What: AI’s below-cost pricing is creating unstable assumptions for workforce, automation, and technology investment decisions.

Why it is important: This risk reinforces the need for disciplined AI governance, investment planning, and workforce strategies as retailers scale automation.

The current price of AI may not reflect what it actually costs to deliver. Ramp data shows that the heaviest corporate AI users spend about $7,500 per employee per month on AI tools and compute, compared with around $16,000 as the total employer cost for a U.S. software engineer, including salary, benefits, and overhead. That gap is driving workforce restructuring across industries, but it is built on pricing providers are losing billions to sustain. AI providers are following a familiar Silicon Valley model: price below cost to build market share, then raise prices once adoption locks in. OpenAI’s leaked financials show $13 billion in revenue against $34 billion in costs in 2025, an operating loss of $21 billion in a single year. Even as token prices fall, newer models and agentic systems consume more tokens, so total bills keep rising. Uber burned through its entire 2026 coding budget by April before capping employee AI spending in June. Companies cutting roles or freezing headcount based on today’s AI costs may face serious budget pressure if prices rise.

IADS Notes: Retailers who are restructuring work around current AI pricing may be building on costs that won’t hold. In June 2026, BCG found that retailers and CPG companies are using AI to improve margins, decision-making, customer engagement, and operational execution, but only a minority are scaling it effectively, which means many may be committing to AI-driven workflows before the economics are clear. BCG’s June 2026 work on AI strategy stressed that governance and accountability matter more than tool deployment, a discipline that is especially important when at least some providers are losing billions to support current pricing. In April 2026, BCG connected AI deployment in pricing, labour, procurement, and forecasting to margin improvement, while noting the capital investment required to get there. BCG’s November and September 2025 work on AI-first retail, covering operating-model redesign, workforce upskilling, and human-AI integration, describes strategies built for a technology whose cost basis has not yet stabilised. The question is whether those strategies are being sized for AI’s current price, or its eventual one.

AI is currently priced below cost. That won’t last.