Building Resilience
Securing the Networks Powering Critical AI Infrastructure
Artificial intelligence (AI) is changing what communications networks must support. From the data center to the edge, AI enables a broad range of applications, including financial services, energy grids, and self-driving vehicles. These complex workloads depend on secure, reliable, low-latency connectivity, as well as hardware, software, and supply chain integrity.
For cable operators, this shift introduces new requirements across access, aggregation, core, and data center networks. It also expands the definition of resilience. Network performance still matters, yet performance alone isn’t enough. Operators must also verify that the components, software, suppliers, and partners supporting AI-driven services can be trusted. This requires verifiable supply chain assurance, not just perimeter defense.
Achieving that level of trust grows more difficult as AI infrastructure adds layers of hardware, software, suppliers, and automation. Hyperscale data centers with dense deployments of GPUs, TPUs, and other specialized AI accelerators increase demand for capacity, throughput, and orchestration. At the same time, AI training and inference workloads are driving network operations toward automated control. These requirements extend to 5G, future 6G, and optical transport networks, which are becoming more software-defined.
AI-enabled controllers can now manage load balancing, network slicing, anomaly detection, and resource allocation. They dynamically optimize traffic flows, reconfigure network resources, and respond autonomously to outages or cyber anomalies. Yet automation creates new dependencies on the integrity of underlying hardware, software, and training data. A flaw in any of those areas can affect performance, security, or service reliability.
Specific risks include data poisoning, model provenance failures, compromised development pipelines, and malicious software packages. For operators, even minor integrity gaps can cause outages, service degradation, missed service-level agreements, or loss of customer trust.
These risks are no longer limited to traditional IT environments. A compromised software library, supplier, or AI model can cascade across operators, networks, and critical services. State-sponsored actors have demonstrated that these threats are operational, not theoretical. U.S. intelligence agencies, for example, have identified foreign-backed groups pre-positioning within communications infrastructure.
As networks increasingly support robotics coordination, machine vision, autonomous vehicles, and surveillance systems, supply chain compromises can disrupt critical operations, economic activity, and national security. To manage this risk, operators need verifiable assurance that every system behind the network remains trustworthy. Assurance must extend to hardware, software, development pipelines, managed service providers (MSPs), open-source dependencies, and third-party partners.
However, self-attestation is no longer sufficient. Communications providers, enterprises, and government agencies need frameworks that verify protections throughout the supply chain. These frameworks should cover supplier vetting, component traceability, software development practices, software integrity, incident response governance, and ongoing risk management.
Industry-led standards provide a practical path forward. TIA’s SCS 9001 establishes verifiable requirements for ICT supply chain security, helping organizations address risk across suppliers, development processes, logistics, and governance. Certification also enables communications providers to demonstrate supply chain integrity to government customers, enterprise clients, and regulators.
Ultimately, networks powering AI-driven services must prove resilience across architecture, operations, and supplier relationships. Standards-based assurance provides a way to demonstrate that readiness in real-world operating environments, ensuring that trust is engineered into critical infrastructure and verified over time.
Scaling AI infrastructure with greater control requires continued industry collaboration. The referenced white paper expands on these challenges and outlines practical approaches for managing AI-related risk across the ICT ecosystem. (Download white paper: https://tinyurl.com/35mv7paj). Organizations interested in contributing to this effort can also engage with TIA to help shape the next phase of AI risk management for the ICT industry. For additional information, please contact membership@tiaonline.org.
Mike Regan
VP, QuEST Forum, TIA
Mike Regan leads the Telecommunications Industry Association’s (TIA) QuEST Forum, advancing business performance improvement standards for the ICT industry. He brings more than 30 years of experience as a senior engineering leader, having led development of complex communications and networking products deployed across global service providers, cloud platforms, and enterprises. Mike works closely with industry, operators, and government agencies to drive standards adoption focused on product quality, cybersecurity, and supply chain security.
Images, Shutterstock
