As Forrester's Top 10 Trends in Edge Computing and IoT, 2025 report confirms, edge and IoT technologies have transitioned from experimental pilots to strategic imperatives.
For operational technology (OT) leaders across Southeast Asia—where manufacturing contributes over 20% of GDP in economies like Vietnam and Thailand—the challenge is no longer whether to deploy edge intelligence, but how to prove its value, secure its sprawl, and embed it sustainably within existing frameworks.
Measuring what matters: The three metric sets that prove edge value
Operational leaders must move beyond vanity metrics to demonstrate tangible business impact. According to Charlie Dai, VP and principal analyst at Forrester, success hinges on tracking three AI-infused metric sets. This begins with measuring outcome KPIs, such as unplanned downtime reduction, first-time fix rate, and predictive maintenance model precision/recall, to assess business impact.
"They should track edge AI performance KPIs, including latency to decision, on-device inference success rates, drift alarms, and site-level throughput to verify real-time capability," he continues.
They should instrument security and resilience KPIs such as asset discovery coverage, Zero-Trust policy conformance, segmentation violations, and mean time to detect and isolate using anomaly detection at the edge."
This tripartite approach aligns with New Relic's 2023 State of Observability report, which found that 65% of manufacturers leveraging observability achieved faster issue resolution by correlating operational metrics with business outcomes.
Operationalising AI at the edge across Asian enterprises
Asia's diverse industrial landscape—from Singapore's smart factories to Indonesia's sprawling logistics hubs—demands contextual deployment strategies. Dai emphasises a structured approach: "Organisations in Asia should operationalise AI at the edge by mapping use cases to the right provider, enterprise, operations, or engagement edge; deploying compact models on gateways or PLCs; and orchestrating placement via edge platforms for workload affinity." Charlie Dai
He adds that: "They should fuse SASE, NAC, and Zero-Trust controls with AI-assisted policy verification, and embed observability that correlates latency, error budgets, and sensor anomalies with model outputs. They should sustain outcomes by co-funding with P&L owners, cross-training IT/OT on MLOps at the edge, and governing vendor integrations end-to-end."
This methodology addresses a critical gap: IDC's Asia/Pacific Security Study (2024) reveals that 68% of regional enterprises struggle to extend zero-trust principles beyond corporate networks to distributed OT environments.
"Enterprises are moving away from legacy transport-layer architectures toward cloud-native, AI-augmented models that embed zero trust, contextual enforcement, and analytics at the core of digital operations," says Sakshi Grover, senior research manager, cybersecurity products and services, IDC Asia/Pacific.
She goes on to add that as infrastructure becomes more ephemeral and user boundaries dissolve, policy-driven convergence frameworks can deliver scalable, intelligent defence across hybrid networks.
Confronting IoT's security triad: Sprawl, controls, and connectivity
Three security challenges consistently undermine IoT deployments across Asian supply chains. Dai identifies them precisely: "The three biggest IoT security challenges are unmanaged asset sprawl, weak LAN-edge controls, and immature multi-bearer connectivity."
He suggests that "asset sprawl should be mitigated with lifecycle discovery, risk scoring, and model-based device classification integrated with enterprise controls. Weak LAN-edge controls should be addressed by extending Zero-Trust with SASE and NAC, plus AI-driven segmentation and continuous posture checks across sites."
He warns that immature connectivity should be countered through secure onboarding practices, end-to-end encryption, and AI-based anomaly detection for east-west traffic across private 5G, satellite, mesh, and Wi-Fi networks.
These vulnerabilities are acute in Asia, where IoT device shipments are projected to grow at a 46.1% CAGR through 2030—outpacing global averages.
Industry-tailored observability: From factory floors to retail aisles
Generic dashboards fail in heterogeneous Asian operations. Dai advocates for role-specific intelligence: "Observability should be tailored with AI that converts edge signals into role-ready insights per industry." He cites examples, such as in manufacturing, where organisations should link anomaly scores to overall equipment effectiveness (OEE) and predictive maintenance outcomes.
In healthcare, he recommends organisations enforce latency and error budgets for bedside inference. In other sectors like retail, businesses should monitor mPOS uptime, offline-fallback success, and computer-vision queues; and transport should fuse location, condition telemetry, and throughput forecasts.
"Teams should unify logs, metrics, traces, and risk data into dashboards that trigger AI‑assisted critical‑event workflows for rapid, distributed response," recommends Dai.
This aligns with manufacturing trends in the region: over half of Asia-Pacific manufacturers now prioritise observability to drive AI and IoT adoption, citing cross-team collaboration as a primary benefit.
Edge-enabled sustainability: Cutting emissions at the source
With Scope 3 emissions accounting for up to 90% of corporate footprints in logistics-heavy Asian economies, edge computing offers immediate leverage. Dai notes: "Edge computing can enhance sustainability by using IoT sensors and on-site AI to optimise HVAC, lighting, water usage, and equipment health in plants, buildings, and transit while reducing backhaul energy."
He posits that local inference enables faster interventions, such as leak detection and adaptive set points. "Extending AI telemetry to logistics and fleets helps cut emissions through smarter routing, condition-based maintenance, and waste reduction. Firms should elevate sustainability KPIs alongside uptime and cost and tie incentives to AI-verified savings," concludes Dai.
IoT-enabled carbon tracking is gaining traction; ambient IoT solutions now deliver 30% greater real-time accuracy in emissions monitoring than legacy systems, a critical advantage for ASEAN manufacturers pursuing net-zero commitments.
