The rapid adoption of artificial intelligence across enterprises is placing growing pressure on existing IT infrastructure, with legacy systems increasingly struggling to meet demands related to performance, resilience, and security, according to new research released by cloud security firm Netskope. The findings highlight how AI-driven workloads are reshaping enterprise technology environments and exposing gaps in traditional infrastructure approaches.
The research indicates that enterprises are experiencing a sharp rise in network traffic, data movement, and application usage as AI tools become embedded across business operations. From generative AI platforms and machine learning applications to AI-powered analytics and automation tools, organisations are relying more heavily on cloud services and internet-based resources. This shift is pushing IT teams to reassess whether legacy architectures can continue to support evolving workloads without compromising user experience or security.
According to the study, AI usage has significantly increased demand for high-performance connectivity, secure data access, and real-time processing. Many enterprises continue to operate on infrastructure models designed for earlier generations of enterprise applications, which were largely static and hosted within controlled environments. As AI workloads rely on distributed cloud platforms and frequent data exchange, these older systems are proving inadequate for modern requirements.
The report notes that performance challenges are among the most immediate concerns. AI-powered applications often require low latency and consistent bandwidth to function effectively, particularly when employees rely on them for real-time decision-making or customer interactions. Legacy networks that route traffic through centralised data centres can introduce delays and congestion, resulting in slower application performance and reduced productivity.
Resilience is another area under strain as AI adoption accelerates. The research points out that many traditional IT architectures lack the flexibility to adapt quickly to spikes in demand caused by increased AI usage. Outages, service disruptions, and bottlenecks can have a wider impact when AI tools are integrated into critical business processes. This raises concerns about business continuity, especially for organisations operating across multiple geographies and time zones.
Security risks have also grown as AI applications expand the enterprise attack surface. The study highlights that AI tools frequently access sensitive data and interact with external cloud services, increasing the potential for data exposure if controls are not properly implemented. Legacy security models that rely on perimeter-based defences struggle to provide adequate protection in environments where users, devices, and applications operate beyond traditional network boundaries.
Netskope’s findings suggest that many enterprises are responding by accelerating their transition toward cloud-native and zero trust security models. These approaches focus on securing access based on user identity, device posture, and contextual signals rather than relying solely on network location. By doing so, organisations aim to improve visibility into AI-related traffic while enforcing consistent security policies across cloud and internet-based services.
The research also highlights growing complexity for IT and security teams tasked with managing AI adoption. As employees increasingly use multiple AI platforms, both sanctioned and unsanctioned, maintaining oversight becomes more challenging. The report notes that shadow AI usage is emerging as a concern, with employees experimenting with generative AI tools outside formal governance frameworks. This can lead to unintentional data sharing and compliance risks if not addressed proactively.
To mitigate these challenges, the study emphasises the importance of modernising network and security infrastructure in parallel with AI deployment. This includes adopting secure access service edge architectures, enhancing observability into cloud traffic, and implementing controls that can adapt dynamically as usage patterns change. Organisations that delay infrastructure upgrades risk facing performance degradation and increased exposure to security threats as AI usage continues to scale.
The findings also reflect a broader shift in how enterprises view IT infrastructure investment. Rather than treating infrastructure as a background function, organisations are increasingly recognising it as a strategic enabler for AI-driven transformation. Reliable connectivity, secure data access, and scalable platforms are becoming foundational requirements for leveraging AI effectively across business units.
Industry observers note that the challenges identified in the research are not limited to large enterprises. Mid-sized organisations are also grappling with similar pressures as they adopt AI tools without the benefit of extensive IT resources. For these organisations, choosing flexible and cloud-first solutions can help balance innovation with operational stability.
The report concludes that AI adoption is accelerating faster than many enterprise infrastructure environments can adapt. While AI offers significant opportunities for efficiency, insight, and automation, realising these benefits requires a corresponding evolution in IT and security strategies. Enterprises that invest in modern, resilient, and secure infrastructure are better positioned to support AI-driven growth without compromising performance or trust.
As AI continues to reshape enterprise workflows, the research underscores the need for alignment between technology ambition and infrastructure readiness. The gap between legacy systems and modern AI demands is becoming increasingly visible, prompting organisations to reassess long-standing approaches to networking, security, and performance management. How quickly enterprises respond to these challenges may determine their ability to compete effectively in an AI-driven business landscape.