 91% expecting similar growth, highlighting the strain on distributed architectures.
Cloud environments are experiencing even sharper increases, with organizations predicting an average bandwidth rise of 49% for cloud connectivity in the coming year. “The cloud is almost always involved,” Leary adds, noting that many organizations are integrating one cloud platform with multiple data centers.
Edge Deployments Set the Next Wave
The network edge is emerging as a critical growth area. Currently, 27% of organizations have deployed AI workloads at the edge, and 54% plan to do so within two years. Butler comments, “Folks who are leveraging AI more extensively are already pushing workloads to the edge,” indicating a trend that will significantly heighten network demands.
With edge bandwidth projected to grow by an average of 51% in the next year, network teams will face increased complexity in managing these distributed environments while ensuring performance and security.
A Shift Toward Autonomous Operations
The research also highlights a shift in organizational preferences for AI usage. Nearly half of the respondents (46%) favor AI systems capable of autonomously determining and executing network actions, while 41% prefer a guided approach. “Two years in a row, the largest group said they want AI to both determine and execute actions,” Butler notes, reflecting the growing trust in automation amid the complexities of modern networks.
Rethinking Platform Strategies
As enterprises face these challenges, there is a noticeable decline in confidence regarding platform-centric approaches. Organizations are increasingly opting for best-of-breed solutions that more effectively address specific needs. Leary acknowledges, “There has to have been some disappointment,” as many platforms have not delivered the expected simplicity and cost savings.
Meanwhile, hyperscale cloud providers are solidifying their roles as strategic partners in AI networking, underscoring the importance of cloud ecosystems in future network architectures.
For network leaders, execution remains a crucial challenge. IDC suggests starting with targeted, high-impact use cases, transitioning from reactive to proactive operations, and leveraging external expertise where internal resources are lacking. “Avoiding a problem pays way more dividends than fixing one faster,” Leary advises.
The increasing reliance on managed services and MSPs indicates that enterprises recognize the need for collaboration with providers to tackle these challenges. As infrastructure demands rise and edge deployments accelerate, the urgency for effective AI-driven operations intensifies.
“This isn’t about whether AI will reshape networking,” Leary concludes. “It’s about how quickly organizations can adapt before the gap becomes too wide to close.”
How Network Support and Use of AI Could Benefit Business
The study reveals that network leaders anticipate several business outcomes from AI in networking. Key expectations include enhancing IT service levels and capabilities (31%) and improving operational efficiency (30%), with increased worker productivity and revenue also highlighted. Interestingly, lowering operating costs ranks seventh, indicating a shift in perspective toward using AI as a means to enhance operational effectiveness.
IDC suggests that targeted use cases, such as automated configuration validation and AI-driven threat response, could yield measurable improvements while fostering the organizational trust necessary for more ambitious AI deployments. For network leaders, incremental progress may represent the most reliable pathway to achieving their anticipated outcomes.
“It doesn’t have to be handing the keys of your kingdom to AI to really get some benefits from these AI tools,” Butler emphasizes, advocating for a balanced approach to AI integration in networking.
Source: Network World News