TL;DR

Major banks are moving away from OpenAI as their primary AI provider, with market share dropping from roughly half to one-third over 18 months. Anthropic and Google are gaining ground, particularly in coding use cases where tangible ROI is demonstrated.

Banking’s AI Provider Landscape Shifts

The world’s largest banks are rethinking their AI vendor strategies, according to new data from AI benchmarking platform Evident. Their Use Case Tracker, which monitors publicly announced AI deployments across 50 major banks, reveals a marked shift in provider preferences.

Eighteen months ago, OpenAI powered roughly half of banking AI use cases that disclosed vendor information. By the end of 2025, that share had fallen to just one-third.

“Since the Gen AI boom began three years ago, banks have been the bellwether for enterprise AI adoption,” says Alexandra Mousavizadeh, co-founder and CEO of Evident. “The way they are behaving now could signal mounting challenges for OpenAI’s position in banking in 2026.”

Why the Shift?

Coding remains one of the few banking AI use cases where over half of deployments show measurable return on investment. Anthropic consistently ranks best-in-class for code generation and review, while Google has used its existing cloud relationships to simplify enterprise deployment.

BNY recently demonstrated this approach by integrating Gemini Enterprise with its internal AI platform, Eliza.

Model Agnosticism Gains Traction

Many leading banks favour a model-agnostic approach, allowing them to switch providers as capabilities evolve. However, Mousavizadeh notes that bankers have expressed genuine enthusiasm for what Anthropic and Google are delivering.

OpenAI maintains successful partnerships with banks like BBVA and Morgan Stanley, so the coming months will test how the company adapts its enterprise offering to retain market position.