Most companies adopting artificial intelligence are ignoring a key limitation: the network that carries AI workloads.
A Broadcom report shows nearly all businesses use cloud strategies and plan large-scale AI rollouts, but fewer than half think their networks can support the load. This mismatch between goals and infrastructure has left many facing performance challenges that hurt AI initiatives.
Hybrid cloud adoption outpaces network readiness
The 2026 State of Network Operations study surveyed enterprise IT teams and found hybrid cloud is now the standard model. Only 23% have added AI-powered networking tools, while 70% are still in early stages of network automation.
Tyron Silk, a senior solutions architect at CASA Software, said the push to deploy AI has diverted attention from the network. “Without full visibility across complex hybrid setups, AI workloads suffer from latency, congestion and unpredictable performance,” he said.
Related: Mandela Day drives skills training, healthcare and jobs
Almost every company in the survey called better network observability vital for AI success. Key priorities include real-time flow monitoring, user experience tracking, and stronger security. Yet 87% of organizations have limited visibility into internet and cloud environments, and 95% lack insight into critical parts of public cloud infrastructure.
Visibility gaps create operational risks
The findings show these visibility gaps have real consequences. Over half of respondents said observability tools are needed to analyze delivery issues, predict capacity limits, and monitor user experience. Many teams face operational pressures that slow modernization.
Security tops the list of challenges, followed by reliance on other IT teams, managing multiple internet providers, and budget limits.
Silk explained that networks now support all digital initiatives. “As AI adoption grows, companies can’t afford fragmented visibility or reactive fixes. Success requires intelligent observability that covers every user, application, cloud platform, and network path.”
The move to distributed workforces and cloud services has stretched networks beyond traditional data centers, creating new gaps.
Related: New camera does it all on the go
Companies with strong visibility tools report faster problem-solving, proactive performance improvements, and less downtime. These advantages matter more as AI workloads need steady, low-latency connections.
For many, the issue isn’t just technical. Network teams often work in isolation with tight budgets and outdated tools. Breaking these barriers will require investment in both technology and training.
Silk noted that AI readiness depends on more than computing power or data quality. “Network performance, visibility, and operational maturity are just as important for reliable AI services,” he said. Companies that ignore these gaps may fall behind in an AI-driven market.
The findings arrive as businesses prepare for a wave of AI adoption in the next two years. Those investing in network observability and automation will likely handle AI demands better, while others may face ongoing performance problems that derail digital transformation.
