LG Doucet and Milk Road lead researcher Martin examine weakening US labor data alongside falling AI-agent costs. Martin argues that most companies remain in an experimental phase, but 2027 could mark the point when adoption produces more visible job losses and sharper pressure on businesses that fail to adapt.
The discussion maps opportunities across the AI stack, from power and compute to applications and the “harness” layer that governs enterprise agents. Martin favors diversification across these layers and highlights ServiceNow, AppLovin, Uber, Nvidia, and Google as examples tied to the trend, while warning that frontier model providers may face commoditization and price competition.