AI Workload Surge Drives On-Prem Data Center Expansion

Data centers were once expected to shrink in size and number due to virtualization and cloud services. However, the rise of artificial intelligence (AI) has reversed this trend, with companies expanding their on-premises data centers to power AI workloads.

Cisco executive Ammar Maraqa notes that enterprises are increasingly looking inward for security and cost reasons, and to maintain control over their proprietary data. This shift towards on-premises infrastructure is also driven by the need for sovereignty and data governance.

The demand for expertise in 'deep-tech' fields such as hardware, networking, and cybersecurity is growing rapidly. According to Cisco's research, AI agents require significantly more load on the network than human workloads, making modernized infrastructure essential. The shortage of professionals with deep-tech skills is a major concern, with Maraqa warning that it may be difficult to meet demand.

The need for specialized security and observability in data centers has increased due to the presence of AI agents. Enterprises must monitor the behavior of these agents and test them against guardrails set up by the company. Outdated infrastructure poses a significant security risk, with Maraqa stating that some infrastructure is so old it cannot even be patched.