E219: Groq valuation projection; OpenAI $12b ARR; Lovable 100x revenue growth in 8 months

0:00All right, fellas, Grok, my favorite private semiconductor company. Okay, Grok makes semiconductors for inference, which is like running AI models. They just did around$600 million at a$6 billion post valuation. This is what I want to talk to you guys about. Where do we think this company can go? Okay, so let me throw some numbers at you. 2025 forecasted revenue is$500 million. 2026 is$1.2 billion. 2027 is$1.9 billion. Okay. This round was oversubscribed. So 600 million on a 6 billion valuation is a 10X revenue multiple. I'm sorry. I think that's a little low. All right. For a semiconductor company.

0:46But if you take a look at the 1.9 billion revenue that they're forecasting in 2027, if they can pull that off at a 10 times revenue multiple, that's a$19 billion valuation. right so that's not bad that's 3.2x in two years time so where do we think these guys could be in like five years time and maybe maybe maybe if we don't have to talk specifically about grok maybe we're just talking about compute on a macro level right um for as it relates to ai because i think some stuff's pulled through on earnings and everything but evan why don't we start with you, man. What do you think about Grok? What do you think about compute, AI demand?

1:28Is this an interesting place to be investing? What are your thoughts? It's absolutely an interesting place to be investing. I mean, this is where the picks and shovels of what is likely to be a technology defining industry are being built. I still think it's quite early. Grok seems to have some excellent technology. Batting on them today has probably some has really nice upside. But it's I would say it's limited to the sense that there is still a lot of competition to be had. There's still a lot of uncertainty on where inference will go. But they seem to have great technology. And I think they've been executing and firing on all cylinders.

2:08I think that there's so much more demand to come in this market that has yet to be even really even, I don't even know the word, like thought about. So I think that their projections, I agree with you, Aaron, are kind of low. I think that they probably are setting them low so that they can crush them. And it seems like a pretty nice opportunity to get in on the ground level, knowing that if you've got to be a long holder here, in my opinion. Okay, got it. Okay. So Clint, maybe we'll kick it over to you. So Evan likes But maybe you could talk just about on a macro level, what you think about compute and like demand for compute.

2:45I mean, I love it. I think this is where the money's going. Right. As we as we say on this call, a lot of you follow the money. And if you look at AI and where the money has to be spent in order to scale AI, chips is one of those areas. And you have to bet that it's going to be moving in terms of market share towards inference over time from just raw compute. So I'm really optimistic on Grok, really optimistic on the space in general. And I think when you see NVIDIA trading at over 30 times sales, really dominating from a monopoly perspective, if you think about chips today and really being the clear benefactor of all this CapEx.

3:30I mean, I just saw that just in 2025, Meta, Amazon, Microsoft, and Google Alphabet are going to spend over$400 billion on AI. The CapEx is real. And I think when you look at where that's going to go and going to continue to go and where it's going to evolve to go, you have to think that inference chips are going to be a benefactor. And so I agree with Evan. And I think they're probably low-balling estimates, right? You've got to manage expectations well. But I think there's tremendous upside here. I think it's an asymmetric bet when you look at a company like Grok because the valuations just not where a name that is expected to grow a lot.

4:17It should be at this juncture. That's probably why it was oversubscribed. It's a good deal. Yeah, it's a good risk-adjusted return. Yeah, there's asymmetry there. Yeah, got it. Okay, so Nick, let me ask you this, man. So like it seems Evan Clint bullish on these types of this type of play. But like I know data center capacity, right? Like you can't build them fast enough kind of thing. And then also like access to electricity. That's the other thing that people are talking about too. Do you think that that's a risk for Grok? Or do you think we'll kind of just figure this out on a kind of national level or global level?

4:52You know, these constraints will kind of be worked out. Like what do you think? I think everyone's working with the same constraints. Maybe some of the bigger players will have better access or can negotiate better terms. But I think they're just looking at the way they've structured this deal. It's priced in. Some of the risks are priced in. So the fact that maybe they won't get as good of a deal as someone that's further along or raised more capital, that risk looks to be factored into the discount. Now, the illiquidity, when you've made the comparison to an NVIDIA or even any of the other AI players that are going to take the same wins, if everyone's pushing more compute, more inference, more everything, everybody kind of wins in that same ilk.

5:39That seems pretty fair, too. And then even the fact that for a lot of this raise, they only took$600 million on a$6 billion, so very crude is 10%. That's not brutal. You know, that seems pretty beneficial for the company. And it's almost like we know we could have taken more. We're drawing a nice line in the sand to show everybody we know we're worth it. We're oversubscribed. So we're going to shout that from the rooftops and then carry on. So I think this is a pretty cool, you know, risk adjusted place to put your bets. All right. So a unanimous over like bullish outlook on Grok. I mean, this to me, like, Clint, I'm right there with you.

6:22I know you're like our picks and shovels guy, but I think everybody kind of agrees with this too. Like this just, the risk adjusted return profile on these types of companies is just awesome. And there's not that many out there. That's the other thing that's in the private space, right? There's not that many out there. So to the extent you can get your hands on them, I think it's an interesting place to look. Okay, let's talk OpenAI, fellas. They're crushing it, crushing it. Okay. Check out these numbers. They came out with a whole bunch of numbers this week, right? Annualized revenue hit$12 billion.

6:57It was 4 billion ARR in December of 2024. That's nuts. Come on. That's$8 billion, right? They were, okay. And then 700 million weekly active users. That's up from 500 million in March. That's 40 % growth in four months. They have 5 million business customers as of July. This is Brad Leich at the chief operating officer of this. They had 3 million at the end of June. That's nuts. That's a 66 % increase in one month. I don't know about you. That's like the definition of product market fit, right? That's insane amount of growth for business customers, right? Right. Okay. And then on the capital raising side.

7:48Evan still looks dour, by the way, for anyone listening. Just the facial expression is, I'm suspect. Okay. All right. And then on the capital raising side, so they were going to come out and do 40 billion. I think they're done. My math says they're done. They raised 10 billion earlier this year. They just announced this past week they raised another 8.5 billion and SoftBank was coming in for 22.5. So I don't think SoftBank participated in this 8.5. They might've done something in the 10 billion earlier this year. So maybe they're not quite there yet. But nonetheless, they've got most of this 40 billion that they were looking to bring in by the end of the year.

8:30And there's some hair on that SoftBank money too. They got to do the for-profit conversion, which maybe we get into that, maybe we don't. But anyway, I got these guys at a$326 billion valuation. I'm seeing them trade up in the secondary market from the convertible note value, the$300 billion kind of post-convertible note value. So thoughts on OpenAI. Nick, maybe we start with you, man. What do you think of this company? It's kind of shaken out that it's like you got Anthropic and Claude, you got XAI, maybe you got some Google Gemini, and it's OpenAI. It's like kind of four of them. I don't really hear a lot of other players for the large language model space, right?

9:10But what do you think of these guys? How do they stack up? What do you think about their future prospects? Sure. Just the hair on it. And I know you said, oh, maybe we'll touch on it. But the conversion from nonprofit to for-profit, the involvement that I really feel Microsoft has their hooks into these guys way deeper than they ever could have expected. That all adds up on my risk profile. And I know we talked about perplexity and anthropic a little bit more last week rather than just the pure comparison to OpenAI. But if I were going to play the sector, I would want to have a bit of each. And actually where I think I'm going to make a better return is probably a perplexity and anthropic relative to OpenAI.

9:54But they're the mainstay. They're definitely bigger. They have a lot more of a captive audience. I never would have expected that 60 % jump from three to 5 million for corporate. That's really impressive in a surprising way. And every time I talk to my close friend over at Google and I say, what about these guys? He wants to dismiss it. He wants to not account for it. He wants to tout how amazing Gemini is. But this is, I'm not going to say an unmitigated threat, but this is certainly the one that could really displace and start to shake up everything that's been happening at Google a little bit more than the others and already have.

10:39So I'm really impressed that they continue to grow at the pace they have. So yeah, they're the ones in the sector you got to watch out for. Okay. So Clint, this type of growth is like, this is a lot of growth for a customer base that's huge. Right. And also to the revenue, it's like translating into revenue growth too. They're not like giving this away. Right. People are paying for it. And, and like, look, there are like 25 times revenue multiple right now. Right. That's the 12 is my number, right? Hold on a second. Yeah, 25 times revenue is$300 billion valuation and$12 billion in revenue, right?

11:21So perhaps by the time the year end hits, they'll be even north of kind of$12 billion ARR. I mean, that's a pretty attractive valuation multiple for a company that's growing this fast, right? I mean, I know it's a lot. $300 billion is a lot. But what do you think about the valuation? Is this like an interesting entry point for OpenAI? Or is it just too expensive, right? Yeah. I mean, I don't think it's too expensive. I just, I think you have to truncate your expectations for 20 X returns at this level, right? I don't think the asymmetry is there, but I think the serious, I think the, the serious opportunities in the business, the B2B side here, I think Microsoft, although they do have their hooks in because early money is expensive money.

12:05I think it's a strategic advantage of them or for them from a data perspective from a getting because of the business to business functions with office and how Microsoft is got significant penetration of that marketplace. That is a natural growth trajectory and growth area for open AI. And I think I mean, I use it all the time. I use other ones too, but for the majority of my work, the open AI pro with the agent mode is insane. And as they continue to add agents, it is, I mean, it does literally everything I need it to do. And not flawlessly, but almost flawlessly. Right. And that's everything from marketing copy to write market commentaries to me uploading every due diligence document I've ever put out and then helping it create investor memos from new subscription documents or private placement memorandums to writing code so that I can develop a form that then spills out an output from conference board data.

13:12Like it is, it's literally saved me a ton of time, saved me a ton of money from a business perspective. And I just think when you multiply the productivity gains that this engine can create for the businesses that partner with it, I think it's amazing. All right. So you're, you think the growth is the revenue growth. It's going to be there. It's going to keep going to be there for a while. And it's all going to be in the business segment. Okay. In my opinion. That's right. Got it. All right. So Evan, so then this, like you, you eyebrow raised, I saw it when I was quoting those business numbers, right?

13:4666 % growth in one month is nuts. That's nuts. Agreed? Sure. Yeah. Okay. And then, and then the question is, is like, they, they must have, they must have done something. And Lightcap actually noted in his kind of remark that he made, he's like, you know, we've, we're working really hard with businesses to get them going. So something magical has happened between OpenAI and their business customers. I don't know what that is. I can probably ask around and find out. Go ahead. Do you think that could potentially just be end of trial or end of credits running out, and that's going to give you your big pop?

14:24That could be as well. And that's not what you're hoping for as an investor. You want this to all be super organic. But it's kind of like when Netflix shut all of your cousin's accounts down. because they were sharing. Well, sorry, I didn't mean to jump over Evan, but there's definitely a lot of that in play, right? But Evan, like when I last time I took a hard look at this, right? It was like a very API driven model with the, for open AIs, like B2B play. But now that to Clint's point, they have all of these other solutions and kind of tools and user interfaces, right? Which I'm sure are available to their business customers.

15:02Is that what's pushing this over the top, do you think? Or like, Like what could be the driver of such a like huge amount of growth for open AI in their B2B channel? That's a great question. And that's the question that causes me to raise my eyebrow. Nothing just happens for no reason. You don't just wake up one morning and there's 2 million new business customers. Sort of what Nick was saying. Something has to have happened. It could have been, you know, subscriptions lapsing. It could have been the agents, which I know. So like a technicality, right? Yeah. I got you. Or the fact that maybe they added a subscriber field, which says, are you using this for business?

15:41And a bunch of people just say yes. Sure. Okay. Sure. Fair enough. To me, and all four of us here have our ear to the ground of what happens in the business world. I haven't seen anything in the last two months that would make me be like, oh, well, that explains that. And so to me, it makes me feel like, and because he specifically didn't cite, and I'm talking about the COO of OpenAI, And I specifically didn't cite any reason for why that happened. It gives me pause to be like, well, maybe this is just, you know, I've worked at companies before. There's lots of ways you can spin news to make it sound better and look better than it may be because you don't have all the content.

16:17So, look, all of this to say, I still go back. I think the product is phenomenal. And like 700 million weekly active users seems like that's probably going to double in the next year or two. so yeah so i don't think that you know i just found the the news to be a little surprising and maybe they're doing it because they see such maybe that news came out because there's so much hype right now around claude code and claude just raised this big round and they got this like they've got the the the engineering business model like really toned in and they just launched their financial services and maybe they were just like well open ai needed to make some noise and this This is how they do it.

16:53Splash, yeah, right. Yeah, fair enough. So I don't, I mean, I think there's a lot of opportunity for all these companies to continue to grow. I think, you know, we're going to see there's billions of people in the world. I would expect at least half of the population of the earth to be using AI on a daily basis sometime in the next three years. So like, it's just like endless opportunity and upside. And the race for the most effective model will continue to go on, the most efficient model. So there's just a lot of upside of all the stuff. And I think for 326 billion, it's probably still a good, so you still got upside, but I agree with Clint.

17:27It's not asymmetric at this point.

17:32Right. Okay. Gotcha. All right. Listen, Evan, you can, you can do due diligence for me at any time. You have a healthy amount of, you know, you know, you just got to be sold. You need to believe. I just, I appreciate that. I appreciate that a lot. Right. He's our BS meter. That's great. That's right. No, it's just like every time I bring somebody, he's like, I just don't know. Let's just get into that, fellas. I think we're all business operators here and we know how much work goes into moving every metric up and down even a point. It's true. And so when I hear someone say, we just gave 2 million business customers overnight, I'm like, that doesn't just happen.

18:09Like, I'm sorry. It doesn't just happen. I know. I just love it. I love how it's not even like, wow, that's amazing. Congratulations. and then you're like but like can we get in and like really digging in on that number please you know it's just straight to the i don't know what come on well early in my career early in my career i was a partner at a venture fund and we were a seed fund and all right like 90 of my job was looking at a deck and being like well what is this metric and what is this metric and why does it like how did you get here because you're a seed company how are you profitable and it's like oh because you're not paying for office space you're not paying yourselves you have no employees oh okay, they're probably profitable.

18:45It's healthy, man. That's healthy. I like that. I like that a lot. I like that a lot. Okay, let's talk about Lovable. So this company is out of Europe, okay? They are an AI app, what I would call an AI app, all right? So they sit on top of other AI models, and they basically help make applications, right? So this company's gone from a million in ARR to 100 million in ARR in eight months. That is insane amount of growth. They're labeling it as the fastest global startup that's ever happened. They have 10 million projects being built. There's 100 ,000 new applications that are being created on this thing every day.

19:35That's insane. And now they just launched this lovable agent, which is helping people, you know, just like OpenAI and everybody else. These agents are like checking and double checking and everything that the apps, the applications, so that there's less errors and more efficient. So I read, I like got into these guys a little bit and they did it. They did a round in July, like last month, right? 200 million at a$1.8 billion valuation. This company has been around for less than a year. So I got into this a little bit and like kind of like what's the big deal with these guys? Like why does everyone love them?

20:08It's not that people are like building an app and like launching it and like starting their business on it. The thing that I've heard that people really love about this is like product teams. Like I used to be a product guy, right? And Morgan Stanley and TD. So like product teams are building actual apps themselves, usual functionality and everything else. And it basically like works. This is where I want the button. This is how the workflow I want it to look like. And they're like negotiating all that between themselves and other stakeholders before they hand it off to the development team. and then the development team can obviously code it up in a proper way and then launch.

20:50So the cycle to deliver new applications is accelerated dramatically. So I find that to be a very interesting use case because as a technology company or even any business, you'll be able to accelerate the release of tech and applications, which should help with efficiency and I think drive more net income. So Lovable is a new company, so I'm not going to assume that you guys know what they are. But what I mean, this is start. Maybe we can just talk about these guys in the in the context of like AI apps. Like, you know, that's interesting. This is interesting. Like if we see more companies like this.

21:30Right. Where they're building on top of other AI platforms, that could be, I think, super compelling. These are super compelling investment opportunities. Right. I mean, what do you say? Maybe we'll start with you. Yeah, trading opportunities. Okay, tell me more. Yeah, so ultimately, I love this space because it reminds me of a quote, and I forgot who said it. And it was essentially, AI is not going to take your job. It's people who use AI that are going to take your job. Yeah. And I think there's a lot of truth to that in the near term as AI continues to iterate and evolve. You're going to have companies that emerge that provide this user layer, right?

22:14This is like a user layer on top of AI that allow product teams to get products launched quickly, allow people to create apps, allow people to create websites. I went on and kind of looked at like, you can create me a marketing page or create me a landing page for this marketing campaign and it creates it. Right. Right. That is awesome. But eventually all of this will be done within the large language models that remain by themselves. And so you have a finite runway. And do you have a big enough established moat to where the substitution is not going to eat your lunch at the end of the day? You might be the fastest startup ever now or whatever they call themselves.

22:55The local startup. Yeah. But ultimately, these type of businesses without major moat development are going to, just as fast as they rise, they're going to fall just as fast, in my opinion. So I think they're a trade. So it's like a victim of their own success. They're so successful that these large language model platforms will just come and develop a solution. You're going to see one person businesses, in my opinion, that that reach over a billion dollar valuation, the age of AI, because you've never been closer to idea to creation in terms of that timeline being squeezed that you are today. And so ultimately it is it's just a matter of time.

23:36Folks who are in an era where fart coin trades at a one point four billion dollar valuation. So you're going to have one person businesses with billion dollar valuations that are going to be layers on top of AI. But that doesn't mean they're sustainable and that you can bet on persistence. And I think when you make an investment decision, that's long term and you don't have liquidity, you're betting on persistence. And I think that's a challenge in these apps. OK, so I'm there with you, too. Like, Evan, like, what do you think of this? This is where I haven't put any chips on the table in the AI app space yet.

24:11Because to Clint's point, I don't know where it starts or stops with these large language model businesses. I mean, you may not know that for another two years or three years or so. Do you think that these AI apps are investable? Do you think that you've got to wait and see? Where do you stand on it? I think it's facts and circumstances. I'll take the other side of what Clint said, just for lovables. as just for the sake of the argument here. And I look at it like, Lovable to me has a lot of the similarities to a MailChimp, which took a business where setting up an email list and it allowed anybody basically a prosumer to come in and build a professional experience where they could write emails that look professional and send them out to people.

24:56And that business got sold from multiple billions of dollars to Intuit. And it was a fantastic business. Yeah, there were like 50 people there too. That was a good result, yeah. I think it really depends on who the customer is, is whether or not this AI business is defensible. Because I think if you're going to the real enterprise businesses, it's going to be very heavily competitive. Because there's the big incumbents that are going to try to come downstream. There's the startups that are going to go upstream, and they're going to fight for... And the contract cycle is just very long. But for Lovable, a business that I would say is much more prosumer.

25:28Tomorrow I wake up and I decide I want to go start at Chipotle for dog food. I can go to Lovable and just... generate a website and it would cost me very little. And maybe in the future, you're right. There's another competitor out there that does it cheaper, or maybe someone like Canva comes in and builds that product. It would, has their distribution. Sure. But I do think that like the prosumer market is a little, they're more price sensitive and they're less like their price sensitive to the fact where they're less brand associated. So they're willing to be like, oh, well, you know, this is the cheapest option.

25:55It does everything I need to do. Like, what do I care? Right. And so I think that there is a big opportunity there for AI businesses to win that market share, the customer who doesn't really care that much about the brand or how effective the tool is compared to the rest of the market. Right. So it just depends on your customer. Like if they're relying on it every day. And I will say the challenge with Hullable is that it builds you and designs you a website, but it may not run that website very well for you. Yeah, that's true. You might churn really quickly. Or maybe you start your Chipotle for dog food and it goes out of business in like six days.