AI Integration in Networking Faces Significant Delays
Despite years of planning, enterprise AI adoption in networking is not progressing as expected. Recent research highlights a growing gap between ambitious AI integration plans and the actual execution, driven by increasing pressure on network infrastructure.
The 2026 IDC AI in Networking Special Report indicates that many organizations, initially optimistic about transitioning from early AI applications to more advanced deployments, have not made significant progress. 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 underscores the challenges facing network leaders.
Stalled Progress and Recurring Challenges
While network teams are eager to adopt AI, they are hindered by persistent obstacles. Security concerns remain paramount, acting as both a barrier to deployment and a primary use case for AI. Brandon Butler, senior research manager at IDC, states, "You have to fight AI with AI from a network security perspective," reflecting the reality that cyber threats are increasingly sophisticated.
Moreover, integration difficulties with existing systems and a lack of skilled personnel exacerbate the situation. According to Leary, many organizations feel their staff cannot adequately evaluate or select appropriate solutions, prompting 81% of companies to increase spending on managed service providers (MSPs) to bolster their AI initiatives.
Infrastructure Demand Is Accelerating
Even with AI adoption lagging, its impact on infrastructure is becoming 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." For example, 89% of data centers anticipate increasing bandwidth by at least 11% in the coming year, primarily driven by AI workloads. This demand extends to inter-data center connectivity, with 91% of organizations expecting similar growth, highlighting the strain on distributed architectures.
Cloud environments are experiencing even more substantial increases, with organizations forecasting a 49% rise in bandwidth for cloud connectivity over the next year. Leary notes, "The cloud is almost always involved," indicating that many enterprises are utilizing a mix of cloud platforms alongside data centers.
Edge Deployments Set the Next Wave
Beyond traditional data centers and cloud environments, edge computing is emerging as a significant growth area. Currently, 27% of organizations have deployed AI workloads at the edge, with 54% planning to do so within the next two years. Butler states, "Folks who are leveraging AI more extensively are already pushing workloads to the edge," suggesting this trend is indicative of the market's future directions.
This shift is expected to further increase network demands, with edge bandwidth projected to grow by an average of 51% in the coming year. As AI becomes more distributed, network teams will face the challenge of managing increased complexity while ensuring performance and security.
A Shift Toward Autonomous Operations
The research indicates a notable shift in how organizations want to utilize AI, with nearly half (46%) preferring systems that can autonomously determine and execute network actions. Another 41% favor a guided approach, while 13% prefer minimal AI involvement. Butler highlights, "Two years in a row, the largest group said they want AI to both determine and execute actions," showcasing a growing trust in automation amid rising network complexity.
Rethinking Platform Strategies
Enterprise organizations are increasingly shifting away from platform-centric approaches in favor of best-of-breed solutions that better meet specific needs. Leary points out that many expected simplicity and cost savings from integrated platforms but found them lacking. This dissatisfaction has pushed organizations to seek strategic partnerships with hyperscale cloud providers, emphasizing the crucial role of cloud ecosystems in future networking architectures.
For network leaders, the path forward involves executing targeted, high-impact use cases, transitioning from reactive to proactive operations, and leveraging external expertise to overcome internal resource limitations. As Leary states, "Avoiding a problem pays way more dividends than fixing one faster." The growing reliance on managed services reflects a recognition that enterprises can collaborate with providers to tackle complex challenges.
As infrastructure demands rise and edge deployments accelerate, organizations must adapt their AI strategies to bridge the widening gap between ambition and execution. The next phase of AI in networking will be determined by how effectively enterprises can turn their adoption plans into tangible progress.
In summary, this situation is not about whether AI will reshape networking but rather how swiftly organizations can adapt before the gap becomes unmanageable.
Source: Network World News