Target rethinks AI rollout as usage-based pricing forces strategy reset

TL;DR:

• US retailer Target is moving from “using AI to running on AI” but is reassessing where it deploys the technology as suppliers shift from subscriptions to token-based pricing.
• Discussions about AI pricing are “at the highest level” in Target’s architecture and technology leadership forums, India president Andrea Zimmerman told Reuters in Bengaluru.
• The company plans to spend an additional $2 billion (£1.5 billion) this year under new CEO Michael Fiddelke on new stores, remodels and AI, even as Target enters a fourth year of revenue decline.

Zimmerman’s comments are notable because they put a named Fortune 500 retailer behind a trend most UK enterprise buyers have been monitoring privately: that frontier AI vendors are quietly retiring flat-fee subscriptions in favour of usage-metered billing, and that this is changing how customers decide where to deploy AI at all.

Context and Background

AI firms including Anthropic and OpenAI have increasingly shifted to token-based pricing that charges per request rather than per seat. The economic logic is straightforward — model serving costs scale with usage — but the implication for enterprise procurement is significant: when AI was a flat subscription, deploying it everywhere was the rational default; under usage pricing, “intentional use and integration” (Zimmerman’s phrase) becomes the rational default.

Target’s tech footprint in Bengaluru employs around 5,600 people across merchandising, digital, stores and supply chain — about 40% of the retailer’s overall tech workforce. India is becoming the test bed for the company’s AI rationalisation, partly because cost discipline is sharper outside Minneapolis headquarters.

For UK enterprise buyers, the relevant context is timing. UK financial services, retail and public-sector buyers signed multi-year flat-fee AI contracts in 2024–25 as vendors competed on land-and-expand terms. As those contracts come up for renewal through 2026, expect the renewal terms to shift toward consumption — and expect FinOps teams to start treating AI tokens the way they treated cloud egress fees five years ago.

Looking Forward

Zimmerman’s quote — “AI is fun, exciting and interesting to think about. Change isn’t going to be immediate, and it is certainly not free” — is the polite enterprise version of the warning UK CIOs have been issuing internally for months. Resultsense expects pricing-driven re-architecture to become a dominant theme in UK enterprise AI procurement through the second half of 2026, with workflows that cannot justify per-token cost being deprioritised regardless of headline capability.