OpenAI vs. Anthropic Enterprise War, OpenAI Ad Prices, Synthesia’s $4B Valuation | Jan 26, 2026

0:13Welcome, everyone, to the Informations TI TV. My name is Akash Pasricha. It is Monday, January 26th. I am coming to you from our San Francisco Bureau for the next few days. I am very excited to be here. Today on the show, we are covering OpenAI's push to take enterprise customers away from Anthropic. The information has exclusive reporting about Sam Altman's pitch to executives at big corporations that his company can become a one-stop shop for everything AI. We also have some new inside reporting about OpenAI's ads push. Our e-commerce reporter has the scoop on its new premium ads pricing strategy.

0:50I'll then sit down with the CEO of AI video company Synthesia. The company announced it closed a$200 million Series E at a$4 billion valuation. And we will wrap the show with some analysis on the NVIDIA CoreWeave investment this morning and a conversation with our cloud reporter who just came back from the Davos of the data center sector in Hawaii. She is bringing back some great insights from the event. It's a great show, so let's get right on into things. OpenAI is revving up its enterprise sales team as it tries to explain to customers that it is more than just a chatbot business. The information reported exclusively that the company is trying to take customers away from Anthropic, which has seen some of the most success in the sector with respect to selling AI to enterprises.

1:41Joining me now is Kevin McLaughlin, one of the reporters behind that story. Kevin, welcome to the show. It's great to have you here. Thanks for having me. So this story begins with a dinner between Sam Altman and Bob Iger last week. What was this dinner about? Well, it wasn't just Bob Iger. It was a number of other Fortune 500 executives who are already OpenAI customers. And the pitch was basically not really new because OpenAI has been selling to businesses for the last couple of years. I think what's new about what's happening now is that OpenAI recognizes that they have to do something. They have to have an answer to cloud code because cloud code is in many ways sort of happening in the enterprise the same way that ChatGPT kind of swept across the landscape three years ago.

2:30This is a product that is spreading by word of mouth. Companies are trying it. People are talking about it. It's going viral. And so OpenAI is just sort of formalizing, I think, their approach to selling to enterprises. So they do have a new product that we talked about in the story. The details aren't entirely clear what it is, but at a high level, the pitch does appear to be that OpenAI is telling enterprise customers, like, we've got everything you need, so why would you need anyone else? I don't think they're specifically saying like Anthropic, we're going to take them out. But I think that is the goal.

3:09So we have OpenAI. They've got a new product. We know that's out there pitching. We don't quite know the details of it, but we know this is part of the enterprise sales push. I want to ask you about the relative sizes of Anthropic's enterprise business and then OpenAI's. I mean, these are two companies at different scales right now, but Anthropic obviously prioritizes enterprise. Is it bigger than OpenAI's or what do we know about the sizing? It's not bigger revenue-wise, but the share, Anthropic's share is 80 % of the revenue comes from stalling businesses, whereas OpenAI's was recently at 40%.

3:46I think the goal or they're saying that 50%, they'll have 50 % by the end of the year. So from the beginning, Anthropic has sort of specifically said that we're going to speak about our AI to businesses in a way that businesses like to hear about AI. So we're just doing one thing. Enterprises love focus. Enterprises don't generally like bells and whistles. And so that has been Anthropic's approach from the beginning is to talk about security, to talk about trust, talk about predictability of its AI. Now, if coding is such an important part of this story, remind me, it's called Codex is OpenAI's coding tool, right?

4:30Yes. And how good is that tool compared to Claude Code right now? Is it getting better? Well, our recent reporting has shown that Codex has made strides in recent months. So certainly, OpenAI is putting a lot of effort into making that product better. In the story, we reported that OpenAI has communicated to consultants who sell to its customers, who work with its customers, that some coming upgrades that OpenAI is working on will put Codex perhaps ahead of Cloud Code. And we've yet to see exactly what that means. But the point is that OpenAI is definitely working on this piece of their business, focusing very hard on the Codex product right now.

5:18So the interesting part to me about this is that the technical specifications and the performance of the product is one thing. But really, I think one of the challenges OpenAI has is that the world doesn't really think about it as an enterprise company. I mean, this seems to be a little bit of a branding challenge, even so much as it is a product challenge, because, I mean, you know, I guess we have businesses that buy ChatGBT for their enterprises, too. But, I mean, I don't really think about it as, you know, one of the top enterprise companies in the world. Yeah, I mean, like I said before, they certainly have a lot of penetration in enterprises.

5:54They've got big name customers like Disney, for example. they just struck a deal last month. I think going back to what I said before, enterprise customers might not love the fact that OpenAI is doing all kinds of different things that are not enterprise software, like some of the fun stuff that they've announced with video and image generation, which of course have business applications, but have gone viral in a very sort of non-business type of way. um so you know this is all sort of tough to really quantify but um like i said before enterprises like predictability enterprises like boring i think might be the best way to say it and by that i mean um not doing things not not pulling things out every week to to show people and just staying focused on a very specific set of products so i mean in the end here if enterprises like boring What are the big questions that you have on your mind as this story moves forward?

7:00You've covered a lot of companies in your career trying to move into enterprise. It's a very, you know, much more predictable segment. What are the key things that OpenAI has to be able to do and what are you watching for? I think in some ways this moment reminds me of what it was like for the major cloud providers like AWS and Google and Microsoft maybe 10 or 15 years ago when their focus started to move from selling to developers to doing larger contracts with business customers, with enterprise customers. And in the course of that push, AWS, for example, they started hiring enterprise executives from other companies.

7:40And you're already seeing that happen at OpenAI. Denise Dresser from Salesforce, Slack unit, for example, is now the chief revenue officer. So I expect to see a lot more of OpenAI hiring seasoned enterprise executives, as well as filling out their ranks with people who've worked at enterprise companies, because that was the playbook that cloud providers followed. You know, on another level, everyone has always been trying to pitch themselves as the only company that you need in the enterprise data center. And it's not always successful. And in fact, enterprises are always a hodgepodge of different products anyway.

8:19So I think it might be tough to sell that sort of comprehensive, broad strategy. But at the same time, I think the battle, the first battles that we're going to see are in AI-generated coding. So yeah, Codex versus Cloud Code. I'm sure that that is going to be a big theme this year. And it will be interesting to see how it plays out. Great. Well, Kevin, I want to thank you for coming on. That is Kevin McLaughlin, our enterprise software reporter. here at The Information. Okay, OpenAI is looking to charge a premium for advertisers who want in on its platform. That is according to a new exclusive story from my colleague Anne Guillen, who scooped some inside details about how much the company is planning to charge brands and what that means for its strategy.

9:05I want to bring on Anne to talk all about it. Anne, welcome back to the show. It's great to have you here. Hey, Akash. So how much is OpenAI looking to charge advertisers? So according to my reporting, the price that OpenAI has been floating with potential participants in the ad pilot is around$60 per 1 ,000 views that the ads get. So they're using a kind of per view pricing model. And yeah, the price that they're hoping to get is roughly$60 per PPM, which is kind of the ad jargon for 1 ,000 views. Okay, so that's the rate. So for those of us who don't know ad speak, is 60 bucks for 1 ,000 views, is that good or is that bad?

9:56Pretty expensive. It's pretty bright. Expensive, okay. So that typically compares a couple ad industry analysts compared it to that's how much you're paying to advertise on like Sunday night football on streaming and like really kind of high visibility, you know, very premium ad inventory. So I think that pricing offers kind of a little hint about what advertisers OpenAI might be trying to go after with this initial ads push. And yeah, it kind of offers a little bit more insight just into how they're approaching this initially. So this is kind of interesting to me because, I mean, I get the premiumness of this all, but also, and we'll talk about this, this is like highly experimental, right?

10:45I mean, like you're charging a premium for something that you haven't really proven out. So, I mean, fine. OpenAI is on top of the world right now, so they can do whatever they want. But I want to ask you about what the advertisers are getting back. You talked about it being in line with, for example, primetime TV slots. They obviously get a lot of data back from who's viewing it and what the results are. Is OpenAI offering that same data? So as of right now, OpenAI is planning to give advertisers in the pilot information about impressions, which is ad world speak for views, basically, and then some information about how many clicks the ads drive as well.

11:30But it's really kind of high level OpenAI in their post announcing the ads pilot has said that these will all be very aggregated insights. I think they want to be very clear that they're not at the stage yet where they're looking at individual users and their conversations or any actions that they're taking after they see ads. And so I understand why OpenAI wants to take that approach initially, especially because introducing ads into ChatGPT seems to be such a thorny problem in terms of how it goes over with users. But I do think that over time, if they keep that very high level, very limited approach, advertisers might not end up thinking it's a really valuable place to advertise over time just because there are so many other digital ad formats where you can get very, very granular insights about where exactly your ad spending is going, who's seeing your ads, and then what actions they're taking afterwards.

12:39which is really kind of the holy grail of what advertisers want to know. But how much initial interest is there from advertisers based on this premium pricing, from what you're hearing? Oh, the pricing is not a turnoff at all. There are plenty of brands that are dying to advertise. I want in! I want in! And ChatGPT, yes, they all want in. And I think that the appeal for advertisers right now, at least, is that ChatGPT just has such a massive audience and so many millions of users. And especially because this is a relatively, this is a brand new ad offering. usually advertisers, they want to be first to new ad channels because the thinking is there usually is a little bit of arbitrage you can get by being one of the first advertisers somewhere.

13:29We saw this with this company, Applovin, which helps place ads in mobile apps like Candy Crush. So the thinking is, while there aren't very many ads in a particular place or app, by being one of the first advertisers, then your ads can potentially be more impactful and you can get a little bit more kind of bang for your buck. And so last question for you, extending that, bang for your buck, arbitrage, being early out of the gate. If they can't measure these things to the degree that they would other advertisements, I mean, success for them is what, just being first or is it learning? Yeah, yeah, yeah, for advertisers.

14:13Yeah, I think for advertisers, you know, the appeal is getting in on this early and also potentially being able to have influence in whatever OpenAI does next with the ad product. I think a lot of advertisers are hoping that by trying to establish some kind of relationship with OpenAI early, then, you know, as they potentially look towards, you know, more click-based advertising, you know, adding advertising into their commerce features somehow. Somehow, I think a lot of advertisers are hoping just because OpenAI has been so hard to reach on the topic of ads specifically, I think a lot of advertisers are hoping that if they can get in some form of regular contact or have some form of relationship with OpenAI, then they can use that to their advantage down the line.

15:11Great. Well, Anne, I want to thank you for coming on. That is Anne Guillen, our e-commerce reporter here at The Information. Okay. AI video company Synthesia raised a new round of funding. The company today announced it closed a Series E of$200 million at a$4 billion valuation. I want to bring on CEO Victor Ripperbelly to talk about where he is going next with his company. Victor, welcome to the show. It's great to have you here. Thank you. Thanks for having me. What are you going to use all this money for? look I think what we've seen over the last couple of years just that the opportunity in AI video is really really big the last five years have been all around making it faster and better for people to create video by using AI instead of cameras actors studios and so on and so forth we think the next chapter of all this is what does AI video look like that's truly sorry AI native video truly look like which means like looking beyond just using AI to create video as we know It means video is going to be interactive.

16:13It's going to be personalized. It's going to be video you can talk to. If you're consuming a training video, for example, rather than just passively consuming five minutes of content, you can now consume some content and go into a conversation with the video. It can pretend to be a customer and do role playing, ask you a question, do a whole bunch of other things. So it's a lot of product development. It's also scaling the core business, which obviously means customer acquisition, hiring more amazing people, and all the new shows. And tell me, how is it that your video generation products, how does it differentiate itself from Google's text to video products?

16:48You know, we've had Runway on the show. There's obviously Mid Journey in the mix. What's different about you guys? So I think all the names you're mentioning here, I'd say it's much more about the AI models and actually generating kind of clips. I think what we've been very successful with over the last five years is to build a true platform of which the AI models is a component but it goes much further than that. So we start helping you write the script of your video, for example. Then we have the avatar technology, of course. We have the video models. You can use video free and Zora insights in PC if you want to.

17:18We have a full video editor. You can add screen recordings, text animations. It's a collaborative platform. It's a content management system. It's even a publishing platform. We have a video player that's built to work with these AI videos that can deliver the video in many different languages. You can make it interactive. And so I think the reason that VGA is winning in the enterprise crisis because we've really thought about how do we make this workflow really, really fast, which is about going from an idea to a final video of which the AI models are important, but let's say they're trying to tell that workflow.

17:48So you're taking on Adobe, essentially. We're definitely taking all kind of more workflow centric companies. The one you mentioned earlier on is much more like model companies. We also develop our own models, but we put a lot of effort into essentially building this like workflow, which involves the video edits, the collaboration platform, and so on and so forth. In terms of the users, I think, looking at a company like Adobe, generally Adobe is more products for experts, so people who are experts in video or image or Photoshop, whatever. What we're seeing with Synthesia is more the PowerPoint user, your sort of average office worker being able to quickly go in and start creating video content instead of creating slides or text documents.

18:30But in spirit, yes, it's all about hunting, creating platforms and tools. How much revenue is the company doing now? So we announced$100 million in AR in April last year. We're not that public about our revenue for this round. But obviously raising this, I think is the signal and growing really, really quickly. And if you look at the breakdown of the revenue then, what proportion of that is consumers versus enterprises? It's very high on the enterprises. We made a very conscious decision five years ago to build for the world's biggest companies. And so I think what appeals to investors around Synthesia and I think kind of the multiples we can demand, because it's very, very high quality revenue.

19:14We have plus 140 % NRR, for example. We work with more than 90 % of the Fortune 100. This is not just individuals kind of signing up with a credit card, playing around, trying to make some cool viral content social media. this really is the world's biggest companies adopting this as a kind of, you know, a standardized way of how to do it. So this is like more than 50 % of revenue is coming from enterprises now? Significantly more than that. Oh, wow. Okay. So it very much is an enterprise video company. For sure. Now, you guys are also pushing into agents. What does agents look like with video production?

19:53Help us, you know, imagine this. Yeah. So I said before, the last five years, we've lived a bit about how can we make video production as we know it just better and faster with AI. And I think clearly that's a huge market. We're attacking it from an enterprise central kind of POV. There's other companies doing it for Schumer. Clearly it's a huge market and we can build a great, really big company based on that. But I think the$100 billion question in this market is what does truly AI native video look like? And I think that means taking into consideration the fact that we're now LLMs. We have intelligence as a utility, right?

20:27We have real-time generation of avatars and videos also coming up inside the platform. And so for an end consumer, what this means is that video will look different. It'll not just be a broadcast where you create one video and then everybody watches exactly the same video. The video will be personalized to every single person that watches it. It will be interactive. So, you know, the example I gave earlier with things like a training video, for example, rather than having your sales team just passively consume content describing your new pricing policy, You can now have that as maybe the first part of the video.

20:55And then it can go into an agentic experience where the agent pretends to be a customer. And for 15 minutes, you have to overcome objections, explain the pricing policy to what is a simulated customer, which both means that the seller will understand the content much better. We know that the best way to learn something is a combination of kind of curriculum-based content you consume yourself with a tutor. and also begins to give management and executives a very, very interesting data set into how enabled is my workforce? Do they actually understand all the different facets of our sales methodology of our product areas?

21:30And that's one of the things that's really appealing in enterprise right now. Everybody wants to upskill their workforce to be able to work with all these new AI technologies, but it's very, very difficult to understand who's good at what and how can we help improve on whatever they're less good at. And this agent product is a world internet. Eventually, it'll scale to other use cases as well. We're excited about things like having, you know, first screening call for recruiters. You could do that with a Synthesia video with an agent in it. You can interact with the viewer, you know, customer support.

22:01There's a lot of our use cases we will enable down the line, but we're kind of starting with the training one. It feels like I'm sort of thinking about all the different ways that video can be used, And it feels like those corporate videos are one of the main use cases because, you know, I'm thinking about personalized video as it relates to content creation. And, you know, I mean, if you think about movies and YouTube videos, I mean, the value of it is having, you know, a single piece of art that everyone can talk about seeing the same thing. So I guess what you're talking about is really much more the interactiveness of, you know, being able to use video as a way to scale company operations, right?

22:42That's definitely the market that we're targeting. I think that said, what will be very exciting is to see what kind of new media formats people come up with with these new technologies outside of CNDC as well. I think you're right that the appeal of video as we know it is the fact that you have one piece of art everyone can discuss. But there's also a lot of people that play computer games. Computer games is within certain boundaries. Everybody has a slightly different experience. They play with different people. You choose a different storyline. And I think it's a lot of those mechanics will just begin to see and seep into video and audio for that matter.

23:14Right. And that's very exciting. There's a corporate version of it, which is the one that we're kind of going for. And that's all about, you know, we've always said that we're much more about like knowledge sharing type of videos, less like marketing, advertising, storytelling, you know, really creative use cases. And it's just very obvious that video you can interact with is going to, you know be so much better almost no matter what use case uh you're you're working with in in knowledge sharing even if it's as simple as you can ask the video a question if there's part of the content that you don't understand now i want to ask you about your costs here because none of this can be cheap certainly when it relates back to the compute costs how are you dealing with that because it it can't be cheap to to run all this in the background we actually have like you know great SaaS margins.

24:03We're not one of those companies that's running it. Like 70 %? 75 % margins? Yeah. Really? This is an AI video company that does real-time authentic video with 75 % gross profit margins. Yeah. Well, the video, the live product hasn't launched yet, right? They'll have a different cost profile. But in general, when you look at Synthesia as a business, it's not just fast growth, it's high quality revenue, and it's really great unit economics. And I think that is, There's a lot of different strategies, but I think building a platform and selling a workflow rather than just selling here's an API call to my model just means that you can capture much more of the value.

24:41And so part of this, of course, also because we develop our own models, which means we can fine tune them for specifically what we need. A lot of these very big video models, they're amazing and creative because they can do absolutely anything you want them to do. But it also means the models have to be very, very big. Right. If you're focusing in on like talking head style videos in our world, for example, it just means that you can get them to much better economics as well. And so what's holding the company back right now? What's the rate limiting step? I think hiring is probably, you know, one of the biggest rate limiters.

25:15Getting people in seed, getting them to, you know, start selling some easier, building some easier, I think is really, really getting great talent to any business. I think it's always fun. Like there's a lot of talk in AI about how AI is going to replace and automate everything. And of course, that's also happening at a really grand scale in Synthesia. But if you look at it, AI companies are the ones hiring the most and the fastest right now. And I think that great people and great talent will always be the core DNA of winning companies. Great. Well, Victor, I want to thank you for coming on the show.

25:45That is Victor Ripperbelli, the CEO of Synthesia here on TI TV. Okay. We had some big news this morning. NVIDIA is expanding its partnership with CoreWeave, looking to accelerate its data center buildout with a$2 billion investment. The deal will support CoreWeave's buildout of more than 5 gigawatts of AI factories in 2030. Here with some analysis is the information's cloud and compute reporter, Anissa Gardizi. Anissa, welcome back to the show. It's great to have you here. Thanks, Akash. Great to be in the same office as you. This is great. Yeah, crazy. I didn't even get to see you in person yet.

26:21You pulled the shade down. I did. Yeah. Well, we got to manage the lights. Okay. So look, I want to get to this CoreWeave deal because you have done a lot of reporting about how NVIDIA is a customer, it's an investor, it's a supplier, it's all these things. So what did you make of the CoreWeave investment this morning? Yeah. NVIDIA invested$2 billion into CoreWeave common shares. So CoreWeave stock was up about 8 % the last time I checked. And what CoreWeave says is that this investment is going to help it build out data centers faster. The number they used was 5 gigawatts of data center capacity by 2030.

27:03I think one big takeaway here is that it's going to take a lot more than$2 billion for CoreWeave to make even a dent in that goal. But I think it's a huge vote of confidence from NVIDIA in CoreWeave, which isn't too surprising, but maybe we'll help CoreWeave get other partners on board as it looks to expand that quickly. And what does this tell us about CoreWeave's long-term strategy in terms of it building its own data centers? I mean, the NeoCloud business, I mean, the roots of this was essentially renting some of these chips and then renting capacity out, right? Totally. So the thing that all of the NeoClouds, including CoreWeave, have been thinking about over the past couple months is, do they need to diversify away from NVIDIA hardware in order to differentiate and have a sustainable business that isn't just solely reliant on being an NVIDIA partner?

28:00And last week at the World Economic Forum, our founder and CEO, Jessica Lesson, actually asked the CEO of CoreWeave, Mike and Trader, you know, are you guys going to consider other NVIDIA alternatives? and I think his answer was pretty telling of the announcement that we saw this morning which is that from CoreWeave's perspective they think that NVIDIA hardware is best in class that's where they're seeing demand so for the time being that's the hardware they're going to stick with but I think a lot of other companies and we've reported on this are considering whether other chips you know maybe from AMD or Google would help them accelerate their business but I think by and large, these startups primarily rely on NVIDIA.