The promise of artificial intelligence (AI) has catalysed a massive wave of infrastructure investment across Asia, with ICT spending in the Asia Pacific (Apac) region projected to reach USD $1.4 trillion in 2025. However, this rising tide is not lifting all boats.
Despite the market buzz, individual chief information officer (CIO) budgets are not immune to macroeconomic headwinds. As we head into 2026, tech leaders are facing rising mandates to deliver AI-ready infrastructure while meeting tightening data sovereignty requirements, often while operating under flat (or shrinking) budgets. As noted by Forrester, this uncertainty is forcing a shift from rapid expansion to high-performance IT. Leaders must now make tough decisions about where to invest, where to cut, and where to experiment.
The only way to move forward is not to spend more, but to spend smarter. To fund the resilience required for this new era, CIOs must divest from legacy cost centres and pivot toward the following three strategic pillars: Containment, Autonomous Management, and Investment Protection.
From perimeter defence to containment
The traditional "castle-and-moat" security model is no longer fit for purpose. The proliferation of unmanaged Internet of Things (IoT) devices and the complexity of diverse vendor ecosystems have pushed the potential attack surface far beyond what a firewall can protect.
Furthermore, malicious actors are no longer just attacking systems; they are attacking people. Singapore recently recorded a staggering 1,500% surge in deepfake fraud cases, proving that human employees, as well as the passwords they hold, are being outmatched by AI-driven social engineering.
In 2026, capital must shift from the perimeter to network containment built on Zero Trust principles. Rather than focusing solely on keeping intruders out, CIOs should invest in architectures that automatically segment devices. By assuming every connected camera, sensor, or handheld is a potential entry point, organisations can "contain" a breach at its source, limiting the "blast radius" before it touches the core.
Crucially, this shift must be governed by a formal AI Security Policy (AISP). Without a framework governing how data is managed by both internal and external AI applications, companies risk "Shadow AI" leaks, where employees inadvertently feed proprietary data into public models.
From static energy to autonomous management
While global headlines focus on the massive power consumption of AI data centres, equal attention should be paid to the enterprise edge. As organisations deploy thousands of AI-enabled devices such as 4K CCTV arrays and Wi-Fi 7 access points, the energy footprint of the physical campus is exploding.
Unlike data centers which are highly optimised for efficiency, enterprise buildings are often plagued by "zombie infrastructure." These are switches and access points running at full throttle in empty boardrooms or guest-less hotel wings at 3:00 AM.
With Singapore expanding its mandatory carbon reporting regulations (Scope 1 and 2), sustainability is no longer a "nice-to-have" CSR metric, having become a regulatory requirement. Consequently, CIOs can no longer afford "static" infrastructure that burns the same energy regardless of demand.
The solution is a move to autonomous energy management. Modern intelligent network fabrics now function as "active energy brokers," using real-time traffic data and presence sensors to adjust power draw. Imagine a network that automatically throttles Power over Ethernet (PoE) to a camera when no motion is detected or puts IP phones into deep sleep after business hours. By treating the network as a dynamic system rather than a fixed utility, CIOs can shave watts off thousands of devices, creating a massive multiplier effect on energy savings and compliance.
Move from “rip-and-replace" and embrace investment protection
This shift toward intelligent, software-driven management allows for a fundamental rethink of the hardware itself. For decades, the industry has been trapped in a three-to-five-year hardware refresh cycle, which is not only capital-intensive, but often an exercise that lacks a clear ROI. Meanwhile, replacing core switching infrastructure on an arbitrary schedule is not just operationally disruptive; it inflates the Total Cost of Ownership (TCO) without guaranteeing performance gains.
In 2026, the strategic move is to divest from the refresh cycle and embrace investment protection. By selecting ruggedised hardware designed for 7-to-10-year lifecycles, supported by modular software-defined upgrades, enterprises can maintain peak operational capability without the constant capital outlay. This approach does more than just save money; it aligns with corporate ESG goals by reducing electronic waste. More importantly, it frees up capital for the software-led innovation and agentic AI capabilities that drive actual business value.
The success of a CIO in 2026 will not be measured by the size of their technology stack, but by how intelligently that infrastructure absorbs pressure. By moving away from legacy security models, non-productive energy consumption, and the "disposable" hardware mindset, technology leaders can free up the capital necessary to build the next generation of resilient, AI-ready connectivity. In an environment of constrained budgets, the discipline of strategic reallocation is the only sustainable path to innovation.
Kenny Ng is the head of Network Business Division for Asia Pacific at Alcatel-Lucent Enterprise
