TechEx North America: AI Needs Security, Power

TechEx North America opened its doors in Santa Clara on Tuesday, May 19, 2026. It's unexpectedly blunt. AI may dominate boardroom conversations, but its future hinges on three things that have nothing to do with algorithms: power, infrastructure, and security. The event brought together tracks on Edge Computing, IoT, Data Centre Congress, and Cyber Security, and across all of them, speakers kept returning to the same question: what needs to be built around AI before it actually works in the physical world?

The Edge Is a Demanding Place

Operating at the edge isn't glamorous. And Ed Doran of the Edge AI Foundation chaired the track that set the tone early, and it means dealing with latency, deployment discipline, and the messy overlap of industrial systems and IT networks. So sessions dug into scaling edge deployments across multisite businesses, agentic network operations, distributed inference across on-prem, cloud, hybrid setups, and representatives from Akamai, Spectro Cloud, Scylos, TÜV Rheinland, the OPC Foundation, Schneider Electric spoke.

Moving intelligence closer to the machine changes the risk profile. Nobody could agree on which direction that risk moves. Faster local decisions reduce dependence on central cloud services, sure. But they also raise fresh questions about observability. If a machine makes a decision in a factory in real time, who is watching? Who owns the outcome? That tension ran through the entire track.

Escaping Pilot Purgatory

The IoT Tech Expo track tackled manufacturing head on. Smart factory trends, AI beyond Industry 4.0, asset management, and digital twins all got airtime. But one phrase kept surfacing, and it carried weight: pilot purgatory. The gap between a working demo and a working deployment is wide, and several sessions made it clear that many companies are stuck in it.

Rockwell Automation and Ford presented together on physical AI and connected asset intelligence. Their session looked hard at scaling projects that perform beautifully in concept and then meet the real world. Old machines. Legacy software. A factory floor that does not care about your presentation deck. The question they posed was direct: how does intelligence enter daily operations without becoming yet another dashboard that nobody owns?

Digital Twins That Actually Work

Digital twins received a similar reality check. The better version is not a visual replica built for demonstrations, several speakers argued. It is an operational model that helps a factory, a city, or a municipal facility make decisions before something breaks. Pre-testing decisions and improving maintenance are table stakes. The harder question is what a modern digital twin should be designed to achieve beyond that. Siemens, Korea's LG CNS, and Boston Dynamics were among the names linking ideas across the show strands.

Power, Water, and Concrete Limits

Now for the awkward part. The Data Centre Congress track brought the conversation down to earth with a thud. Construction chaos. Power procurement. Cooling. Water. Permits. The keynote speakers and roundtable guests made one thing plain: AI economics change fast, but the infrastructure stack takes years to mature. Those timelines do not align.

Santa Clara, the host city, shared its own data centre journey with early visitors. Water and power constraints, speakers noted, can cut straight through the rhetoric around the scale of AI. The data centre is where AI strategy becomes physical. Not abstract. Not a whitepaper. Concrete and copper and cooling towers. The boardroom's considerations are practical, and the limits are material.

Shadow AI Expands the Attack Surface

The Cyber Security and Cloud Expo track put its own urgent spin on deployment. Day one covered security culture, compliance, ransomware, shadow AI, data exfiltration, legacy systems, open-source dependency issues, and the CISO relationship with the C-suite. A general consensus emerged fast.

Existing security weaknesses don't diminish when the business wants faster, smarter tools.

That message was repeated across sessions. Shadow AI was a standout concern. Staff use AI services inside business workflows, sometimes without approval and usually with no facility for logging their activities. Data governance and cyber governance become the same conversation. And the cybersecurity track's concerns about legacy systems found echoes on the IoT and Edge stages, where modern intelligence meets aging plant equipment.

What the Tracks Shared

The benefits of one conference hosting complementary tracks became obvious in several moments. Here is a short list of the recurring themes that crossed stage boundaries:

• Legacy systems remain the friction point where smart AI meets stubborn reality
• Speed of adoption is frequently the enemy of security
• Unplanned and disorganised AI implementations do not fit the modern enterprise
• Critical infrastructure in transport or energy means cybersecurity cannot be an afterthought

And this is where it gets interesting. The TechEx North America day one tracks that focused on infrastructure gave the entire conference a dose of reality. AI may be discussed in terms of agentic automation and autonomous decision-making, but deployments depend on networks, data centre capacity, and cybersecurity. Edge and IoT sessions showed how carefully intelligence needs to be applied when it reaches actual machines. The data centre sessions exposed the material limits of physical construction. The cybersecurity sessions showed how a desire for speed can create the very vulnerabilities that bring systems down.

TechEx North America drew thousands. Day one wasn't a parade of shiny software; it reminded them that putting AI in production isn't about switching software on, but instead relies on the mundane matters of buildings, grids, networks, and security. Companies that understand those dependencies are more likely to deploy the latest technology successfully. So it's the bigger picture.

The conversations that started in Santa Clara will continue at upcoming TechEx events in Amsterdam, California, and London.