
The episode discusses practical AI use cases in Australian manufacturing and the hardware considerations for scaling from pilot to production.
Four real AI use cases in Australian manufacturing and the hardware decisions behind moving from pilot to production. In this episode, Troy Smithells from Xenon Systems (Australia's only Nvidia Elite partner) walks through where AI is actually earning its keep in Aussie manufacturing, and why production workloads often need to leave the cloud. We cover: Hand-drawn to CAD file conversion: 4x faster, near-100% accuracy, $750K per year saved at a steel fabricator Physical AI in pipe inspection: tiny robots mapping up to a kilometre of pipe for leak detection and preventative maintenance Predictive maintenance from sound: a listening device on a conveyor belt catching wear before failure Computer vision for construction OHS: PPE checks, exclusion zone alerts, and yes, crocodile alerts on the docks at Cairns Why cloud AI costs scale exponentially and on-prem or hosted GPU hardware can land at roughly a quarter to half the spend Data sovereignty, the US Cloud Act, and the space, power and cooling reality of running GPU clusters (132kg, 110dB, and a burn-in you need earplugs for) When to bring in infrastructure specialists: after the proof of concept, before you try to scale Whether…
Hosts: Shane Williams, Paul Mason
Guest: Troy Smithells
Organizations: Xenon Systems, Nvidia
Places: Australia, Cairns, Australia
Explore listener stats, chart rankings, contacts and more on the Manufacturing Tech Australia podcast page.