Can OpenAI build a $1 billion ad empire?

Kamika Bukhoi (0:10)

Hello, hello, and welcome to another episode of The Digiday Podcast, a show for anybody wondering if OpenAI is going to hit its ad revenue goal this year. I'm Kamika Bukhoi, Senior Marketing Reporter here at Digiday.

Tim Peterson (0:23)

I'm Tim Peterson, Executive Editor, Video and Audio, Digiday Media. Kamika, you know who we haven't had on the show in a bit?

Kamika Bukhoi (0:30)

Is it our resident OpenAI reporter?

Tim Peterson (0:34)

It is. It's our Senior Platforms Reporter, Crystal Scanlon, who, Crystal, we gave you the summer off basically, but now Q4 is about to start. We're getting you back in the mix.

Crystal Scanlon (0:44)

I feel like I'm very grateful for those few weeks off.

Yeah, I feel like from now on, it's just going to be chock-a-block with anything OpenAI. For the caveat, obviously, still do write about other platforms, but just this has taken a whole bunch of my year up.

Tim Peterson (0:58)

For good reason. I mean, this is the year OpenAI famously launched ads, and it's been busy. Like Kamiko, how many times have we had to talk about OpenAI and kind of refresh ourselves on all the things OpenAI has launched this year?

Kamika Bukhoi (1:15)

I'd actually need more fingers and toes to count if I was using fingers and toes. It's been a shit ton.

Crystal Scanlon (1:23)

Yeah. I mean, every other week, they're launching new things even on the ad side of it.

When I've been speaking to the team, even they said we're getting to a point where we're just wondering we're releasing things every other week. Like is that too much for even you to keep up with? And I'm like, I mean, it's great to see you're doing new things, but not every single individual feature will necessarily warrant an individual story on that. There's just too much at this point.

Kamika Bukhoi (1:48)

Crystal, you recently reported back in August that ChatGPT, which is OpenAI's ad business, obviously just hit its $1 billion revenue run rate after introducing self-service over to Europe. That's a really big number. How did we get here?

Crystal Scanlon (2:07)

Right. Well, first of all, I mean, yes, it sounds amazing and it sounds glorious. Like, you know, they've already said like 1 billion annualized runway in terms of dollars.

But that doesn't necessarily mean that they're already making over a billion, like at all. Like an annualized run rate is literally just a financial metric that's used to predict a company's total yearly earnings by taking a short-term revenue data.

So, for example, typical ways that they like will work this out. If you're using monthly data, they'll multiply one month of revenue times 12, because 12 months in a year. Or if they're doing it quarterly, they'll multiply what they've made in one quarter and times that by four quarters, because we've got four quarters in a year. That is how they get to that figure. So what that basically means is, on the trajectory that we're on with how much money we've made, obviously we don't know specifically whether they've made this massive announcement that they've created it via either monthly or quarterly. But however much, whichever way, whichever quantity that they were using, they've obviously multiplied it and then decided that, provided nothing changes between now and the next 12 months, this will be what we are projecting that we are going to be able to make on an annual basis.

Tim Peterson (3:16)

Crystal, are you saying that an unprofitable tech startup is using fuzzy math?

Crystal Scanlon (3:22)

In a very roundabout way, possibly.

Tim Peterson (3:25)

Do you know what the basis is for this? Because I've known even companies in the past who used one week's revenue to project annualized revenue, which is the most bullshit type of number possible.

Crystal Scanlon (3:37)

Yeah. I mean, no surprise. They want to go public. And something like this is obviously going to make sure that they want to make sure they're in favor with investors.

As much as investors might be like, yay, new technology, AI, albeit the mood on that is definitely changing. At the same time, investors are also not stupid enough to go and invest in a company that has no positive trajectory to give them returns. So for them to actually get investor buy-in, especially if the mood on AI is changing, they need to be able to guarantee that they're going to get really good return if they do invest with them.

Tim Peterson (4:12)

But is that billion-dollar annualized revenue run rate, which again is not a real number. It's a hopeful number. My run rate looks a lot better on payday weeks versus non-payday weeks. But didn't they have a bigger number in mind for 2026 in terms of how much ad revenue they expected to bring in this year?