AI’s Control Problem: Agents, Costs, and the Robot Gap

Bullish Gator

Executives discuss runaway agent costs, control and kill switches, inference economics, and the U.S.-China robot deployment gap that could sink AI businesses.

Editorial companion

What stands out

Jeetu Patel, Cisco's president, told CNBC that autonomous systems need a kill switch, reframing the AI debate as one about control and cost not just capability. The panel laid out a harsh math: artificial intelligence (AI) agents can multiply compute bills until a $20 task becomes a $50,000 invoice, and founders warned inference economics and a U.S.-China robot deployment gap are already reshaping business models. For Cisco, which sells the networking and security layers enterprises rely on, the conversation is about who gets paid to stop the runaway systems.