Wedbush: Missing ROI Metrics Raise Risk for Enterprise AI Rollouts

Many enterprises have deployed AI pilots without building a way to measure whether they are getting a return on that investment. Wedbush Securities analysts said that lack of measurement now threatens to slow further AI spending, according to a Friday note from the firm.

Dan Ives and his team learned at the Wedbush Disruptive Technology Conference that customers face growing pressure from boards and CFOs to show actual returns. Without that proof, enterprises struggle to justify more investment and identify what works, the analysts said. The inability to answer the ROI question presents a real barrier to additional spending on long-term AI buildouts, Ives added.

The risk is not theoretical. If enterprises cannot measure returns, they are more likely to cut or delay AI projects. That would hit companies selling AI infrastructure, software, and services. NVIDIA, for example, has ridden a wave of enterprise AI deployment. A slowdown in that deployment would pressure its revenue growth. A recent report noted similar concerns about AI spending ROI questions mounting across the sector.

PYMNTS Intelligence data supports the concern. In a survey, more than eight in 10 enterprise executives said it could take three to 10 years to see positive payback from generative AI. Another PYMNTS report found that 71% of executives see organizational readiness – people, processes, data quality – as a bigger constraint than AI technology itself. They cited an average of four to five barriers, including budget limits and governance processes.

The Wedbush note adds a financial layer to those operational concerns. A company that cannot articulate ROI cannot persuade the CFO to fund the next phase. That creates a circular problem: limited deployment, limited data to measure, limited case for more deployment.

For traders, the key question is whether the narrative around AI spending shifts from "must have" to "show me." Signs of that shift would include more cautious guidance from AI hardware and software vendors, delays in large enterprise contracts, or increased disclosure of AI spending metrics in earnings calls. The Wedbush note is a reminder that the AI buildout depends on enterprises solving a measurement problem they have not yet solved.

Wedbush plans to track how companies address the ROI question in their upcoming earnings reports.