Everyone in media is talking about OpenAI like it’s the next Google. And maybe it will be. But “maybe” doesn’t belong in a media plan when your client has a Q3 sales target and a finite budget.
Let’s be direct about what’s happening here.
OpenAI is staffing up an ads team, courting brand partners, and dropping rates to get buy-in early. That’s a classic land-grab playbook. It’s the same one Google ran in the early 2000s and Meta ran around 2012. While it worked for them, it also took years before either platform could reliably move product at scale for most CPG and DTC brands.
And even with a similar playbook, the outcome isn’t guaranteed.
The Shiny Object Problem Is Real
We’ve watched brands burn meaningful budget chasing early access to TikTok Shop, Pinterest Shopping, and connected TV inventory before the measurement infrastructure existed to justify it. Some of those bets paid off. Most of them produced a lot of learnings and not a lot of revenue.
OpenAI entering the ad market feels similar. The intent signal is genuinely interesting. People searching in ChatGPT are often further down a decision-making path than a Google keyword search suggests. That’s worth paying attention to.
But interesting intent signals and proven ROAS are two completely different things.
What We Actually Need Before We Allocate Budget
For any new channel to earn a place in a performance media mix, we need three things: reliable reach, measurable conversion, and clean attribution. Right now, OpenAI has none of those at the scale our clients require.
Reach is growing fast. ChatGPT has real daily active users and that number is climbing. However, “growing fast” for a tech platform and “sufficient scale for a national CPG launch” are not the same thing.
Measurable conversion is the bigger question. If someone sees a sponsored result in ChatGPT and buys a product three days later on Amazon, how does that get attributed? Nobody has a clean answer to that yet, including OpenAI.
This Doesn’t Mean Ignore It
We’re not telling clients to sleep on this. We’re telling them to watch it with their eyes open and their wallets mostly closed. For now.
The brands that will win on OpenAI’s ad platform are the ones who test small, early, and with clear hypotheses. They’re not the ones who pour budget in because a sales rep promised premium placement, and not the ones who wait until CPMs triple once the platform matures.
Small tests in Q4 2025 (if the product exists and the measurement hooks are there) make sense for the right brand. Think high consideration purchases, strong brand story, and audiences who actively use AI tools. A DTC supplement brand or a premium food company with a complex product story could find real signal here.
Alternatively, a commodity CPG brand trying to move units at Walmart? Probably not the right fit yet.
The Meta and Google Question Nobody Is Asking
Here’s the take that most agencies won’t give you. OpenAI entering advertising is less interesting as a new channel and more interesting as a pressure campaign on Google.
If advertisers have a credible alternative search-adjacent platform, Google has to respond. That might mean better pricing, better tools, or more transparency in how search ads are placed and measured. Any of those outcomes benefits performance media buyers more than OpenAI’s ad product will in the short term.
The holding companies with deep Google commitments won’t say this out loud. We will. Competition in the ad market is good for brands.
Our Honest Advice for CPG and DTC Brands Right Now
Don’t restructure your media mix around OpenAI in 2026. Do assign someone on your team (or ask your agency) to monitor how the ad product develops. Do ask your current agency what their plan is for AI-native search advertising, because if they don’t have one, that’s a problem.
And if any agency is telling you to shift significant budget to OpenAI ads right now based on potential alone, ask them whose interest that recommendation actually serves.
Performance media only works when it’s accountable. New channels earn their place by proving they move product, not by generating headlines.
We’re watching. We’ll tell you when it’s time. Learn more about our other Services.
Chris Pyne, Founder, Junction 37 – 30+ Years in performance media.