Check out our latest blog article: From component to enterprise – modular robotics done right.
Lights Out: Why Fully Automated Factories Are Closer Than You Think
Ben Tan is a Product and Business Development leader with a track record of launching new consumer and enterprise products and services in a world of choices, identifying growth markets, and forging alliances that drive broad adoption. He currently serves as Product Management Director, Industrial and Supply Chain AI at Intel, where he leads a product-led growth strategy through the application of easy-to-use, deployable AI/ML solutions, addressing the needs of both end users and developers. Across his 8+ years at Intel, Ben has also served as Market Development Director for Health and Life Sciences, developing new partnerships and go-to-market routes in healthcare for the rapidly growing remote patient monitoring segment. He builds teams around the belief that measurable small wins strengthen the fabric of development toward the long-term goal, and that success depends on crafting win-win strategies for internal and external partners, an approach that has helped him grow beachheads into sustaining, double-digit-growth businesses, always starting with a deep understanding of user experience and the key value exchange points.
Stay Connected:
- https://www.linkedin.com/in/ben-tan-3907
- https://www.linkedin.com/company/intel-corporation/home
- https://www.linkedin.com/in/techris
Stay inspired and ahead of the curve by subscribing to Forging the Future. Share your thoughts on this episode with the hashtag #ForgingTheFuture or tag us online!
🎧 Episode Highlights
[00:02:32]: Why edge AI is having a moment — sovereignty, tokenomics, and latency
[00:05:08]: What a “lights-out” factory actually looks like, and how close early pilots already are
[00:08:33]: The biggest mistakes organizations make when starting an AI initiative
[00:16:00]: Reframing performance vs. cost — and why “good enough” often wins
[00:19:18]: The shift to smaller, domain-specific models and performance-per-watt
[00:26:51]: Unexpected generative AI use cases on the factory floor, from PLC code to synthetic defects
[00:35:15]: A first look at SuperClaw, Intel’s hybrid edge-to-cloud token strategy
[00:38:04]: Why defining success for the business — not the technology — is what most people miss
🔑 Key Takeaways:
- Edge AI is being pulled forward by more than latency. Data sovereignty concerns and the economics of cloud tokens (“tokenomics”) are now just as influential as real-time performance needs in pushing workloads back to the edge, especially in manufacturing and industrial settings.
- The biggest AI failures are organizational, not technical. Top-down rollouts that leave factory-floor workers out of the loop, ungoverned pilot sprawl across departments, and unresolved friction between IT and OT security policies are far more likely to sink an AI initiative than the technology itself.
- Bigger isn’t always better. The industry is shifting from massive, general-purpose frontier models toward smaller, domain-specific models — Ben calls it “big brain to small brain” — that deliver higher performance-per-watt and only need to know the one job they’re built to do.
You may be interested in:
