OpenAI’s Financials Leaked — The Losses Are Staggering

0:00Support for the show comes from Odoo. Running a business takes everything you've got, and a lot of the tools out there that are supposed to make your life easier just aren't great at talking to each other. And that means you end up having to toggle between a dozen different apps and services just to keep the lights on. Enough of that. Now there is Odoo, the all-in-one, fully integrated platform that might actually help you get it all done. Thousands of businesses have made the switch. So why not you? Try Odoo for free at odoo.com. That's O-D-O-O dot com.

0:38With Finn, we've built the number one AI agent for customer service. It solves up to 90 % of queries for businesses, tops all the performance benchmarks on the G2 leaderboard, and it comes with a million dollar guarantee. Check it out at fin.ai. Support for this show comes from Odoo. Running a business is hard enough, so why make it harder with a dozen different apps that don't talk to each other? Introducing Odoo. It's the only business software you'll ever need. It's an all-in-one, fully integrated platform that makes your work easier. CRM, accounting, inventory, e-commerce, and more. And the best part?

1:16Odoo replaces multiple expensive platforms for a fraction of the cost. That's why over thousands of businesses have made the switch. So why not you? Try Odoo for free at odoo.com. That's O-D-O-O dot com.

1:47Welcome to Prof G Markets. I'm Ed Elson. It is June 17th. Let's check in on yesterday's Market Vitals. The S &P 500 and the Nasdaq declined as chip stocks sold off. Meanwhile, the Dow hit another all-time high. Brent crude fell lower. The yield on 10-year Treasury slid ahead of the Federal Reserve's interest rate decision due this afternoon. On Calci, the odds that the Fed It holds rates steady, are at 99%. And finally, SpaceX stock popped another 15 % early in the day. Before paring back most of those gains, it is now roughly as valuable as Amazon. More on that later. OK, what else is happening?

2:32OpenAI's financials were just leaked, and the numbers are wild. The company hit$13 billion in revenue last year, up more than 250 % from 2024. But the number that has everyone talking is how much OpenAI lost last year. The answer,$39 billion. Meanwhile, the company has just filed to IPO and plans to go public later this year, which begs the obvious question, is this kind of spending sustainable? So to dig into these numbers, we are speaking with the man who actually broke this news, the guy who found the financials, and that is Ed Zitron, author of the Where's Your Ed At newsletter and host of the Better Offline podcast.

3:19Ed, it's great to see you. Thank you for joining us. Once again, I know you've had a busy day because you literally just reported these financials. I'm not going to ask how you got your hands on them, but what I do know is It's been independently verified, audited by the Financial Times. These are real numbers. You found these numbers. Take us through them. What should we know about these financials? So last year, OpenAI spent about$34 billion to make about$13.01, sorry,$07 billion and had about$22 billion in cash at the end of the year, lost about$21 billion. dollars that 38 39 billion dollar number is what i would describe as gap voodoo there is some very strange stuff and i'll be going into this in future episodes going on in the balance sheet and everything with this company they turned from a non-profit to a for-profit though they've remained profitless uh last year which means that they have their net loss is quite strange it's like 38.5 billion dollars but the number i'd like to come back to is they lost 21 billion dollars from operations and they spend astronomical amounts on sales and marketing they spent 7.5 billion dollars on cost of revenue 19.18 billion dollars in r &d but that sales and marketing cost that 5.73 billion dollars is one i really like to hammer on because what it's like it's just like it's like excuse me well how much did you spend because the thing is here there could be i think that there's a potential that that could be partly passed through to Microsoft with revenue share.

4:57I truly don't know. This is just me guessing. But nevertheless, that's an extreme cost. But another thing to note is$867 million. So about 6.6 % of OpenAI's revenue last year came from SoftBank. Now, I truly don't know what that's for. They have a thing called Cristal Intelligence. I'm not kidding. It's actually called that, spelled the same way too. But that's a very alarming number because that's a huge amount of revenue from one partner that I don't think was probably causing their cogs to spike at all. That's a large amount of their revenue coming from a single partner that is more than likely just, I mean, I can assume not, like, Crystal Intelligence is not launched anywhere.

5:39It's only, it took a few months to even get out the door. It was announced in February 2025. Took until the end of the year to hear anything about it, but they still paid nearly a billion dollars. It's very concerning. And it also suggests that OpenAI's growth is not quite as fast as we believe to. Just going back to the losses here, because this seems to be the most important thing for investors to understand is how much money are they burning, really? The number that I saw, the net loss that was reported by you and by the Financial Times independently, I know you guys looked at the same document, but there are a lot of numbers on there and it gets a little confusing.

6:15the number i saw was 38 and a half billion dollars in net losses up eightfold year over year now there were there was some nuance in here um because as you pointed out when you take into account these other factors like interest income and interest expense the number goes up to 60 billion dollars and you pointed out that they lowered that number by taking that stuff out but then the financial Times also spoke with someone, quote, familiar with the matter. And this person who was familiar with the matter said that the reason that that number was so high was because as OpenAI's valuation went up, they had to create this$30 billion charge in investor rights, which I kind of understood to be stock-based compensation.

6:59And so their argument was, when you adjust for that, the real net loss is closer to$8 billion in losses. All of this is obviously very confusing, and I don't fully understand it. So I just want to get your view on how do we make sense of this? Because there's a lot of numbers, a lot of nuance, and I'm not fully sure what it all means. So I did not speak to the person familiar with the matter. I do not know who they are. I think that's wank. I'm just going to be completely honest. You spent$34 billion to make$13.07 billion. You didn't just lose$8 billion. Stop. When I read that comment, I did a big chuckle and I knew all the boosters would all get hot and heavy and be like, aha, the proof we need.

7:44The numbers to look at are the actual costs and the actual revenues. These companies, OpenAI in particular, have lived high on the hog, spreading weirdness around numbers. The numbers to focus on how much they're burning. $1.57 billion just on general administrative, just on people. And I don't even think that that includes all the people. I don't have any other knowledges to this because I don't have the definitions of these terms from within the sheets. But the thing is here is this is a company with spiraling costs, dramatic costs, way beyond what we thought, that is spending. And they put research and development in there and they market that, I believe, personal belief here, because they want people to think that this is a temporary situation, that this is something that will go away, that when there's enough R &D, they can stop.

8:35No, no, no. R &D number going up very high. Everything is increasing. Every cost is increasing. And I think it's just a kind of a frightening sign of what the innards of these companies look like. And think about it like this. OpenAI has already said as part of the Musk trial that they're going going to spend$50 billion on compute this year alone in 2026, their losses are going to be astronomical. I don't have any privy knowledge to the 2026 numbers, but based on these, they could be burning$80,$90 billion. It's genuinely horrifying. I'm glad that I got to tell this story before the S1 is filed, because I think people need to see what the innards of this company look like before they get a chance to massage them.

9:21Yeah, I agree that actually one of the craziest numbers was the sales and marketing number up to$5.7 billion. It jumped 418 % in one year. That's 44 % of total revenue. And we just look for context. When you look at Facebook's marketing, it peaked at 28 % of revenue in 2008. Google's hit 11 % in 2003. Open eyes at 44%. So they're spending astronomical amounts of money not to build AI, but to sell AI, to make it profitable. And that does seem to be a real problem. I mean, just looking at the SpaceX IPO as a comparison point, when the S1 came out, I wrote about it and a lot of people talked about it.

10:10I'm sure you talked about it as well. The numbers did not look great from a profitability perspective. We called it a money furnace, cash incinerator, etc., because that's what was happening and continues to happen. The company is losing astronomical amounts of money, building mostly AI, also rockets, but AI is the real money loser. And yet the stock is flying right now. It's as valuable as Amazon currently. So I guess I'd be interested to hear what you think will happen when the S1 finally does come out for OpenAI. Maybe these numbers aren't concerning to people, or maybe they will be. I'd be interested to hear your views.

10:53If OpenAI had SpaceX's losses, they'd be so happy. But then again, SpaceX doesn't have OpenAI's revenues, so who knows? I think the reason SpaceX has flown in the way it has is because of the Musk reality distortion field and his ability to play Goldman and JP Morgan against each other, both of them, everyone there should be ashamed of themselves for pumping this, for what they're just like, oh, AI revenue will increase 300x in four years. Putting all that aside, these are fundamentally different companies. Even though SpaceX has its horrible burning AI element, it still has Starlink, it still has the rocket ships, and it also has the social network, but I don't know if I'd say that was a plus.

11:34It still has a functioning business. Open AI has chat GPT and API. That's it. Like they have, they can dance around being like, oh, we got Joni Ives nonsense. Oh, we got like a consumer device. Oh, we got, we're going to do an everything app. But they're just dancing around the fact that they have one product. They have one product with a few offshoots. They don't have ways of investing more money to make money. You mentioned profitability earlier. I don't think that's possible. I don't, like their costs are so severe. that I don't think profitability is actually possible for any AI lab. I think that Anthropic is going to look similarly horrible.

12:15I think that the markets believe for some reason that things won't look this bad. I've been speaking to people all day about this and everyone's saying, wow, I didn't know it would be that bad. And I genuinely, I got a laugh at it. It's like, did people think I was kidding? Did people think I was chuckling to it? I guess they were bad. No, of course they were this bad. So I don't think that OpenAI and Anthropic have the nuance and also the aggressive bully mentality of Musk, who I really deeply dislike, to be clear, to do an IPO with the amount of bullshit that will be necessary to sell this dog.

12:53I don't know how, because remember, in the S1, they're also going to have to do growth trajectory. They're going to have to say what they could grow into. I think it's going to be difficult to float. They could, theoretically, but considering how SpaceX is done, I think they have a higher chance than none. But at the same time, SpaceX's numbers were bad. These numbers are terrifying. These are scary numbers. And watching the Copa-Lympics on Twitter, there's people like, well, actually, it's okay. You saw the quote that said$8 billion. It's okay. They didn't actually spend$34 billion. I only spent$8 billion.

13:35It's okay. Nothing bad happened at all. Nothing weird. It's all very silly because this is a company that spent$34 billion to make$13.07 billion. And also$867 million of that was what looks like given to them by SoftBank. And that is not a stable or thriving business. This is a business that is just consuming capital at an alarming rate with no sign of stopping. They just raised$122 billion at the beginning of this year. I don't think they're doing that because they're going to get profitable. It does seem as though when this company goes out, when it goes public, and it appears it is going to go public, they have filed, plan is to go public in the fall.

14:16the valuation will be similar to SpaceX, almost entirely dependent on the CEO's ability to tell a compelling story that enough people buy into and believe. Elon is someone that you could arguably bet good money on doing that because he has a proven track record of being an incredible storyteller specifically to the markets on what his companies will achieve. sam altman i'm not sure i back him uh to tell a compelling story or to capture the imaginations of investors in the markets because it's clear that the numbers aren't going to do it for him well also the other day i just i had to go and check on this while we're talking here's the reason i don't think so i agree with you fully i don't think sam altman's got the head for this because when his customers started complaining about the cost he said in a panel yeah, AI costs are a huge issue now.

15:09You don't say that, Sammy. You don't. That's not what investors want to hear. You make some crap up about how we're doing cost adjustments, but people are really excited to see what they can do with AI. You don't go, yeah, that's a huge problem. What's he going to do on the roadshow? They're going to be like, yeah, you keep losing money. Yeah, you know, you've got to lose money to lose money. Yeah, it's certainly very concerning. Well, we're going to wrap this up. But Ed Zittrain, we always appreciate it. It's amazing that you got your hands on this. It really is so important in these companies, the size and the valuation and the sense in which they are becoming in a lot of ways systemic to this market and to expectations of this market is so important.

15:55And now we've finally gotten some transparency into what's really going on here. So we appreciate it. Ed Zittrain is the author of the Where's Your Ed At newsletter, also the host of the Better Offline podcast. Ed, appreciate your time. Thanks for having me. After the break, SpaceX stock keeps going up. And for even more markets insights, you can subscribe to my weekly newsletter, Simply Put, at simplyput.profgmedia.com.

16:31Support for the show comes from Odoo. Running a business is hard enough. So why make it harder with a dozen different apps that don't talk to each other? One for sales, another for inventory, a separate one for accounting. Before you know it, you are drowning in software instead of growing your business. This is where Odoo comes in. Odoo is the only business software you'll ever need. It's an all-in-one, fully integrated platform that handles everything. CRM, accounting, inventory, e-commerce, HR, and more. No more app overload, no more juggling logins, just one seamless system that makes work easier.

17:08And the best part, Odoo replaces multiple expensive platforms for a fraction of the cost. It's built to grow with your business, whether you are just starting out or already scaling up. Plus, it is easy to use, customizable, and designed to streamline every process. So you can focus on what really matters, running your business. Thousands of businesses have made the switch, so why not you? Try Odoo for free at odoo.com. That's O-D-O-O dot com.

18:01up to 100 % of school-certified costs. Not just undergrad and graduate tuition, but SOFI can also help cover housing, books, food, and other education-related expenses. SOFI's private student loans are all about flexibility. They offer competitive fixed or variable rates, multiple term options, and monthly payments, so you can build around your budget. The application process is completely online, and you can check your rate in minutes. And for families exploring their options together, adding a qualified cosigner may help improve chances of securing a lower interest rate. And there are zero fees required.

18:35That means no origination fees or late fees, no surprises. What you see is what you pay back. Head to SoFi.com slash ProfG student to check your options and get your education funded the smarter way. That's SoFi.com slash ProfG student. Originated by SoFi Bank, NA member FDIC. Terms and conditions apply. Please borrow responsibly.

19:00Support for the show comes from Granola. We'd all take a slightly longer meeting if it was actually productive, especially over a bunch of short ones that accomplished nothing. But the trade-off is this. Our brains can only hold on to so much information, and it's easy to forget the details when the meetings keep going. So when those longer meetings roll around, and you're in need of clear, actually useful notes, you'll be glad to have Granola. Granola is an AI-powered notepad built for the way real people actually meet. You can take notes like you normally would in the background, while Granola securely transcribes the meeting.

19:31Then, after you wrap up, it turns everything into clean, structured, actually useful notes. Granola also works through your device's audio, which means it integrates seamlessly into the video conferencing tools you already use. It's your standard meeting setup, but enhanced. You get to actually listen and contribute instead of trying to capture every word or summarize in real time. You can walk away knowing exactly what was decided, who's involved, and what comes next. If meetings are eating up your day, Granola is a no-brainer. You can try it totally free for three months. Just head to granola.ai slash markets.

20:02That's granola.ai slash markets to get your time back. Get three months free at granola.ai slash markets.

20:18We're back with Prof G Markets. In its third day of trading, SpaceX became the fifth most valuable company in the world tied with Amazon. The stock popped more than 15 % early in the day to reach that milestone and briefly eclipsed Microsoft, too. That rally followed an announcement from SpaceX that it is acquiring AI coding startup Cursor for$60 billion. The stock gave up most of those gains through the day, closing up just 5%. Still, shares have rallied 48 % from their IPO price on Friday. So clearly, there is still a lot of excitement around this company. So we wanted to talk to someone who has actually run the numbers and attempted to value this stock.

21:03So joining us to discuss SpaceX, we are speaking with Nicholas Owens, equity analyst for Morningstar. Nicholas, thank you for joining us. You are someone who has actually done a discounted cash flow analysis of this company. You valued the company. You wrote about it. And your number that you reached was$780 billion, which is, I mean, we compare it to the current valuation,$2.6 trillion. It is notably a lot lower. Take us through how you got to that number. The headline number that you just mentioned is the weighted average of three scenarios. And the way we valued SpaceX was in three parts, essentially, the rocket piece, Starlink, and the AI piece, which is itself kind of a set of moving parts.

22:00I would argue that the rocket and Starlink is more straightforward. It's a more mature part of the business. And in most of our scenarios, they're worth, say,$611 billion. in terms of enterprise value, some$40 a share, let's say. The AI, we think there's a lot of riding on some unproven outcomes. And I'm not really talking about the science or engineering behind data centers in space, though there's some debate about that. It's really the financial benefit of having a data center in space. Is there some kind of operating cost advantage versus a terrestrial data center? We modeled this business primarily as an infrastructure play, so renting out computing capacity the way they've started to do with Anthropic and Google.

22:50You know, if Grok takes off, then we would say that we're sort of indifferent between them renting out the capacity and using it for lots of Grok jobs and then monetizing that some other way. We would assume there's some market rate for that, and that's how we modeled it. And so the difference between our valuation and the market price really has to do with how probable do you believe it is going to be that both the Starship rocket is highly reusable, like in hours or days, and that data centers in space will be a, call it a compelling bargain in terms of their operating costs versus terrestrial.

23:29Both of those might be true, but we think there's only about a 7 % probability that they will both be true at the same time. And that results in a much higher valuation that's closer to the market price here. And so I think that the market investors are essentially saying they believe these will both be true for sure, but we don't. This is the thing that I think is really important, which is that your valuation is weighing multiple probabilities at the same time. And it isn't ruling out what you describe as the moonshot scenario, which I think is an accurate portrayal of some of these businesses, orbital data centers, down the line civilizations and cities on Mars.

24:11I mean, how do you even value that? But the point being, you've actually taken all of that into account, weighed up the probability, and you've landed in a number that is less than half what the company is currently trading at. And I think that that is striking. I want to get to your reactions to how the stock is trading in a moment, but I want to stay on your valuation for a moment longer. You also said, you said, quote, we value SpaceX at$780 billion with a Morningstar economic moat rating of narrow. I've seen a lot of people pushing back, criticizing the moat rating. A lot of people would say this is an extremely wide moat.

24:57They're building rockets. Not many people can do that. Take us through why you landed on that. I'm really glad to have the moat methodology behind me, so to speak. At Morningstar, we had quite a good discussion about this a couple weeks ago. And I will say this. The SpaceX business that I, as the aerospace and defense analyst, was planning to cover until February has almost all of the characteristics of a wide-mobile business. A very, very prodigious cost advantage, both through the R &D and how they do it and the economies of scale. They've just done it more than anyone else. And so the satellite business and the rocket launch business kind of reinforce each other in terms of them just marching down this cost curve.

25:44They're a decade ahead of anybody else and will continue to do so, assuming Starship continues to add, you know, the bigger loading bay of Starship means that you can put more stuff in it. So your cost to launch a kilogram of stuff goes down. You're sort of just dividing by a bigger denominator. And that's amazing. The company gets a narrow moat rating from us because of the AI piece, which is, I would say, indeterminate at best in terms of what its moat would be. And they're investing very aggressively in that business. So you're taking these returns on capital that are evident from the Starlink and Rockets business, and you're saying, I'm going to write a$60 billion check to go get Cursor, I do similar size of investments to build data centers in space.

26:34And while I think there could be moats in AI, we don't see evidence of them here. So, you know, Grok is not one of the leading models. I do think there's a pathway in this, let's say, moonshot scenario, and even in what we call the minimum viable product scenario, where they can extend some of that cost advantage into data centers in space. If they can, then that would be, let's say, Modi. But it's too soon to say, frankly. Yeah, this is part of the problem with this stock, which is it was a space company. And then a couple months ago, they turned it into something else by buying XAI and then saying that actually it's an AI company.

27:21or that that is the$28 trillion or$26 trillion opportunity, according to the S1. That was the total addressable market with AI. So, I mean, I appreciate you clarifying that. It's like there are different businesses that we're evaluating here. The space business might have a wide moat, but this is no longer a space business, apparently, because that's not what they're pitching. And that's supposedly not what the valuation is predicated on. Let's go to the stock price. Do you agree with that? Yeah. And I think segueing to the stock price is appropriate because I think the, let's call it, packaging of this company as an AI business is a very rational way to tap into investor appetite for AI.

28:09Right. So the IPO price was$135. We're looking at$200 a share at the moment. That does not make much sense to me. I assume it doesn't make much sense to you. But when I think about reasons why that might be the case, the first thing that jumps out to me is that only 4 % of the shares are available to the public. The float is incredibly small, which to me means this has high potential to become a meme stock and seems to look like it already is a meme stock. Is that the explanation? I think so. I've thought about it in terms of supply and demand, as you point out, small float. And then you have this additional, let's say, the market is looking forward to sort of, let's call it structural demand from passive funds, ETFs, and others that track a handful of indexes.