AI adoption is rising fast but turning those early successes into something that runs at scale is proving far more difficult.

On Industry Insight, Lynlee Foo speaks to Mingcheng Lim, Country Head, Singapore, ST Telemedia Global Data Centres about why many AI projects stall when they move from pilots into real-world deployment.

While most organisations have already started their AI journey, far fewer have been able to scale it in a way that delivers consistent value. From infrastructure limits and rising compute demands to less visible constraints like power, cooling, and operational complexity, the conversation explores what actually gets in the way — and what it takes to move beyond experimentation