The OpenAI IPO?

0:28Good morning, good evening, wherever you are. Um, just kind of got me, you know, very focused on some things, wasn't able to contribute as much to the show. So I appreciate those two guys stepping up. Um, if you've been, uh, paying attention to some of the little details, uh, there's been a lot that's been kind of evolving in show notes, in show format, in how we're presenting the show and putting it out, uh, both in audio and visual format. So if you're checking it out on YouTube, you'll see some changes there. Aaron's been carrying the, uh, the load quite a bit there. So I appreciate him, uh, stepping up and doing a ton of things there, but, uh, just me today, kind of an old school way of doing this one, just me going solo.

1:02I thought it'd be interesting. One of the big things that's been talked about over the last couple of months is the potential IPO for both Anthropic as well as OpenAI. And this is, again, I want to be very clear. This is in no way financial advice. This is in no way financial suggestions. But I thought it'd be sort of interesting to dig into kind of where both companies are. I'm going to start with OpenAI and today's show, dig into where they are sort of pre-IPO, what the pitch they might be making to investors around the world, around the country, and maybe where they've got some areas they might need to clean up because their timing of their IPO is always very important.

1:40And the timing right now seems like it's a little more unsure on the OpenAI side. Anthropic feels like they're kind of moving forward. But again, neither one of them have necessarily put a specific date out there. There was obviously a lot of anticipation earlier in the year for SpaceX. Anthropic and OpenAI were sort of expected to be the sort of two and three of the big three this year. It's not exactly clear if either one of them will actually IPO here in 2026. Lots of moving pieces and parts. So I thought I'd dive in a little bit on the OpenAI one today. I think I'll conquer Anthropic for next weekend, sort of weekend perspective, but going to dig into that right after the break.

2:24Today's show is sponsored by Nasuni. There's a growing gap in AI right now between what's possible in theory and what successfully works at scale inside an enterprise. The difference comes down to unstructured file data. Many AI initiatives struggle because the file data they depend on is scattered, unstructured, and disconnected from where and how work actually happens. Nasuni changes that. It brings your unstructured file data into a single, secure foundation so AI, both generative and agentic, can access it with the context, governance, and performance it needs in production. Bring AI to where your unstructured data lives.

2:58See what it takes to activate your data for AI and request a demo at nasuni.com slash AI.

3:07What does the outside world already know about your company? Cyber criminals could be seeing leaked credentials, compromised session cookies, exposed infrastructure, and even impersonations of your brand and executives. Nord Layer Intelligence by Nordstellar gives security teams visibility into those threats across the deep and dark web. Data breaches, attack surfaces, and brand impersonation all in one platform. Find out what attackers know before they can use it. Visit nordlayer.com slash intelligence slash enterprise AI and use code enterprise10 for 10 % off your Nordstellar plan. And we're back.

3:53And as I mentioned at the top of the show, I want to do a little bit of kind of old school weekend perspective style. Dive into where OpenAI is prior to their IPO. Again, recording this September 27th, so sort of end of 2027, end of September 20, end of September 2026. Got that all out. OK, so just to give you some perspective from timing, they have not announced an official time frame on this. I know they've gone back and forth a little bit about when they might. You know, there's also some speculation right now that they are potentially in the private markets looking for another round of funding, looking to try and get to more than a trillion valuation.

4:34And this would be pre any sort of public market IPO. So just to put that into timing perspective, it's interesting. I was kind of looking at, okay, what's the best way to sort of frame this conversation? Do we look at this from a financials perspective? Do we look at this from a momentum perspective, a technology perspective? I think it's important to sort of look at that from all of those things. And then also look at what some of the other macro level things going on in the industry, in the economy, kind of in the world that may have some impact on whether or not OpenAI eventually does go IPO.

5:11I think that's a perceived perceptive that they will eventually go IPO. But timing is always very, very important. And what's going on in the world that may have external effects on what they're doing is also going to be important. So I'm going to try and break down a bunch of those different things. Obviously, I'm going to put a bunch of things in the show notes as best we can. But if we sort of start with the basics, right, if you're going to go IPO, a couple of basic things that people want to understand is, you know, what do your numbers look like? And a lot of this is speculation. You know, there have been some reported leaks, you know, folks like at Zitron, Financial Times, others, you know, have tried to, you know, dig into exactly what their numbers are, both, you know, trying to get it through leaks, trying to, you know, do channel counts and other types of ways that you try and figure out revenue.

6:01You know, we've seen OpenAI talk about some numbers, whether it is Sam Altman talking about numbers or their CFO starting to talk about some numbers. Obviously, we hear some leaks as they are, you know, going around doing, kind of feeling out the market for, you know, where they think investment can come from. They're having to share some amounts of numbers. So let's start with the numbers where we are today. As of, let's kind of put this a little bit historically in perspective. The revenue that was sort of announced, and again, I'm going to kind of stop saying things like sort of announced and we're not exactly sure.

6:39I'm just going to kind of put some numbers out there that the industry has kind of rallied around. 2024 revenue was$3.7 billion. So let's just call it$4 billion in 2024. In 2025, revenue was$13 billion. So we went from$4 billion to $13 billion, a little over 3x sort of ramp up number, let's call it 3.5x from 2040, 2024, to 2025. And right now, OpenAI is sort of signaling to the market that they're on about a$40 billion run rate for 2026. So again,$4 billion in 2024,$13 billion in 2025,$40 billion in 2026. That last one would be a little more than a 4X, probably, again, about a 3.5X kind of jump in numbers.

7:34So those are the kind of numbers that you typically are looking for prior to an IPO, right? So sort of 3X, 3X, 4X, kind of in that range, you're not looking for necessarily less than like 2X, you know, going into an IPO, because you want to be able to show the market that there's still a significant amount of growth in the business. And again, you know, the business would eventually be then expected to grow, maybe not just in the business that you're in today, the segments that you're in today, the products that you're in today, but expanded into other markets, into other products, and so on and so forth.

8:12So, you know, again,$4 billion,$13 billion,$26 billion, or I'm sorry,$40 billion would be sort of the numbers for 2024, 2025, and 26 that they're going to be fundamentally talking to Mark about. Now, the flip side of that is the expense numbers on OpenAI have been probably perceived as a negative, sort of too high at this point. So just to put those in perspective, on that roughly$4 billion of revenue in 2024, it's about$13 billion in expenses for 2024, so a loss of roughly$10 billion. dollars. That$13 billion revenue number in 2025 was on expenses of about$34 billion. So again, now we're looking at about a$30 billion loss.

9:01So we went from a$10 billion loss in 2024 to about a$30 billion loss in 2025. And we don't necessarily know what those numbers look like for 2026. But if, you know, if things are, you know, sort of going on that same sort of ramp, you know, you would expect on a$40 billion revenue number in 2026, that expense number is probably somewhere in the$70,$80 billion, maybe more than that, just for the annual for the year, right? So that's the sort of the yin and the yang of revenue and expenses, right? They are still significantly higher in expenses on an annual basis to deliver the service that they are, you know, that they are driving large amounts of revenue on, but they are doing it at a very, very high level.

9:52And those expenses are everything from, you know, the cost of training models, the cost of the team itself, which is a very, very, very high-end team, the cost of training models, so GPU cycles, data acquisition, all those sort of things, running data centers or renting data centers, the cost of selling, the cost of their internal cost of expenses in terms of things like salary and equity that invest over time and so forth. So still a very unprofitable business, but that's not necessarily ever prevented a company from going IPO. So again, I'm not trying to make a claim or a case one way or the other for the viability of this IPO, just trying to give you some sense of what that is.

10:41Now, to put this in perspective, funding-wise, up until

10:50up through March of 2026, they took another round. They have taken somewhere on the order of about$180 billion in funding. So this is from May of 2021, Series A, through, let's see.

11:15Don't panic. This podcast is a work of fiction. From the Global News Network, it's 2029. Breaking news. A hacker calling himself the Purple Yak may have used advanced AI to gain control of a Russian nuclear warhead. He's reportedly offering to launch it at a target of the highest bidder's choosing. Get the news from the end of the world. Follow the clues. Unravel the mystery as civilization collapses. 2029, The War for Our World. A science fiction news series from the future. Premiering October 30th. Follow wherever you get your podcasts.

11:56um yeah through let's see there was series f in march of 2025 another round in august of 2025 which was sort of a follow-on to that and then what they called a mega round 122 billion um in march of 2026 so over about five years they've taken roughly 180 billion dollars that might be closer to 200 billion dollars um and you know again if we go back to you know the numbers through what they're burning. We saw in 2025 expenses were$34 billion. So they are burning through that$200 billion number. And then the second part of it that people look at quite a bit is they are carrying roughly$650 to$700 billion in what people are calling, or what's described as off balance sheet purchases.

12:51So commitments to future data centers, to future purchases, to Google, to NVIDIA, to Stargate data centers and things like that. So there's quite a bit of not only funding that they've taken in, funding they are burning through at a pretty healthy rate. And I mentioned earlier, there's talk of them even pre-IPO taking another round of funding to try and get them up to a higher valuation. So maybe that's another$100 billion of funding. But it does feel like they're burning through that funding at about$30 or$40 billion per year. So you can sort of do the math on the burn rate. And I may have that slightly off, but you can do that.

13:35And then they've got commitments of between$600 and$700 billion. And those numbers might not be exactly right, but I believe those are probably right in terms of, yeah, about$650 billion in terms of future commitments on spend. So again, you sort of go through that and you go, okay, you know, the revenue way it looks right now is while the revenue is growing, the revenue is not growing as fast as expenses. So they're going to be, you know, tapping into funding rounds that they've taken. So that number has to sort of catch up somewhere. They either have to reduce expenses, increase revenues, or get those numbers going at a different velocity, sort of in the up and the down rate.

14:25They've got, like I said, about$650 to$700 billion for the future obligations through about 2020 to data center companies, to equipment companies, and so forth that would have to be paid out through profits, through revenue, less expenses. And right now, that doesn't necessarily look like the math necessarily works out. Now, they would argue that over time, their expenses are going to come down, their revenues are going to grow, those types of things. But that's sort of the high-level financial picture of what OpenAI looks like as of September 2026, prior to any sort of IPO. Now, you know, you look at their strengths.

15:07I mean, their strengths, fundamentally, they've got two or three major strengths that they're bringing to the market. The first is massive brand awareness. Obviously, they were sort of the company that defined what generative AI looked like. ChatGPT is, and probably still is to a certain extent, you know, perceived sort of in the same way that Google is in terms of it almost being a verb, right? It's sort of the thing that you describe when you're talking about generative AI, except for sort of the niche-y technical class of people who know the difference between a clod and an opus and all sorts of other tools and capabilities within the AI space.

15:46But if you're just talking to sort of the general public, they know what ChatGPT is. There's more than a billion people now that use ChatGPT on some sort of weekly basis. So massive, massive brand awareness, massive ChatGPT user base. Now, they're not necessarily monetizing it terribly well. Most of it is either free accounts,$20 a month accounts, some ad-supported capabilities or ad enhancements to those things. But that's sort of the unusualness of having a massive consumer business that is also trying to transition into an enterprise business. But again, first and foremost, massive brand awareness, massive existing user base, again, whether that's a paid user base or not.

16:36Their second gigantic strength is just incredible depth and experience of their frontier labs. So, you know, perception that they are in the top one or two, depending on, you know, whether you're comparing them to philanthropic, whether we're comparing them to some junk of the team that may have just joined or may have left or something like that. But they're very much perceived to be in the top, probably the top two in terms of engineering talent, of depth of their lab. So the perspective there of them being able to create great things, of them being able to create what the future of Gen AI looks like, whether that is around safety or whether that is around new capabilities that They're going to unlock value for humanity or for any individual or for a corporation.

17:25They are perceived to have the technical capabilities to be able to go out doing that. And then thirdly, in terms of sort of their ability to reach the enterprise, while it's a different partnership level than it was a couple of years ago, their co-pilot installed base with Microsoft, there's still ongoing partnership with Microsoft, although it will end at some point in time in the near three, four, five years, or at least will diminish in the next three, four, five years, is still very, very powerful. You know, most enterprises' first engagement around generative AI was with Microsoft using Copilot.

18:04There's still a lot of Copilot that's out into the marketplace that's embedded contracts, you know, embedded large ELA deals and so forth. So they do have both a massive channel into the consumer space, as well as a massive channel sort of secondarily through Microsoft into the enterprise space that that could evolve over time. You know, the co-pilot brand, the ChatGPT brand has some brand conflict. But again, I think most people perceive that the thing they're using is open AI in one way, shape or another. So that's really their strengths. Their strengths are, you know, first mover advantage, you know, big user base, both on the enterprise side as well as the consumer side, an incredibly deep, experienced, you know, evolving, innovative team on the frontier lab side.

19:01So it's sort of the engineering and technical side of things. From a weakness perspective, there are a number of things, right? Their sort of early mover advantage has slipped. Anthropic has taken over a good chunk of sort of defining what one of the biggest markets is, which is around engaging around software development, software developers with the enterprise. So, you know, there is that. There's quite a bit of, and this is somewhat outside of their control, but also sort of within their control, a big negative sentiment about AI and data centers within the United States. So that has not only potential implications in terms of people's willingness to make massive investments in AI, and that's going to sort of ebb and flow because individual users' perception versus what large corporations would like to do are very different buying centers.

20:00But there is a negative sentiment and has been a negative sentiment here for at least a year or so within the United States around AI and data centers. So that's going to be negative headwinds that they're going to have to deal with. There are the associated sort of political ramifications of that. You know, here in 2026, the current administration, you know, is still very pro AI, very, you know, pushing the idea of a big data center expansion within the United States. if the elections in 2026, the sort of midterm elections in the United States, shift in terms of which party is in power, we may see that sort of negative sentiment turn into actual policy that slows down some of the things that could have impact upon open AI.

20:48So whether that's just increased regulation, increased cost, more difficulty to get permits to be able to drive some of these data center expansions, whether it is having to do with open models versus their closed proprietary models or something along those lines, that sentiment, which is more dominant in the US in terms of policymaking versus end user kind of opinion, that could shift in 2027. I know OpenAI has been one of the largest contributors to the current U.S.

21:30presidential and so forth. They are one of the biggest contributors. They're trying to drive policy through that relationship. That may not be as much of an avenue for them beyond 2026, again, if the midterm elections sway the balance of power in the United States from where it is today to where it could be. in the future. Getting back to things that they have some control over, they are what feels like have gone through a constant amount of corporate restructuring over the last 18 months or so. They went from being a almost entirely consumer company, almost entirely a one brand company around ChatGPT to a big shift over the last 18 months.

22:13We've seen them try and get into hardware. We've seen them trying to get into ads. We've seen them try and get into a lot of specific verticals, so whether it is law or finance or other types of things. We've seen them bring in a number of leaders. We've seen a number of leaders leave. So a lot of corporate restructuring. It does seem like it is now rolled up for the most part under Greg Brockman, who is one of the original founders, sort of now has control over the product strategy. It does feel like to a certain extent that a lot of the other things that they were working on, so whether it was hardware or ads or even ChatGPT as a standalone thing, feels like those have become sort of secondary focus areas, if they're still focus areas at all.

23:00It does feel like they are very much sort of chasing or trying to mimic the anthropic strategy. So we've seen much more focus around coding agents, much more focus around sort of a unified experience from chatbots to coding assistants to things that impact day-to-day white-collar work. So that corporate restructuring, that product restructuring, I think has to get sort of stabilized. It has to be a little clearer as to where do they have immediate ambitions versus where do they have future ambitions that'll have an impact on the way people look at how much potential is in their story versus how much are they looking at immediate financial numbers and so forth.

23:46They have had some issues and they've been right in the middle of, as many of these frontier labs have been, kind of the increased safety and security concerns. They had a very, very well-known agent breakout in Hugging Face. They got a lot of press, went through the normal rinse and wash cycles within the press, whether it was this was OpenAI's fault, this was Hugging Face's fault, this is normal, and this is what we should expect. This is very scary and completely unexpected. I don't think from a PR perspective, they've quite figured out how to explain that. Because on one hand, it does show the power and potential of what these agents can do, both from a negative as well as a positive perspective.

24:34I don't think they've quite figured out sort of the story because there is so much kind of polar opposites between people that are the quote unquote accelerators. You know, AI needs to move forward. AI is going to do great things. And the, you know, what are perceived to be sort of the doomers and the security related people, you know, again, people that are looking for things to potentially slow down. So, you know, we have seen them make some statements about being willing to slow down. They tried to be somewhat in lockstep with Anthropic and some of the other frontier labs. But I think, again, in that same vein of getting the story right between sentiment about AI, sentiment about AI taking jobs, sentiment about AI being safe and secure, they are not necessarily owning the story as much as they might be able to.

25:25I think they get lumped in with Anthropic, who is probably even a little more extreme about how they're telling their stories. And when we get into IPOs and we talk about price of a stock and so forth, it's very much kind of the intersection of what do your actual financials tell and what's the strength of the potential story that you're telling, the future of that story. And it does feel a little bit like OpenAI doesn't necessarily have control over the full story that they're trying to tell the marketplace. It sort of feels like in some cases they do a little bit, maybe at the product level, but it does feel like at the macro level that so much of that story is either getting sucked up into general politics and animosity and so forth, or Anthropic is kind of telling the story.

26:12and Anthropic, as we've talked about a lot of times in the show, hasn't necessarily been the best purveyor of what the potential generative AI story is. They tend to fall back to scaring the market more than they do trying to inspire the market. And then I think finally, the last thing that they're going to have to sort through in terms of how they're going to pitch what they look like is, what does their future look like? There have been a lot of things going on there. People are trying to figure out, you know, where are they making bets? What's their prioritization? You know, not just future commitments.

26:48So commitments into future data centers, right? Do they have any flexibility there? Are they going to be completely hamstrung by four or five years of financial commitments? But also, you know, one of the other things that had been talked about before is the hardware strategy still a possibility. You know, the thing that they've talked about so far, they've made announcement with Johnny Ives. They've made investment around Johnny Ives. But the initial announcements have been pretty uninspiring. You know, a speaker or a home device, not terribly, terribly inspiring. You know, what happened to the to the ad strategy?

27:19Right. The ad strategy was the thing that was going to, you know, give them this this huge installed base of revenue to go compete against a Google. You know, if you've got a billion people and the fact that you haven't been able to monetize that is a pretty big indictment. So is the is the ad strategy still front and center? Is it still really important? Is that consumer business just going away? I think those are the big questions that they kind of have to answer, you know, that right now feel like they're a little bit unsure. So anyways, hopefully that helps a little bit. You know, I think, you know, people, whether or not they use OpenAI's technology or not, I think, you know, we're going to see, you know, so much of the market right now, if you look at the overall financial market, as a lot of people have said, the whole financial market right now feels like a giant bet around AI.

28:13And so when we look at the four, five, six, seven, 10 big players that have a huge influence on that, OpenAI is going to be one of those, along with NVIDIA, along with Anthropic and Google and many others. And so kind of understanding where they are, what their timing looks like, what they're telling the bigger market, I think is going to have some downstream ramifications that people might be interested in as it relates to their own financial investments, what the bigger market investment looks like, but also timing. Timing if you are making an enterprise purchase, if you are looking at what your company's strategy looks like over the next two or three years, it's important to figure out where some of these big companies are, where they've got strengths, where they've got weaknesses, and whether or not that impacts you in any way, shape, or form.

29:00So hopefully this was interesting. It's good to get back and kind of do a weekend perspective. I know I haven't done these in a little while. It's good to be back. Thank you all for being flexible with us over the last few months. I know we've tried to make some changes, made some changes for the good, I believe. They may not always be obvious to you, but we think over time they're going to help us keep putting out a better and better enterprise AI show. So with that, I'm going to wrap it up. Thank you all for listening. Thanks for telling a friend. Thanks for continuing to send us some great guest ideas.

29:29We've got a big backlog we're working through right now. We're going to get to that very soon. And so with that, I will wrap it up. We'll talk to you next week.