Morning BriefAI Search & Discovery

OpenAI and Anthropic Are Fighting Over Who Becomes Your Buyers' Research Interface. They Don't Control What Gets Recommended.

OpenAI is offering PE firms 17.5% guaranteed returns to embed ChatGPT across their portfolios. Anthropic is running the same play with Claude. Neither of them decides which vendors show up in the answers.

Jaxon Parrott|Mar 31, 2026

OpenAI is in advanced talks with TPG, Bain Capital, and Brookfield on a joint venture to deploy ChatGPT across hundreds of PE-backed portfolio companies. The deal includes a guaranteed 17.5% minimum return, well above standard preferred instruments, to lock PE firms in before Anthropic does. Reuters broke the terms last week.

Anthropic is running the same play. Blackstone, Hellman & Friedman, and Permira. Same structure, no guaranteed return. "There's a big race to lock in as much enterprise, as many desks as possible," Matt Kropp at BCG's AI unit told Reuters.

Both companies are right about the strategy. Once an AI model is embedded across hundreds of enterprise systems, switching costs are prohibitive. That stickiness is what makes the distribution fight worth $10 billion.

Here's what neither of them can control: what the model says about you.

When your buyer's AI assistant becomes ChatGPT and they ask it which vendor to use, the answer is not determined by who won the PE deal.

It's determined by whether you appear in the publications ChatGPT was trained on, and whether you're appearing in the sources Perplexity retrieves right now. Two different mechanisms. Both driven by one thing: earned media in publications AI engines treat as authoritative.

Forrester's 2026 State of Business Buying research, drawn from nearly 18,000 global buyers, found that 94% of B2B buyers now use AI during their purchasing process. Twice as many named generative AI as more meaningful than any other single information source, more than vendor websites, more than product experts, more than sales reps. The buying journey has a new first step, and it happens before your SDR sends the first email.

If ChatGPT becomes the research interface for every company in a PE portfolio, that number gets more consequential. The AI's recommendation layer, which vendors get surfaced and which get ignored, is built before the PE deal closes. It's built from years of editorial coverage, citation patterns, and what the model encountered during training.

OpenAI's enterprise distribution play doesn't change any of that.

The finding that cuts through the noise comes from Muck Rack's analysis of over one million AI prompts: 85.5% of AI citations come from earned media sources. Not brand-owned blogs. Not SEO-optimized pillar pages. Third-party editorial coverage in publications with real editorial standards.

Ahrefs' study of 75,000 brands put a number on the mechanism: brand web mentions, the output of earned media coverage, correlate 0.664 with AI visibility. Backlinks, the currency SEO programs are built around, correlate 0.218. The ratio is roughly 3x.

Signal type	AI visibility correlation	What it means
Brand web mentions (earned media output)	0.664	3x more predictive of AI citation presence
Backlinks (SEO currency)	0.218	One-third as predictive as earned media
Paid media	~0	Not cited by AI engines at all

This is not an argument for PR over SEO. It's a description of how AI engines decide what to recommend. They use independent editorial coverage as a proxy for credibility, the same signal that made PR valuable to human readers for decades. The mechanism didn't change. The reader did.

Every PE-backed company that adopts ChatGPT or Claude as an enterprise research tool extends this dynamic to their entire procurement workflow.

The procurement officer at a portfolio company asks Perplexity: "What are the best workflow automation platforms for a manufacturing company with 500 employees?" The AI composes an answer. Your company is either in it or it isn't. That evaluation happened before the question was asked. It happened in the months and years prior, as editorial coverage accumulated or failed to accumulate in the publications AI engines trust.

OpenAI closing the PE deal with TPG doesn't give your company a seat in that answer. It just changes which AI is asking the question.

What actually determines the answer: earned authority in trusted publications. According to AuthorityTech's citation research, earned media distribution delivers a 325% lift in AI citation rates over owned content distribution. The kind of presence that takes years and thousands of direct editorial relationships to build at scale.

Here's how the two distribution strategies compare in terms of what they actually produce for AI citation:

What it is	Who it serves	What it produces
PE joint venture with OpenAI/Anthropic	AI companies seeking enterprise stickiness	Locks in which model runs the research interface
Earned media in trusted publications	Brands competing to be recommended by that interface	Determines which vendors appear in AI answers

The PE deal and the citation competition are running simultaneously. Most companies are focused on the first and ignoring the second.

The practical implication:

The distribution war between OpenAI and Anthropic is real, and the winner matters for which AI your buyers use. But the winner of that war does not decide what AI recommends. That's a separate competition, and most companies are losing it without knowing they're playing.

Forrester's AI visibility research describes what they call the "visibility vacuum": as research moves into AI answer engines, marketers lose visibility into buyer questions, intent, and activity. The buyers still arrive at your site eventually. But they arrive already having decided, based on what the AI told them, and you had no visibility into that conversation.

As I've written about at jaxonparrott.com, this is where the strategic gap opens. The companies building earned media infrastructure now are establishing citation architecture that compounds over time. Each editorial placement, each independent source that mentions the brand, each piece of coverage in a publication AI engines treat as credible accumulates as the training and retrieval signals that determine which vendors get recommended. Christian Lehman has covered the tactical execution layer of this shift from the GTM side.

The gap between companies that started building this in 2024 and companies that start in 2027 will mirror the gap that opened in SEO between 2010 and 2015 adopters. Different mechanism. Same compounding dynamic.

OpenAI and Anthropic are fighting over the distribution channel. The channel does not pick winners. Machine Relations is the discipline for becoming one. If you don't know how your brand currently appears in AI answers, that's the first question to answer: run the visibility audit.

OpenAI and Anthropic Are Fighting Over Who Becomes Your Buyers' Research Interface. They Don't Control What Gets Recommended.

Related Reading