AI

For all the hype around generative AI, the bottom line is looking bleak.

A sweeping new study from MIT Media Lab found that 95% of enterprise GenAI projects show no measurable return on investment. Companies across industries rushed to spin up pilots—chatbots, code assistants, automated marketing—but most of those efforts stalled before reaching scale or delivering bottom-line value.

The report estimates tens of billions have already been poured into generative AI, yet only about 5% of initiatives generated “millions in measurable business impact.” The failures weren’t primarily about the technology itself, but about execution: poor data pipelines, murky goals, and a tendency to chase buzzwords instead of designing for specific business problems.

The divide is widening between startups and focused teams—who often succeed with small, targeted AI deployments—and large incumbents that try to blanket their organizations with tools but lack the clarity to integrate them. That mismatch has rattled investors: AI stocks dipped following publication of the MIT findings, raising new fears of a bubble forming around the sector.

Experts caution that this doesn’t mean AI is snake oil. Rather, it underscores the need for discipline. Projects tied to a clear KPI, human-in-the-loop review, and clean data are already showing traction. Without those fundamentals, enterprises risk spending heavily on shiny demos that never deliver.

Between the Lines — The Readovia Cut

The MIT study is less a eulogy for AI than a reality check. The next phase of the AI boom won’t reward fast followers or corporate hype machines—it will reward precision, patience, and proof of value. For Wall Street and Main Street alike, this is where the story shifts from promise to performance.