Xiaochao Pu

Title of the Talk: AI Total Cost of Ownership, an Unsolved Problem

Abstract :
Most organizations still evaluate AI investments using cost models built for traditional software — and those models are breaking down. AI’s true cost structure is harder to pin down: retraining cadence, model drift, data pipeline maintenance, human-in-the-loop review, and the compounding integration debt of stitching AI into legacy systems rarely show up in a business case, yet they routinely dwarf the initial deployment cost. This talk unpacks why AI TCO is structurally harder to model than conventional IT cost, drawing on real enterprise deployments across energy, healthcare, and professional services where hidden costs derailed otherwise sound AI initiatives. The speaker will introduce an early-stage TCO governance framework with a working structure for naming and tracking these cost categories systematically.

Bio :
Xiaochao Pu specializes in enterprise technology transformation and AI-driven systems development. His work focuses on navigating the complexities of AI integration within large-scale infrastructure and spearheading organizational technology shifts. A mentor in the tech ecosystem, Xiaochao coached multiple startups on technical product roadmaps and sustainable business modeling. He earned his MBA from the University of Michigan, focusing on the synergy between emerging technology and corporate strategy.