AI activity is pushing Gulf networks toward breaking point
86% of Saudi firms expect campus or branch capacity limits within two years
#UAE #SaudiArabia #networks — New research from US networking giant Cisco, conducted with research firm Foundry across 200 IT leaders in the UAE and 200 in Saudi Arabia, finds that enterprise networks in both countries may be on a collision course with AI demand. In Saudi Arabia, 86 percent of organisations have already hit or expect to hit campus and branch network capacity limits within 24 months. In the UAE, 81 percent expect to reach that point within three years. Both countries expect AI-driven network traffic to more than triple over the same period, with agentic AI carrying the biggest load.
SO WHAT? — The Gulf is adopting agentic AI faster than almost anywhere else. Saudi Arabia already has more than 40 percent of surveyed organisations running enterprise-wide AI agent deployments, ahead of the global average of 33 percent, and 99 percent of UAE respondents expect agentic AI use to grow within 24 months. The problem is that AI agents don’t behave like human users. They fire dozens of API calls and database queries per second, generating dense traffic that legacy networks weren’t built to carry. The infrastructure bill for that ambition is coming due faster than most IT budgets are prepared for.
KEY POINTS:
Cisco and Foundry surveyed 200 IT leaders in the UAE and 200 in Saudi Arabia, about the accelerating impact of AI on campus and branch networks. The survey found that AI-driven traffic is expected to more than triple in both countries within three years. The data is an extract from a global survey of 3,472 IT leaders across Asia-Pacific, Europe, the Middle East, Latin America, and North America.
In Saudi Arabia, 86% of organisations have already hit or expect to hit network capacity limits within 24 months, compared to a global average of 73 percent, with agentic AI workloads projected to drive a 116 percent increase in traffic alone.
In the UAE, 81% of organisations expect to reach campus and branch capacity limits within three years, with agentic AI workloads expected to drive a 126 percent traffic increase, the highest of any single workload category.
More than 40% of Saudi organisations already have broad, enterprise-wide AI agent deployments, ahead of the global average of 33 percent, while 34 percent of UAE organisations report the same.
Wi-Fi is emerging as a primary bottleneck in both markets, cited by 46% of Saudi respondents and 50% of UAE respondents as the area driving the greatest increase in capacity requirements.
Security is deteriorating alongside capacity pressure: 93% of Saudi respondents and 91% of UAE respondents say they’re struggling to keep up with AI-related security challenges, and 87% and 89% respectively say AI is already causing damage.
Visibility is a growing problem too, with 58% of Saudi and 54% of UAE respondents saying they lack adequate visibility into AI-related traffic flows across their networks.
Budget is constraining the response: 42% of UAE IT leaders say budget limits their ability to modernise to a great extent, above the global average of 31% , while 34% of Saudi respondents say the same.
Despite the pressure, a confidence gap persists: 74% of Saudi IT leaders and 76% of UAE IT leaders say they’re more confident in their AI strategy than in their network’s ability to deliver it.
ZOOM OUT — Cisco and Foundry surveyed more than 3,400 IT decision-makers across 15 countries and found that campus and branch network traffic tied to AI workloads has already grown 36 percent in the past 12 months, with another 96 percent growth expected within the next year alone. Over three years, traffic is projected to reach three times current levels. Only 15 percent of organisations globally said their networks are flexible enough to support AI at the necessary scale, according to Cisco's own AI Readiness Index, and 73 percent already face or expect to face campus and branch capacity limits within 24 months.
[Written and edited with the assistance of AI]
Source: Cisco