Why Spending $1 Trillion on AI Is Rational

In 2026, the AI narrative has shifted from basic prompt-and-response to Agentic AI, multi-step, autonomous systems that are finally narrowing the gap between technological promise and financial reality. Duncan Stewart, Director of TMT Research, Deloitte Canada, explains why the global CapEx surge to $1 trillion is a rational enterprise play, driven by "token maxing" and the race for sovereign compute.From the "Search Cliff" threatening web referral traffic to the "RAMageddon" memory shortage fueled by a Persian Gulf helium crisis, we dive into the high-stakes forces reshaping the regional and global tech landscape.Tune In To Find Out:The Narrowing Gap: Why the investment-to-revenue ratio is dropping from 100x to 10x as AI monetisation hits the $100 billion mark.Agentic vs. Generative: The move from simple prompts to autonomous, multi-modal code-writing systems for enterprise optimisation.Token Maxing: Why software engineers are now burning $250,000 a year in tokens to drive a 5x to 10x productivity gain.The Search Cliff: Why media and retail sites are facing a 50% to 80% sudden drop in referral traffic as AI overviews take over search.Sovereign AI: Why nations are spending $100 billion on local data centers to prevent total dependency on US and Chinese infrastructure.RAMageddon: Why memory prices have spiked 400% and why the shortage could last up to five years.The Helium Factor: How the conflict-driven force majeure on Persian Gulf helium is threatening Asian chip and memory manufacturing.Jevons Paradox: Why making AI tokens cheaper will actually lead to an explosion in consumption, not a reduction.Tech Talk is brought to you by Maxis Business. Building Tomorrow's Business Today.See omnystudio.com/listener for privacy information.