Glean's AI Data Wars, Serena Ventures' New Fund, and Google's New AI "Fixer" | July 31, 2025

0:13Welcome, everyone, to the Informations TITB. My name is Akash Pastrucha. It is Thursday, July 31st, and it is a big day for tech markets. We have Figma debuting today. Meta and Microsoft shares are also up on their earnings from last night. We've got a bunch of guests coming on the show today. Arvin Jan from Glean is coming on. We've got Serena Williams Venture Fund coming on the show. We've also got our Google reporter who is going to tell us about the company's new chief AI architect. That was a profile that she published this morning. And we've also got a big six-figure funding round from another generative AI media company.

0:50We've got the founder of that company coming on the show. Before we get into earnings and into headlines, I want to alert you about a big story that we published last night. The information was first to report that OpenAI has hit$12 billion in annualized revenue. We've got this chart for you. Just to put the number in context, I mean, OpenAI was doing$500 million in annualized revenue per month. That was at the start of the year. It's now doing about a billion a month in annualized revenue. And so it has effectively doubled its top line over the course of 2025 alone. I should mention, you know, with that revenue growth, the company is also burning more cash.

1:29It's now projected to burn$8 billion this year. That is$1 billion more than it projected earlier this year. And so that$40 billion fundraise that the company is trying to close that we've talked so much about on the show that will very much go to good use. That story is live on our site right now, and I encourage you to check it out. A couple headlines for you. Meta and Microsoft reported earnings last night. I want to give you the one number that you should know from both of those prints. For Microsoft, the big number is$75 billion. That is how big Satya Nadella said Microsoft Azure's annual revenue was over the past year.

2:08Now, this is the first time that we've learned at all how big that business is from the company. I mean, Microsoft has typically only told us what the growth rate is for this business line called Azure and other cloud services. That business line did have its second straight quarter of acceleration. And so it is growing. Investors were cheering, but it was finally the first time that we got a sense for how big Azure is on its own. For Meta, the number that you need to know from earnings is 22%. That is how fast revenue grew in the quarter. And crucially, it was well above even its own projections.

2:44The company is obviously spending a lot on AI right now. It's very expensive. And so the fact that its top line is growing and is booming is a good sign for shareholders means they can fund all of that investment. Last but not least, we are watching Figma shares. The company is debuting today. It is a marquee IPO for Silicon Valley. We'll have more on tomorrow's show about how the stock closes, but we are eagerly waiting to see how it trades. Okay, our first guest, Arvin Jen, has quite the resume. He spent 11 years at Google, and he was a distinguished engineer. He then co-founded cloud security company Rubrik, which went public last year.

3:23But for the past eight years, he has been working on Glean. The company is building an AI platform that is supposed to make it easier for businesses and their employees to search through all the hordes of data that they have across all their various applications. The company was valued at$7.2 billion last month, and Arvind is here with us today. Arvind, welcome to TITV. Thanks for being here. Thank you for having me. So look, we've written a lot about Glean at The Information. We have dedicated entire stories to the ways in which companies are actually making your life a lot harder by restricting Glean's access to data.

4:02You know, one of the stories that we wrote was that Salesforce has limited Glean's access to Slack data. I mean, you know, I want to ask you directly, give us the view on the ground. What is going on here? Well, when you think about the AI industry today, a lot of AI companies that are serving businesses, they need to access business data to add value to their customers. And that's what we do as well. Our company, as you mentioned, Glean, we connect Glean with all of the enterprise systems. And then we build an AI assistant, a chat GPD, or cloud-like experience over your enterprise data for our customers.

4:47So for that, we need access to all the information and data that lives in systems like Microsoft or Google or Salesforce or Classian. And what you see in the stories is the questions emerging today about who owns that data and do companies that are actually hosting this data on behalf of their customers, should they actually keep it open for other companies like ours to also work on to add value for those customers. And so that's the crux of the debate here. And so where do you fall on that debate? Yeah, well, for us, we believe that customers, they own all the data that they host in each one of these SaaS applications.

5:49and they have all the right to actually do whatever they want to do with that data. They have the right to take that data and make new AI applications work on it to add value for them. And I think that's the right thing for the industry moving forward. I mean, if you think about SaaS as an industry, a lot of success for SaaS systems came, like how it was able to replace the previous generation of on-premises system was basically the openness of the ecosystem. Every SaaS application interoperated really well with other systems, which actually created a lot of value for customers, which in turn then pushed customers to adopt these new applications.

6:31And the same thing is happening with AI. Today, customers can actually have really amazing experiences, a lot of automation AI agents that they can build so long as they can actually have, you know, all the startups sort of like, you know, be free to work on that customer data. So what's right for the industry is, you know, is being open and being interoperable. Right. Have you talked to Mark Benioff about this and told him this? We haven't talked to him directly, but, you know, we talk about this publicly all the time. I mean, this is like, you know, more like, you know, we talk to our customers and they tell us that, you know, it's really important for them to have access to their own data, you know, with no restrictions.

7:19Right. And, you know, I guess what I'm trying to get at, you know, there was a comment actually on the story that we wrote about Salesforce restricting the Slack data access. And the comment was, we are moving into the gloves are off phase of AI. And so, you know, to your point, to this debate, I mean, how do you think this gets resolved ultimately? Well, ultimately, the customers decide, you know, what happens, you know, in this particular area. If you think about what they need, customers are looking for free access to all of their information. is clear to them. There is no confusion in their mind.

8:03You know, we talk to a lot of large companies. We work with a lot of large companies and they believe that, you know, they have full rights to their data. And they also have a choice. Like if you think about the tech industry, you know, we are a hyper-competitive industry. There are a lot of players that provide same, you know, kind of products and services. So if I'm a customer and, you know, I'm looking for a specific product and two different vendors, and one of them, you know, it's going to actually restrict me from accessing my own data. Another one that doesn't, it's the choice that's clear to me in terms of, you know, who I'm going to go with.

8:38And so that competition itself is going to actually force, you know, that openness, you know, in like how all of these applications are built. And it will basically put us in the right direction, which is that, you know, the enterprise SaaS ecosystem will remain open. Right. You know, speaking of enterprise software, you know, you have built an enterprise software company, you know, before sort of AI became very popular in Rubrik. And so you have sort of an interesting perspective here, you know, having been in both worlds for so long during two separate eras. This narrative here that AI is going to eat enterprise software.

9:22What's your take on that? I actually don't believe in that as much. I mean, like, look, AI is a big enabler for all of us, including all the existing software SaaS companies. Think about, think about like, you know, where people spend time today. You know, if you are an engineer, like, you know, you spend time in engineering systems like GitHub or Jira or Confluence. If you're a salesperson, you spend time in CRM systems. And so the existing application vendors, you know, they have like a lot of amazing stuff going for them. They have a lot of valuable data from their customers. They have the attention of the users.

10:04Users spend a lot of time within those applications. And I think the opportunity for them to leverage AI to create better experiences is right there. So I think for me, it's not about, well, the existing companies are going to become irrelevant and new AI companies are going to be the ones that are going to become more powerful. We don't see it that way. The thing is who's going to win our companies that are able to leverage AI today to create more powerful experiences. And you have to remember that people don't change that fast. People don't change their habits that fast. So even if you build new products, it takes a long time before you can actually start to have users change their habits and start using you and change their workflows.

10:55So to me, the opportunity for every existing large company is huge. They can keep that engagement that those users, so long as they're actually building great products. And so what I hear you saying is that, yeah, AI might make it easier to develop some of the software on your own, but it will take a very long time to actually adopt any of that software or have people learn how to use it, essentially. AI is going to make it easy for you to customize things for you, but I think the core systems that technology companies build, those are hard to build. So AI doesn't necessarily make them that easy either.

11:34Right. So you are, I think, eight years into building Glean now, if I'm asking right? Six years. Six years. Six years. So six years into building Glean, you guys have been able to raise a lot of money. You have good top line growth. What do you see as your exit strategy for the company? We don't have one. We don't think about exit strategy. This is the time to build. we feel quite fortunate that we built a product that all enterprises in the world are looking to buy and so our business is growing, we're doing well there's a lot of things to build and we see a vision for the future which is very bright for all of us as humans as workers AI is going to actually really fundamentally change, you know, like how we work.

12:32And in the future, you can expect a lot of like boring work that takes a lot of time for us is going to be done by AI. We're all going to have really amazing personal assistants and companions that are going to help us really sort of dramatically change how we work and how productive we are. So we're right in that mission. And so the plan for clean is to keep building towards that future vision of AI at work. Great. Well, thank you, Arvind, for coming on the show. We really appreciate it. I think you have a very interesting perspective having been in this space for so long. And I think your approach to the corporate data wars is one that we are all watching.

13:11So thank you so much for joining us. We look forward to having you again on the show. That is Arvind Jan, the CEO of Glean. Our next guest has the exciting job of managing a venture fund that is closely affiliated with one of the most successful athletes in the world. Serena Ventures is the venture fund started by Serena Williams. And a little over a year ago, the company brought on Beth Ferreira as the general partner. The fund invests in everything from edtech to Web3 to fintech and, of course, sports and wellness. And Beth joins us here today. Beth, thank you for being here on TITV. Thanks for having me.

13:47So, look, I just got to ask, do you play tennis? I don't play tennis. You don't play tennis? And it's a mediator, but I do not play tennis. I mean, have you played any kind of a sport with Serena Williams at all? Not with Serena. Not even pickleball? Not even pickleball. Not yet, but I'm sure that will happen. Okay. I guess we don't even know if she plays pickleball. All right. Well, look, let's talk about the fund. You've been at the fund, you know, formally in this role as general partner for a little over a year now. And, you know, I have some specific questions for you, but, you know, just give us the pulse check here.

14:23I mean, how has the year gone for you? What have you changed about the fund in the last year? So overall, it's been great. I would say from the changes perspective, not much. We have stuck to the original thesis, which is all about cultural demographic and value changes and where technology intersects there. So think access, opportunity, and ownership. And so that's been the lens that we've been looking at. I'd say we are investing slightly later stage. So we did a lot of pre-seed before we joined. And when you think about the cultural relevance and where Serena sits in the ecosystem, it's just hard for us to insert that value at a company so early.

15:19So we need to get too close to product market fit before we can really insert Serena in really valuable ways, which is whether it's amplification or connections to key partners or customers or things like that. That's where she sits gives us sort of the magic and accelerates our business. And you kind of hit on this point, which is exactly what I wanted to ask you about, is how do you think about the competitive advantage for a fund like this? Because obviously you've got Serena's name and you've got this high profile sort of name around the fund. But I mean, how do you think about the strategy and where you can actually stand on the investment side?

16:07Yeah. So, I mean, it's not hard for us to add value. We have a team that has deep operational and venture capital expertise. So we have the base there. But then we have something sort of magical on top of that where Serena not only is involved on a day-to-day basis and all of our companies have a direct line to her, but she's also in a place where she opens doors at a different level than most of our competitors. And so I'll give you a for instance where we had, you know, we're an investor in a company that had, you know, not just one, but multiple large Series A players around the table. You know, for three months, they were saying that we're going to make an introduction for the company.

16:50And finally, I was just like, OK, Serena, this is our this is our place. This is our turn. By the next day, she had not only the introduction and the meeting set up, but the following week she joined the meeting. and help them get that customer over the finish line. Got it, got it. And talk to me a little bit about her vision for the fund. I mean, you guys raised, I think this was before you were formerly involved in the fund, but you raised a fund a couple of years ago. I think it was a little over$200 million. How much of that fund is deployed now? So the fund was raised in 2022 and it was$111 million.

17:32$111 million. Yeah, so she likes the number one. So there's a play on that. Right. I mean, I think where we're thinking about the fund is, and the reason why I came on, was that her aspirations are not small. So we're thinking about a much larger platform and where we can leverage all of our experiences and competitive advantages across many areas. Have you guys had any exits? We have not had any exits in the portfolio to date. But soon to come, maybe. At least soon to come. Look, I want to, you know, go back to sort of your career building in venture. You have actually built a venture fund on your own a little while ago.

18:24You built the venture arm for William Morris Endeavors. That was about 10 years ago. So what has changed about building a venture fund in the last 10 years compared to when you did then? Yeah. So when we were building WME, one of the key messages that we were sending was that we believed that founders were more exacting about who they wanted as partners. And that brand, you know, large brand, while important, wasn't the only consideration for founders. And fast forward 10 years with the democratization of information and people talking about what's been working for them and what's not been working, we see that as a competitive advantage for us.

19:14So we can come in and we can add value from the first moment we're in the company. And founders see that. And so that gives us a competitive advantage when, you know, if we miss something and we're at the tail end of a round, we can get those companies to either take more dilution or open up the round to get us in. We can, we have more success, you know, cold emailing or reaching out to companies that we do is super interesting. So then sort of create rounds for ourselves. And so that's just part of the whole competitive advantage. And, you know, very quickly before we let you go, I mean, you know, you talked about the ambition for the fund to sort of move a little bit later stage because that's really where Serena's profile, you know, can really help sometimes open some doors.

20:10So, are we going to see Serena Ventures start to compete for some of these very high-profile AI deals that are now starting to get raised in the coming months? So, compete, yes. But we also, given where our value proposition is, we can play very nicely with some of the large players. So, you know, we are very focused on sales, marketing, communications, whereas many of our competitor, our large competitors in the market do not have that expertise or that's not where they're leaning. Right. And so, which is great because they have expertise in other areas that they can add value to the company and we can sort of insert ourselves in to those rounds as complimentary versus competitive.

21:03So when you say play nice, you actually feel you can be more competitive in those deals because of this communications marketing sort of focus that you bring. Yeah. And along with all the other things that we've just talked about. Right. Right. Last question for you. I mean, the fund was raised three years ago. Are you guys planning to raise another fund anytime soon? We are. So we hopefully will be out shortly. When can we expect it? Probably the fall. Okay. And do you have a number that you're going towards? Not disclosed yet, but it would probably be a little bit bigger than the one that we had.

21:42Great. Well, I'll tell you what, Beth, when you close that fund, come back on the show, and then we can ask Serena what technology she's also obsessing over these days. That'd be a lot of fun. Yeah, it would be. That is Beth Ferreira from Serena Ventures. Thank you so much for joining us on the show. One of the hottest areas in AI right now is video and generative media. And this week, another company in that space raised another six-figure sum. FAL, which stands for Features and Labels, raised$125 million from Meritech Capital was the lead investor. The company's round was also supported by Salesforce Ventures and Shopify Ventures.

22:21Shopify Ventures, I should say. Joining me now is the co-founder of the company, Gorkum Yortsevin. Welcome to the show. hi atash why are you i'm doing well thanks for being here yeah thank you for having me it's great to be here so look gorkum i i have to be honest i i tried for about 45 minutes last night to really understand what the company does and there's a lot of big words on the website and so the way i think of it is you know generative media which is video and images what does your company to just help us understand. Yeah, that's correct. So we are an infrastructure provider for generative media.

23:02And so if you are building a product using an AI video model or AI image model, you use our platform and use our easy to use APIs to build on top of some of the models that we have. And on the other side, we work with some research labs to provide them infrastructure. And the way we do this is we use our inference engine to speed up their models and expose them as easy to use APIs for developers and enterprises building on top of these models. So you're not actually like a runway or a mid-journey where consumers are using your app to generate the video. You would be selling your technology to an app that hopes to allow consumers to make the video?

23:50All our revenue and customers comes from B2B. So we sell to other companies, other applications, some of the biggest enterprises and some AI-native startups to build products on top of the models. So it's purely through an API. And tell me, I mean, who are you selling to right now? We have some of the fastest growing AI native companies, but also some of the biggest names in e-commerce and retail. For example, we work with Shopify, Canva, Perplexity, I can count as one of the AI native companies. And then we have really interesting companies like PostRam Partners in the architecture space. And recently we announced a partnership with Adobe as well.

24:41obviously one of the giants in the design and productivity industry. So we have all sorts of companies that we are working with who are utilizing these models. So the thing that I'm trying to figure out is you're working with all these big companies. I mean, you make it easier for them to build their AI video and image generation tools. Why do they need to use you at all? I mean, I've got to imagine that Perplexity and Adobe and Shopify have their own teams that know how to use all these image and video models on their own. So, I mean, I guess I just, why do they need to use you at all? Yeah, exactly.

25:21So what we've been building for the past few years is what we call an inference engine. So we are able to run these models a lot faster than if they were to run these on their own. So that's definitely one of the big plus. And then we also have a big range of models that are production ready from day zero. The industry right now is really fragmented and there's a new model every single week. And people want to be using the latest and the greatest models all the time. And our platform makes it really easy to utilize a wide array of models and change it over time as the new models are released. Got it.

26:05You guys raised$125 million. What was the valuation? Valuation was$1.5 billion. $1.5 billion. And how are you guys doing in terms of revenue right now? We are approaching$100 million run rate. I think we closed July, close to 95. So we'll make sure to do a big celebration when we reach 100. But we are almost there. So$95 million, that's annualized revenue. So that's monthly revenue annualized? Correct. I'm just multiplying the revenue we made in July by 12. Right. I'm calling it run rate. Great. Well, and very quickly, I should say the growth for that, where was your run rate at the start of the year?

26:51At the start of the year, I think it was around 25 maybe. And then last year around this time, I think it was a couple million. Got it. Well, I mean, 25 to 95, that's a pretty good growth rate. So congrats on that. Congrats on the round. Thank you for coming on the show, Gorkum. We're really excited to have you and look forward to having you on again. That was Gorkum Yurtsevin from FAL. Google just named a new chief AI architect last month. Corey Kavakshalu, who is also the CTO of Google DeepMind, has quietly amassed quite a bit of influence inside the company. and my colleague Erin Wu wrote a great profile about him that just published this morning on our website.

27:34I want to bring on Erin. She is our Google reporter. She joins us from San Francisco. Hi, Erin. How you doing? Hey, thanks so much for having me on. So before we get to the profile, I mean, did I say his name right? You were really close on saying his name. Oh, man. Okay, how do we say it? Kare Kavakcholu. Kavakcholu. Okay. Okay, got it. Because you did put the pronunciation actually in the story. but it didn't have the emphasis on syllables. So that's what I'll have to work on. He's from Turkey, I think, is one of the things you mentioned in the story. Okay, so explain to us who Corey is and why he's so important.

28:11Yeah, so he's a really long-term Google DeepMind researcher. He was actually at DeepMind even before it was acquired by Google. So he's been there since 2012 and he's really risen up the ranks. Most recently, he's CTO of Google DeepMind and he is chief AI architect for the whole company. And so in that role, he's really become like the operator, the fixer, the like guy who gets things done, who's in charge of organizing all of the different disparate teams working on Gemini. He has more than 2000 people reporting up to him. And he's the one who figures out if a researcher needs to be moved to a team, who to prioritize, like where to put compute.

28:46And so he's really become essential to the Gemini operations in a lot of ways. And so why did he get this promotion? Yeah, so I think how we can think about this is Google is moving into this new phase where it still really cares about the models, obviously. That's really important to it. But it's also trying to figure out how do we put this into products that people are actually using. And so in the memo announcing his new role, Sundar was talking about we want our products to evolve just as quickly as our models. And so that's really why Cori got this new role because he's seen as like the operator guy for the Gemini models, and now they're hoping he can do the same for all of Google.