HS128: Planning for an AI Bubble Burst

In this episode of Heavy Strategy, Jonna Johnson and John Burke of Numerties explore the pressing question: is AI a bubble, and if so, how should enterprise IT leaders prepare? Drawing parallels to the dot-com boom, they argue that while AI is a transformative capability, the current investment frenzy—marked by circular funding, massive venture capital commitments, and unprofitable AI ventures—signals a classic bubble. They highlight that most AI value today comes not from selling AI as a product, but from using it as a feature to enhance existing services, with only a few players like Google and Palantir profiting directly. The hosts warn that the economic foundation of AI is shaky, with costs far outpacing actual profits, and caution against overreliance on external AI providers like OpenAI or NVIDIA. They urge organizations to build flexible, vendor-agnostic strategies, prioritize use cases over rigid business cases, and consider on-premises AI deployment using smaller, task-specific models to reduce dependency. They also debunk the myth that LLMs equal AGI and stress the importance of understanding AI's limitations, especially hallucinations and error propagation in complex workflows. Despite the risks, they affirm AI’s long-term inevitability and advocate for strategic agility over reactionary panic.