Turley testified OpenAI planned its own index from day one

Tomek Rudzki published a LinkedIn post on 4 September 2026 describing ChatGPT's retrieval stack, which he says is internally called Labrador. He lists a family of indexes covering general web, PDF, YouTube, news, arXiv, Wikipedia, local, finance, legal, medical, shopping and images.

The load-bearing quote is from Nick Turley, Head of ChatGPT, testifying under oath in a 2025 US antitrust hearing. Rudzki reports Turley said an in-house index was the plan from day one, and that OpenAI wanted Google data to improve its own index. Rudzki adds that recent OpenAI job posts mention exabyte-scale database systems.

Why the framing matters

Most practitioners have been planning as if ChatGPT rides on Bing. Sworn testimony that a proprietary index was seeded at launch reframes the licensing deals as a bridge OpenAI always intended to cross.

The vertical split Rudzki lists, with separate indexes for medical, legal, shopping and the rest, points to retrieval tuned per domain rather than one web crawl reranked. We think the testimony and job posts are directional. They speak to intent and capacity, not to how much of today's answer set comes from a proprietary index versus a licensed one.

How to test your own exposure

Pull the queries where you rank in Bing's top ten for your category. Check which of those surface you as a citation in ChatGPT and which do not.

The gap is your exposure to the shift Rudzki describes. If ChatGPT cites you where Bing ranks you, a Bing-to-Labrador transition is a smaller risk to you. If it does not, Bing rank is not carrying you today either, and the work is to become the source a careful specialist in your vertical would pick on the merits.