We are headed towards the most epic run of IPOs of all time with SpaceX, Anthropic, and OpenAI all coming out within the next calendar year. What's going to happen? Let's ask the man that took Twitter public right after this.
Welcome to Big Technology Podcast, a show for coolheaded and nuanced conversation of the tech world and beyond. Today, we're joined by Dick Costolo. He is the former CEO of Twitter who took the company public in 2013. He's also a partner at 01 Advisors and one of the members of a new podcast, the Nick, Dick, and Paul Show. Dick, great to see you. Welcome to the show.
Good to see you. For clarity, I'm Dick in the Nick, Dick, and Paul show.
Oh, it's good to know that. It's good to have that clarity up from the beginning.
Thank you for establishing that.
It's a very interesting time that we're about to head into. We are in, right? But in terms of the financial markets, three major IPOs are on the way. SpaceX, OpenAI, and Anthropic could all come within, who knows, maybe even the next six months, which is crazy to think of. You were the CEO that brought Twitter public, and in fact, I showed up to San Francisco in May 2015. I got one week in and then week two you stepped down as Twitter CEO.
That's right. That's right.
And then away we went. So it's good to be... We actually spoke the morning after. We're going to talk about Twitter in a bit, but you're somebody who knows not only about the going public process, but how important I think the message is when you go public, how important that message is for what the rest of your company lifetime is going to look like and how you're communicating with your own team and company about it inside the company. Because an interesting thing about an IPO for any company, but particularly for tech companies and particularly in a bull market, is there's the time before the IPO where you raised your Series B and you were valued at a billion dollars and that's what the stock price is and what the valuation of the company is until 18 months later whenever and you raise your Series C and now you're valued at $2 billion. And so there's not really any concern about the stock price if you're a fast-growing company before the IPO. You're like, oh, my options are priced at a dollar and now the stock's $2. My options are priced at... now the stock's $3. It's all kind of fine, you know? And then you go and think about not just Twitter in 2013, but more recent examples like what Dylan Field had to deal with, I'm sure, at Figma and Cerebras now dealing with as well.
You know, it took us, you know, seven years, six years, seven years to get to the price of, you know, $30 a share and then in one day it went to $110 a share and then, you know, two months later it's back down to $40. So there's a roller coaster. You go from, yeah, whatever. It's not something you think about having to reinforce with the team or talk about with the team regularly to whiplash and, you know, what happened and why is the... So my advice to people who haven't taken a company public before and are about to do it is you need to prep the team for, hey, we're about to go into a world where the price of the stock can change even though nothing particularly happened today. Like the stock can go up by 15, 20% in value based on nothing or down 15, 20% based on nothing.
I think the public underrates how turbulent that could be inside the company. And this is kind of what I was getting at. The early days are really important in terms of the way that that turbulence actually lives out in your time on the public markets. So let me just make the distinction between the private markets where, you know, the numbers matter to a degree, of course, but they're not the ground truth, right? It's a lot of it is feeling and potential. You go public, those numbers matter a lot. You have to report quarterly. And for now, we can talk about it, right? So there's conversations that it might be once a half, but you have to report quarterly. And ultimately, even if you're not showing profitability in the early years, investors, they need to see a path toward it. And investors who bought in on the IPO promise, they need to see that promise live out. What are future cash flows going to look like?
Exactly. I remember before, a year before we went public, I was at a summit at GE, General Electric, and I was on stage with Jeffrey Immelt, the CEO, and he asked me, you know, how do you think about how you're going to manage your investors once you're a public company? I said, well, we're just going to continue to focus on the long term like we have now. And everyone in the audience laughed, you know, like, yeah, good luck. You know, let us know how that goes once you're a public company when all the, you know, sell-side analysts are asking you every two days what the next quarter numbers look like.
No, you're right. And so, look, I'll tell you what I remember covering because I was a Twitter beat reporter for a while and what I remember covering because I would go on CNBC and talk about Twitter earnings and I would always look back to the sort of the messaging at the Twitter IPO. Yeah. And now I'll caveat it. Well, no, I won't even caveat. I'll talk about it. What you said was, look, we see a path that this is a global service. We'll get a billion users. And what's the thing that dogged Twitter every quarter? It's like, you know, 200, 250, 300 million monthly active users. But I bought in at this promise that there's going to be a billion to a billion.
Now, this was what I was going to caveat, but now I'm going to sort of talk about... We could even have beat topline and grown, you know, bottom line more than we said we were going to grow it. But if we missed an MAU number, the stock would be down.
Exactly. And we could miss topline. I mean, we only missed topline one quarter by like 1%. But that quarter, MAU were up over consensus and the stock went up and by a little amount.
Yeah. And so here's the point that I was going to bring to it, which is the hype that you came out of the door with at Twitter is basically going to be nothing compared to the hype that we're going to see SpaceX, OpenAI, and Anthropic come out with. I mean, talk a little bit about what it's going to look like when these companies hit the public markets and then start reporting these quarterly earnings. SpaceX might come out 1.5, 2 trillion. OpenAI, Anthropic, certainly trillion each, at least. This seems like it's... they have obviously the potential to change software, we talk about that all the time, change tech, but until they do, they're going to be sort of measured against the expectations they... And that could be a very rough road on the public markets for a while.
100%. However, I think you got three very different personalities and stories associated with each company. And let me just say what I mean by that. Elon has the benefit of already knowing what this is like. And like him or, you know, like the way he runs his companies or not like the way he runs his companies, he does a great job of the narrative of, yeah, this is what's going on right now. But let me tell you what's coming. I mean, look at Tesla. You know, they can hit or miss delivery numbers and they can have 37, you know, 63 robo taxis on the road in Austin versus the a year ago number of a million, estimated a million. But the narrative is, yeah, but this is where we're really going and he's just out ahead of it with the public market and the retail investors and his fans and even the analysts and the sell-side analysts and the people who run hedge funds. So I think with Elon, even though the initial valuation of SpaceX as a public company, if you mapped it to the amount of stuff they have to launch into space and make productive, is going to be like, wait a minute, the math doesn't... that math doesn't...
I mean, forget Mars, they're going to have to get to Neptune to...
And build it. He just does such an expert job of... Yeah. But if you play this out 10 years, and I think he'll be... I think that... so I think he'll be actually fine even though there will be the people that are like, look, these numbers just don't make any sense compared to, you know, Starlink has to grow at this crazy amount over the next 10 years for this to make any sense.
Right. Okay. Before we move to OpenAI. How much money in the market is there in this belief that Elon will figure it out? Because so you have Tesla, right? So you have Tesla and there's this belief that he'll figure out robotics and then you have SpaceX with this belief that, I don't know, he'll figure out space. And I think SpaceX... I am one of the people that believes SpaceX will buy Tesla, right? And so Tesla's 1.29 trillion. Yeah. It's a... I get... I don't know, maybe I'm misreading the market, but I get like a trillion-dollar company on belief of Elon, but then if he buys Tesla, we're talking about a $3 trillion company in the belief that Elon will figure it out. Is that possible?
I think... so again, when SpaceX goes public, and I think it'll actually... you could... it could end up being worth over $2 trillion on day one. Whether that's a fake price or not and, you know, three days later it's something else, people are going to... there's so little float. My understanding is there's so little float and the float is, you know, how many shares are being issued, etc., as a percentage of the overall number of shares of SpaceX, and there's going to be so much demand that the day one price is going to go to some insane number. So, and where it puts out... I don't think is... I don't know. I just think that his ability to frame a narrative on a small float, you know, he doesn't need... he doesn't need, for example, a bunch of sovereigns to come in and put 400 billion each into the company. He just doesn't. So, he's been able to sustain that retail narrative with Tesla and I think it'll continue for some time with SpaceX. We'll see how long that lasts, but I do think he'll be able to do it.
Okay. What is the narrative before we move to OpenAI? What is the narrative?
I mean, it's just going to be, you know, Starlink is going to be... Starlink is going to power all internet connectivity.
So it's a Starlink story.
It's a Starlink story and then I'm sure it'll be data center space, which Sam has. And on our podcast we've talked about it's a long, long way off from being true. I think data centers in space is actually like full self-driving. If you remember Elon in 2013 said, you know, maybe it's 2014, by next year, your Tesla will be able to back out of the garage, drive you to work, pull into the garage at work, navigate downtown, all without you ever touching the steering wheel. He eventually was kind of right, you know, he was just like 11, 12, 13 years off. The data centers in space thing is the same thing. The amount of acreage of cooling systems you would need to launch into space right now...
Well, the argument is you don't need cooling because it's space.
Well, you actually do, but that's a longer conversation. The amount of stuff you have to launch into space right now to even power a one-gigawatt data center is something like... foots out to something like 16 acres of stuff in space. Also, so you can't just... the capacity to launch that right now is non-existent.
I don't see it as ever possible. I mean, a piece of space junk hits your thing, you have to send not just an astronaut, a fleet of astronauts. You think a data center never has a plug come undone because it wasn't like properly plugged in? It's like, ah, we have to send an astronaut.
Let's at least agree that it's at least 10 or 12 years away. But the story will be, here's what's coming. And I just think, look, if you would have asked me 10 years ago or how long are people going to buy the Tesla story where, yeah, they missed this, but this is what's coming, I would have said it would have fallen apart a little while ago and it just doesn't. So I think the same will be true of SpaceX for a while.
I have this theory that the Tesla stock will... or the Tesla shareholders who've been through this bumpy ride with Tesla, like they've been told a couple stories. Tesla, the electric car, the self-driving car, the Optimus robot. They will move their money, the retail in particular, will move their money to SpaceX.
I think that's probably true. Yeah, I agree. And then SpaceX will buy Tesla, probably.
Yeah. Okay. So, if Elon is somebody that can sell a vision, Sam Altman is not.
No, I don't think it's that he can sell the vision and Sam can't. I think it's that the narrative around OpenAI is already... Sam has gone out and written checks that Sarah Friar and team... not to put Sarah's... the CFO, so it's not just her... Sam has gone out and written checks that the rest of the organization can't cash, if you will. The compute commitments, the public commentary about, yes, we're going to spend this much on compute but we're going to spend three times as much on training the models. And, you know, when you go... I mean, there have been plenty of podcasts with OpenAI investors on all these topics, etc. But when you go and run the numbers, it adds up to crazy, over a trillion dollars worth of commitments that don't have the... and you don't see the revenue model or growth yet to back it up. So he's kind of mortgaged a bunch of the company, if you will, that now Sarah and the rest of the executives have to go out there with an S-1 behind them and answer questions, right? It's one thing for the CEO to be out there and again, Elon's got this history of I'm talking about things that are four or five years out. Sam has made monetary commitments, right? Here's how much so-and-so is going to put into the company. This is the deal for this particular data center. This is a deal for this kind of compute, actual dollar figures attached to them and commitments. So Sarah and the team now have to go explain to the public markets and get out there during the roadshow and, hey, help me, I'm not getting how the bill adds up here, right? I'm not seeing how the dollar commits over here are mapping to revenue over here and when we're going to see profitability. So I just think they're going to have a much, much tougher time. It will still do really, really well on day one because people are going to be like, it's the next... it's the next Stripe, it's the next Cerebras, I can't, you know, I haven't been able to buy secondary before, I want to be an investor, you have to own one of these things. But I think they'll be under a much bigger microscope quarter to quarter than Elon will. Is that fair? No. Is it probably what will happen? Yes.
So when I spoke with Sam about this, his perspective was basically like it's tough for people to grasp exponentials and we are in the middle of an exponential. So you might see our revenue today and you might see our planned expenditures. But what you don't fully grasp potentially is that we see it going that direction and it... let it... you know, that argument is a lot of people don't buy it.
Yeah. The counter on it, the counter to that counter and then I'll let you answer. No, you just keep doing the arguments yourself and then we'll end up somewhere.
This is typically what happens. I just debate myself. All right, but let's just talk about it. If you've seen this play out over the past few months, you've seen Anthropic hit capacity limits and OpenAI is starting to look wiser. Because it planned for that exponentially.
Yeah. So I think first of all, I think Sam is right that people are... I'll just insult all your audience right now and say I think people are generally innumerate and don't understand exponentials. And it's true if you look back even six months ago with the actual committed data center buildout, people were saying there's going to be way too many data centers. We're not going to ever use all these things. And six months later with agentic coding and Codex and Claude Code, etc., you're like, oh, yeah, we actually need more because these things to both do, you know, forward and backward work and the token usage has grown exponentially. And so he's proven to be correct. And yet the size of the commitments, the debt load that some of these people are taking on to fulfill some of these commitments and the gap between those commitments and current OpenAI revenue growth is, you know, he's just got a hard... Sarah and the rest of the executives in the company are going to have a harder time. Help me map how you get from here to here. I just think they will. They've got hard numbers associated with them that Elon doesn't necessarily have. Elon kind of does a lot of, here's where we're going to go and this is what's going to happen next and people are like, great, I get it.
Well, SpaceX has basically given up on the AI project and I know that might sound extreme but when you're renting data centers to Anthropic, they've given up. So is it possible with OpenAI's case that OpenAI really delivers on the product because it has more of these data centers but still doesn't get it right economically?
I think so. This is my admittedly opinion from the last couple weeks. But one of the things that I see happening that works in OpenAI's favor is Codex is getting more and more popular. I think if you asked people a month ago, you know, if you actually looked at Google Trends a month ago for Claude Code, like what Google Trends is, you know, show me the volume of search results for this term over time. Claude Code looked like one of these inflection points. And Codex is kind of bumbling along. In the last couple weeks, I've heard more and more folks say, we were 95% Claude Code and a couple Codex fans, and now there's a rapid growth in the number of people moving to Codex. So all of that works in Sam's favor and should that continue and you see a sort of, you know, balance of, I don't know, 50/50, 60/40, whatever, and with the growth in agentic coding across the board that's going to continue, that could be what saves... that could be one of the things that saves OpenAI.
Anthropic... I just think, I mean...
No, I think Dario... I mean, that's why I think they're three different things. I think Anthropic has just been, maybe as a result of just not being the lead horse so they don't have the spotlight on them, they were just more focused on, we're going to go attack the enterprise. We're going to focus on the enterprise. You know, they weren't even initially going to release Claude Code. I mean, the initial... they initially, my understanding from folks inside Anthropic, is we were just going to be like, wait, this is like a huge competitive advantage to be able to increase the rate at which we can accelerate our own work. And then decided, well, actually we should release it because people will find problems with it out in the wild that we didn't find internally and that'll make everything better, etc. There are probably other narratives internally but that's my understanding anyway. I just think their focus on the enterprise, their consistency of story and the public's perception that Claude Code is like... irrespective of what I just said about Codex two minutes ago, that Claude Code is like, this is the great unlock and going to be the real game changer, makes their story maybe the most middle of the road of the three where I think SpaceX will build castles in the sky and people will buy it and Sam will be put up much more against the, help me understand how this maps to this over here, and Dario's threading the needle between the two and will probably get a little bit more of a, this is the most consistent one of the three I can bet on for long-term growth and focus.
But the finances aren't going to look that different or they will?
I think they'll have big losses from all of them. I get it. My experience as a public company executive is that the narrative and the story and your ability to help the market think about the way you want to tell the story is just as important as the specific numbers in the quarter. And I think, you know, that's more true now than it ever was.
Does it matter then who goes first because you get a chance to set the narrative as opposed to...
Yeah. I think it would be great. I think it would be great for Anthropic if they can go first because for this reason, like we can tell our story, it's not going to be compared to what Sam says and we just saw in the opening bell on CNBC that they have this much, they had to spend this much in compute, why don't you have to spend this much in compute? You want to go first. I mean, I actually think part of the Elon lawsuit that we now all know the verdict of was, you know, when the verdict was announced, some people were like, oh, I'm surprised he pursued that, it was going to be that kind of obvious that he was going to lose. I think part of it was just going to throw some wrenches in the works over there and slow them down and I can get out first before any of these things and I'll be first out and I can suck a little bit of the air out of the room. Right. And that'll be great.
Now, you mentioned that most people are innumerate, so let's keep running with that.
Yeah. This is going to be the... unfortunately this is going to be the... I can already see the clip. This... unfortunately many, most people are innumerate and then someone will go find somewhere I made a simple math error and then they'll...
Oh, our team will do that.
But we're very generous with...
Excellent. Just kidding. We won't do that to you. So, but let's talk a little bit then about the money in the market that can... I mean, we're going to talk about one and a half, two trillion for SpaceX, a trillion for OpenAI, trillion for Anthropic. Is there... I mean, obviously that is just the market cap, not the money going in, but there's a lot.
But the money has to come from somewhere. Again, another advantage to going first and or going second and not going third is there's not infinite capital out there. You're going to want to be in, you know, one or two of these things and then probably not all three. Although we'll see. And they're going to require a lot of capital and I think the advantage of going first is you have access to everybody's liquidity and then you go second, you have access to some of that liquidity and then when you go third, like, well, we put this much into SpaceX and this much into Anthropic and we, you know, I'm sort of capped out on the percentage of money I devote to public equities, etc., etc.
Do you think the other thing... let me ask you, the investors, the public market secondary effects on... but they must be making contingencies, right? They go through these planning cycles where they're like, I'm going to want a piece of all these companies, but so I'm going to have much more money available than previously, but maybe not.
We'll see. We'll see. There's such... as you already said, even though those market caps are not the amount of money that's sloshing around and on the supply-demand side of the shares that are floated, it's still... there's still big numbers and there's not infinite capital out there. The interesting other thing that will happen is there's been so much illiquidity in the private markets. You've got all these university endowments and venture LPs, etc., that haven't had their big liquidity events, that are now going to be, you know, flush with distributions of SpaceX and Anthropic and OpenAI that are going to be able to turn around and pour a bunch more money into the early-stage venture capital again, that have been maybe... people who, hey, like, SpaceX is marked up to this, it's on a markup basis, it's now this percentage of my portfolio, it's only supposed to be seven or eight percent, it's now 15 percent, can't invest in new venture funds. That becomes liquid, you get your big SpaceX distribution, you got a lot of money to go invest in venture capital again. So there will probably be six to 12 months down the road a great time to go raise venture capital from LPs.
Right. And then the sort of question is, well, what can you invest in that doesn't get eaten by these companies? I mean, something you have to be thinking about if you're a venture investor for sure.
Yeah. All right, let's talk a little bit about where this story can sort of unravel, where it could go wrong. Let me give you one scenario here which is I was speaking with an investor recently about the margins that these big AI companies are getting. And I thought, well, they can easily raise the price because the economic value that they're creating is so great that you could double the price and most people would pay that price. This investor said, no. They're so close to parity that you're much more likely to see a price war than you are to see a across-the-board price raise. And then I had to agree with him. And so if that happens, then how do you justify these massive valuations for the OpenAI and Anthropics of the world?
I agree with that sentiment. I actually think there's a bigger challenge for them and it's something that is growing in the news, which is the public backlash against where's all the compute going to come from and who's... whether are these data centers that have been already committed to and approved by the county commissioners and...
Are they really going to keep happening? The one that's been in the news a lot is the Utah data center just north of Salt Lake that Kevin O'Leary, the Shark Tank investor from the Canadian Shark Tank, whoever's behind this thing, and I actually don't know who it is. Why they picked Kevin O'Leary, a Canadian Shark Tank guy, as the face of this is anybody's guess.
It's funny, we talked about this before. You mentioned these Canadians. So like, why are you in a US infrastructure project? Fine. But I would say the bigger problem is he was the face of FTX.
Yeah. Not the guy, right? Somebody picked this guy and he's going out and doing big public interviews on both the right. You know, he's on the Tucker Carlson thing. You've got the right that really doesn't want this thing to be built. Progressives and others who are like, it's going to do this to the environment and it's really not going to create a bunch of jobs. It's going to be just this huge problem in Utah. And he's not prepared for these interviews. The Tucker Carlson interview was not dissimilar to the Ted Cruz Tucker Carlson interview. He just got crushed.
So, I think there's actually going to be a growing backlash against 'hey, not in my backyard' on these data centers. The compute has to come from somewhere. These things have to get built and that's going to be a real problem for these companies should that become a...
This is like the only thing I can think of that unites the right and the left in the country right now is 'I don't want you to build one of these things in my backyard.'
Yeah. Recent Gallup poll said 9 out of 10 Americans don't want them. The other three people don't know what one looks like. It's probably the other three got a check in the mail or something like that because they're widely unpopular. If you give a commencement speech and you say anything about AI, they'll boo you out of the stadium like we've seen recently with Eric Schmidt. So that could be a real... the idea that there could be moratoriums.
There was one that was proposed in Maine. It got vetoed by the governor. But it's being brought up by national politicians. The idea that moratoriums could happen, maybe not tomorrow, but it's going to be on the ballot in 2028. There are going to be pro-moratorium candidates, you can guarantee that.
So the interesting thing about all this is, because you've got... what has Silicon Valley spent the last two weeks gossiping about? That Elon, Sam, Brockman... the Brockman diaries. I'm sure he regrets having a diary. You don't have a diary, do you?
I don't have a diary. I had one and burned it. No, I don't have a diary.
While these guys have been like, 'I know you are, but what am I?' in court, the real challenge is they should be helping Americans understand this is why it's important we win. Here's what happens if we don't build these data centers. It's not the most difficult challenge in the world. If we don't do these things, it's a fairly straightforward story. We don't do these things, China wins. If China wins, we lose. If we're the follower, we don't get to vote on a lot of things happening or not happening. They're going to tell us whether that happens or doesn't happen. And they're not telling that story right now to people.
The China thing can be too abstract for, let's say, a college student who's like, 'I don't know if I'm going to have a job.' Man, if I was sitting at University of Arizona and Eric Schmidt told me about AI and I didn't have a job lined up coming out of college, I would be booing. I didn't have a job coming out of school as it was when I graduated. And if I heard anything like that, oh man, I'd be ready to boo.
Yeah. You better start telling the story and there's no time like the present. You haven't been telling it in the past. And again, Americans' perception of these guys is they're like bickering about who's going to be more rich in court and name calling. And that coupled with now you want to go build this thing next to my house is a bad look. And so they just need to be a lot more focused on the external narrative about this is why this is important to the future of our society and it's not happening at all right now.
So let me challenge the economic benefit part of this and see what you think. Another place this could go wrong is are these companies going to create enough benefit for others using them. Chamath had an interesting tweet.
Wait, Chamath had an interesting tweet?
