NetOps teams struggle with AI readiness

As more enterprise organizations adopt artificial intelligence, research reveals a growing gap between AI plans and network preparedness, according to Broadcom.

While 99% of organizations have cloud strategies and are adopting AI, just 49% believe their network could support the bandwidth and low latency that AI requires, according to Broadcom’s 2026 State of Network Operations report. In addition, 95% of 1,350 IT professionals who were polled for the research said they lack visibility into network segments, particularly public cloud environments. The same 95% reported that they need more information from their ISPs.

“Everyone feels they have all the insights they need from any device on any protocol—only inside their data center. They can see everything, but they are not running everything only through a corporate data center; it’s all public networks,” says Jeremy Rossbach, chief technical evangelist of network observability at Broadcom. “This trend of expanded visibility outside of the data center is really catching on, and that really just means collecting more data. The more data you have, the better your network operations are going to be.”

Some 87% of respondents indicated that internet and cloud environments are creating network blind spots in many areas. Half of organizations reported a lack of adequate insight into public clouds, 44% of respondents indicated transit and peering networks created blind spots, and 43% said remote work environments lack visibility. Other areas that create visibility issues include: private cloud (39%); ISP last-mile (38%), and on-premises work (30%).

The visibility challenges extend to ISP relationships as well. Only 5% of network teams say they receive all the information they need from their ISPs. The vast majority want access to data on DNS issues (50%), path latency (49%), historical performance by path (45%), DDoS attack locations (45%), node or hop issues (44%), and packet loss (43%). This lack of visibility matters because it directly impacts operational efficiency. Survey respondents said visibility is essential for analyzing network delivery issues (56%), understanding the impact of route changes (49%), predicting capacity issues (48%), and tracking user experience (48%).

“If Google Cloud is hosting a bit of your AI initiative, whatever that is, and you can’t see into their network, how can you triage that network path for anyone that’s complaining? There are no excuses for black holes like that anymore,” Rossbach explains.

The visibility challenges also impact network operations. Respondents said that visibility is essential for analyzing delivery issues (56%), understanding the impact of route changes (49%), predicting capacity issues (48%), predicting capacity issues (48%), and tracking using experience (48%).

The disconnect between AI adoption plans and network readiness is creating network operational challenges. Survey respondents shared that several factors impact their organizations’ success with AI:

• Network congestion: 46%
• Insufficient visibility (monitoring, observability): 39%
• Congested traffic flows: 388%
• Latency: 37%
• Unpredictable traffic bursts: 34%
• Low capacity (bandwidth): 32%
• Packet loss: 26%
• Low reliability (uptime): 24%
• No network challenges impacting AI: 6%

“The way to approach AI for network operations is to not put the cart before the horse, which is everyone saying, ‘Let’s adopt AI.’ But wait, I only have three wheels on the cart. I have to add the fourth wheel first.’ That fourth wheel is maturing your network observability practice,” Rossbach says.

The State of Network Operations, 2026: AI and its Effect on Enterprise NetOps was sponsored by Broadcom and conducted by Dimensional Research.