The keynote, delivered by Struan Buchanan (Partner, Power, Utilities & Renewables Consulting, Deloitte), framed AI less as a technology story and more as a strategic and structural one. His central argument: the constraint on AI value isn't the tools — it's execution. Organisations capturing real value aren't the ones with the best models, but the ones that have fundamentally changed how they operate.

A key market signal he raised is the collision between AI and electricity demand. Data centres are becoming one of the fastest-growing sources of new load across APAC, with investment potentially reaching hundreds of billions of dollars by 2030 and electricity demand from digital infrastructure multiplying several-fold within a decade. This is reshaping capital allocation, grid planning, and where infrastructure investors choose to deploy capital — energy and AI strategy, he argued, can no longer be planned separately. At the same time, AI-enabled efficiency gains across the power value chain represent a meaningful counterbalance, meaning the AI-energy relationship is best understood as circular rather than purely additive demand.

Buchanan's broader thesis was that most organisations are stuck in the "adoption gap" — access to tools is not the problem; integration into how work actually gets done is. He positioned scaling as requiring commitment across leadership accountability, disciplined value focus, trust, redesigned operating models, resilient technology foundations, and genuine workforce readiness — with leadership conviction as the single largest differentiator between companies that scale and those stuck piloting indefinitely.

The panel discussion, moderated by Buchanan, brought together perspectives from across Deloitte and Databricks, and the conversation stayed largely strategic rather than technical:

Jonathan McCormick (Data & AI Partner, Deloitte) reinforced that the missing piece in most AI programmes is rarely the model itself — it's the surrounding machinery of trust, governance, and change management. He framed AI adoption as fundamentally a leadership and transformation challenge rather than a technology rollout.

David Owen (Cyber Partner, Deloitte) brought a risk and geopolitics lens, raising an increasingly board-level question: what happens if access to a foreign-hosted model is withdrawn, or if the integrity of external data feeding critical decisions is compromised? He linked this directly to the rising strategic importance of sovereign AI and on-premise control as insurance against external dependency, particularly given shifting geopolitical dynamics and the cost differentials emerging between AI providers of different national origin.

Kristy Delaney (Human Capital Partner, Deloitte) argued that AI scaling is, at its core, a workforce transformation. The real gap isn't tool access — most employees have that — but genuine AI fluency: knowing when to trust AI, when to override it, and how decision rights between humans and AI should be explicitly designed rather than left ambiguous. She noted that without intentional organisational redesign, adoption stays shallow regardless of how sophisticated the technology becomes.

Kelly Savannah (Databricks) added an economics dimension, noting that AI costs scale non-linearly and are becoming a genuine CFO-level agenda item — the right question isn't "what did the pilot cost," but what the full economics of an AI-enabled process look like at scale.

Closing the session, panellists converged on a consistent message: over the next 12 months, the winning capability won't be a specific tool, but organisational — grounded data foundations, clear governance, workforce fluency, and the ability to connect AI initiatives to a coherent, business-owned transformation narrative. AI, they agreed, is not a technology problem. It's an execution problem.