Inside OpenAI’s $20B Revenue Discrepancy & China’s Increasingly Crowded AI Race

Akash Pasricha (0:13)

Welcome, everyone, to The Information's TITV. My name is Akash Pasricha. It is Friday, October 9th. Today on the show, the AI race in China is getting more crowded. We're going to talk with our Asia Bureau Chief, Jing Yang, for the latest on that race. We'll then dig into revelations that OpenAI's revenue is not what we thought it might be. We're also going to bring on the CEO of ARENA to talk about the company's latest funding round at evaluation north of $3 billion. To close out the show, we've got one of Europe's brightest minds in AI, the CEO of Domyn coming on for a conversation. It's going to be a great show, so let's get right on into it.

The AI model race in China is getting more and more crowded. My colleague Jiro Osawa from The Information's Asia Bureau wrote a good piece outlining the current state of play and I want to bring on our Asia Bureau Chief Jing Yang to share more in this week's edition of The Editor's Cut. Jing, welcome back to the show. It's great to have you here. Hi, Akash.

Okay. So I want this Jing to be the explainer that everyone has really been asking for on the model race going on within China because there are names that we talk about a lot. We talk about DeepSeek, we talk about Moonshot, which of course has the Kimi models. We talk, what's the third one? Zipoo with the GLM models. We talk about them. We don't talk about the big tech companies in China and what they're working on and some of the new entrants that Jiro wrote about. So give us the current state of play. I mean, outside of those three, who are the other model making companies that we should be paying attention to?

Jing Yang (1:55)

Yeah, so it's getting really, in my opinion, quite unsustainable.

So as you mentioned, right, there's a few upstarts that have attracted a lot of attention globally.

But there's actually a lot of Chinese tech incumbents, which have existing businesses that are printing cash and they use that money to bankroll their effort to develop advanced AI. So there's Alibaba, Tencent and Bidens, which already have been doing this for the last 30 years. Then what we reported is some smaller names, but also existing tech companies that are entering the race. So there is Meituan, which is the biggest food delivery, grocery delivery platform in China. And then there's Xiaomi, which is a leader in smartphone, but then the company also sells electric cars. And then finally, there's Mihoyo, which is a video game publisher, video game studio. And they have published some hits that you may have heard of, such as the Genshin Impact. So all the last three names I mentioned are all entering the LLM race, which is making things pretty crowded.

Akash Pasricha (3:15)

Right. So let's go through these names. We'll focus on the second three here, because they're kind of the most exciting. But just very quickly, I mean, Alibaba Tencent and ByteDance, broadly speaking, why aren't their models the ones that we talk about the most? Are they not as good? Have they not, I mean, they have loads of capital to invest, right? And so what are the core reasons here why we don't talk about those models as much?

Jing Yang (3:43)

Well, we definitely have written actually many times about the three. But to give you a quick sort of summary, Alibaba is quite strong in their open source models named Quen. They are still in like sort of the top tier, they're still the top tier players in China. And then Bydowns is a global leader in video generation models. They're a C-downs video. If you remember, actually shocked Hollywood and also sparked a lot of lawsuits or complaints from Hollywood studios.

And then with Tencent, Tencent is considered a laggard among the three, but they are also quickly catching up. For example, we talked about, they last year hired a very young researcher from OpenAI who's now leading their AI effort.

Akash Pasricha (4:39)

Right. And I should say, I mean, when I say we don't talk about it, we write about it. We do talk about it on the show too, but I feel as though it doesn't come up in conversation, broadly speaking, in the AI story.

Jing Yang (4:56)

I just have to come back to the show more often. Problem solved.

Akash Pasricha (4:59)

There you go. Well, I'm not going to argue with that.

We will welcome you here every day if you want. We'll just have Jing's corner. How about that?

Let's talk about the three newer entrants, maybe you could call them, or entrants we want to pay more attention to. The lead that Jiro wrote in his story, which I thought was a good way of putting it, is imagine if Epic Games, if Apple, and if DoorDash were pouring colossal amounts of capital into producing their own models. This is effectively the case with the three names you mentioned in China. Are these models any good? How much competition is there?