scte PRESENTS
FUTURE OF TECH Panel 1: AI usage in both B2B and B2C
work. Describing work already under way in that area, he said: “We found those problems and it saved us roughly 94% of the cost and time it would have taken to do the exact same work by hand.”
deployment. “None of this works without the plumbing underneath,” he said.
AI success will depend on power and people, not hype Artificial Intelligence may dominate the technology conversation, but speakers on the opening panel at The Future of Tech argued that its future will depend less on software breakthroughs than on the more difficult questions of infrastructure, energy, governance and trust. AI usage in both B2B and B2C brought together perspectives from nuclear energy, telecoms infrastructure, media technology and AI investment. While a speaker from the nuclear industry might feel outside of the SCTE’s normal area of interest, the central point remained the same. AI is no longer just a tool sitting on top of existing systems. It is rapidly becoming part of the underlying infrastructure that industries and public services will rely on.
He argued that while AI platforms can evolve in weeks or months, the networks, energy systems and support infrastructure beneath them operate on far longer timescales. “AI isn’t actually the hard part. Software is really, really easy. The infrastructure that supports it is the limiting factor,” he said. “The lowest common denominator is always the limiting factor.”
Rich Welsh, the Chairman and Co- Founder of Volustor, warned against anthropomorphism and treating AI systems as though they reason like
humans. Welsh said his business wasn’t an AI company more that AI was the tool that facilitates what Volustor does. “People attribute human qualities to AI all the time, especially chatbots, especially agents,” he said. “They’re not people. They are statistical algorithms making decisions.” That distinction matters, he argued, not only for consumer applications but also for enterprise adoption, where businesses may begin to rely too heavily on systems that appear confident and human-like but remain unpredictable at the edge cases. Welsh said that risk becomes particularly serious when organisations start using agentic systems in customer interaction, internal decision-making or sensitive support scenarios. He also warned that intellectual property protection is becoming a much bigger issue as AI systems become more capable of reproducing style, structure and commercially valuable creative assets. “Protecting your IP is your problem — it is not a legal problem,” he said, arguing that companies will increasingly have to think of AI as both “a threat and an asset”.
That infrastructure challenge was picked up by Michael
Drury, who linked the AI boom to the enormous increase in power demand from data centres. Representing
nuclear technology firms Terra Praxis and Lucid Catalyst, Drury set out the demands of data sector amid challenging projections in future decades. “2050 is not that far away, and that is an awful lot of energy to start providing to data centres,” he said. Drury’s central argument was that the debate should move beyond abstract arguments over whether nuclear power is desirable and towards the practical issue of whether it can be deployed in ways that suit the timescales, economics and operating models of fast-growing industries such as data centres. “The real challenge isn’t whether nuclear is good or bad. The real challenge is getting it to align with the type of industry you’re trying to align it to,” he said. The issue is not just generation capacity, but also grid connection, long lead times for transformers and substations, planning and licensing cycles, and the difficulty of matching large, capital- intensive energy projects with fast-moving digital businesses. “It takes nearly 10 years to licence it, so we have to find ways to speed things up,” he said. Drury argued that AI itself could play a role in reducing those delays, particularly in accelerating planning and licensing
“AI is becoming infrastructure – the infrastructure that runs our country,” said Mike Hewitt, founder of Breakwater Technology, arguing
that organisations need to stop treating AI as a bolt-on. “AI cannot be a bolt-on. We have to change the way we run our organisations.” Hewitt said the biggest mistake businesses make is to approach AI as a technology project delegated to a single department, rather than as an operating model question that has to be addressed at leadership level. “If we want to be AI-native organisations over the next five years, you need to step back and ask: how am I going to structure my organisation?” he said. In his view, the issue is not simply deploying tools but deciding what should be automated, what should be redesigned and what processes may no longer make sense at all.
Concern over rights and provenance was echoed by Dr Emma Young, Senior Director of AI Production Technology
at Deluxe. Young described how AI is already being used in media workflows to support time- consuming localisation, mastering and production tasks, and how generative tools are increasingly helping creative teams move faster in pre-production.
Phil Scott of Technetix highlighted the
gap between the speed of software
development and the much slower pace of
physical infrastructure
16
MAY 2026 Volume 48 No.2
Made with FlippingBook - Online magazine maker