Corporate AI spending hits a wall of skepticism

The era of open-ended AI budgets is closing. Tech giants are now actively curbing internal usage; Amazon decommissioned its token-tracking leaderboard to discourage waste, while Walmart has implemented strict usage caps. Uber COO Andrew MacDonald and Cisco product chief Jeetu Patel have openly questioned the current pricing models, with Patel arguing that token costs frequently exceed the actual value generated at scale.

Consulting firms face a unique pressure in this climate. After rushing to integrate AI to fend off the threat of displacement, firms like KPMG and McKinsey are now pivoting toward granular oversight. McKinsey, which currently deploys 25,000 AI agents alongside its 40,000-strong workforce, represents the aggressive end of this transition. However, broader data from the Boston Consulting Group suggests a looming paradox: while companies plan to double their AI spending by 2026, reaching 1.7% of total revenue, the return on these experimental billions remains notoriously difficult to quantify.

Russell Fradin, CEO of Larridin, notes that while organizations are not yet cutting budgets, they are obsessed with instrumenting their spend to avoid the unsustainable cycle of 10x annual growth. For the consulting industry, the mission has shifted from mere implementation to justifying the bottom-line impact of every agent deployed.