AI at the edge: the IoT is moving to local decision-making

The rapid development of AI tech is changing what's possible in the consumer and industrial IoT. Devices are no longer "data collectors" but decision-making agents. So, there are important questions to be answered around where these systems are deployed - and where the data is processed.

Across CES announcements and keynotes, it became clear that there's a move in favour of edge-native intelligence. This is driven by factors such as:

• Latency constraints
• Bandwidth costs
• Reliability requirements
• Data sovereignty and privacy expectations

Underpinning this change is the huge volume of data generated by sensors in AI-powered IoT applications. Edge (on device) computing reduces the need to transmit this data to remote servers. This reduces latency and conserves bandwidth.

But AI systems are power and compute-hungry. In response to this challenge, chip makers have launched Edge AI processors that emphasise battery efficiency and are designed for applications in constrained environments.

So how might these factors play out in a real-world scenario? Consider the smart temperature sensor in a warehouse setting. It can detect potentially ruinous safety issues in real-time and resolve it without sending data to the cloud. Local IoT processing is also helping to accelerate predictive maintenance and the development of autonomous robotics.

Overall, CES 2026 revealed how on-device intelligence is becoming a baseline expectation for IoT stakeholders. However, it is not all upside. Local processing opens up new questions around data leakage and security. As intelligence moves closer to the edge, the need for reliable, secure connectivity becomes even more critical.

Security and regulation: not burdens, but business enablers

CES 2026 underscored that connectivity is the 'new infrastructure'. The combination of local processing, dedicated IoT chipsets, improved battery life and ubiquitous connectivity tech is changing the connectivity game for enterprises.

Global reach and seamless roaming are opening up mission-critical IoT use cases. The challenge is to manage massive device fleets across multiple geographies - and the security issues that come with this.

There's also pressure from external parties. Regulators are acting. A key example? The EU Cyber Resilience Act (expected in 2027). This is landmark legislation and, on stage at CES, experts argued that it will most likely set It establishes mandatory cybersecurity standards for connected products. that the rest of the world will follow.

Growing regulatory expectations reinforce what many in the IoT ecosystem already know: security is no longer optional. But regulation should be an incentive to innovate, not a burden. Providers must focus on the fundamentals: secure device identity, robust authentication and the encryption of data at rest and in transit. And they should apply this across the device lifecycle from factory to field.

The good news is that great cybersecurity can serve as a true market differentiator. More and more IoT customers are asking not just what a device can do, but whether it can be trusted. And can this trust be scaled as IoT deployments grow to millions?

Happily, IoT providers are innovating to support manufacturers. An example from CES 2026 is the new post-quantum-ready security chip from Samsung Electronics. It embeds Thales' secure operating system and represents a step forward in protecting connected devices against today's cyberattacks and tomorrow's quantum-era threats.

Take away: CES 2026 signals progress towards a more mature IoT era

The noise from CES suggests new approaches to local processing, device/data security and lower-power connectivity technologies are helping to birth a new era for the IoT. The forecasts back this up. A recent report estimates the number of connected IoT devices will reach 39 billion in 2030, and 50 billion by 2035.

As the IoT moves from experimentation to operational scale, the winners will be those who embrace intelligence, trust and efficiency - and see regulation as a motivator. The next phase of connected innovation will be defined not by how many devices we launch, but by how intelligently, securely and sustainably we deploy them.