TBPN Live - Disney Partners with OpenAI, Breaking Down SpaceX’s IPO Plans, Oracle Slips | Diet TBPN

Episode Date: December 12, 2025

Our favorite moments from today's show, in under 30 minutes. TBPN.com is made possible by: Ramp - https://ramp.comFigma - https://figma.comVanta - https://vanta.comLinear - https://linear.appEight Sleep - https://eightsleep.com/tbpnWander - https://wander.com/tbpnPublic - https://public.comAdQuick - https://adquick.comBezel - https://getbezel.com Numeral - https://www.numeralhq.comPolymarket - https://polymarket.comAttio - https://attio.com/tbpnFin - https://fin.ai/tbpnGraphite - https://graphite.devRestream - https://restream.ioProfound - https://tryprofound.comJulius AI - https://julius.aiturbopuffer - https://turbopuffer.comfal - https://fal.aiPrivy - https://privy.ioCognition - https://cognition.aiGemini - https://gemini.google.comFollow TBPN: https://TBPN.comhttps://x.com/tbpnhttps://open.spotify.com/show/2L6WMqY3GUPCGBD0dX6p00?si=674252d53acf4231https://podcasts.apple.com/us/podcast/technology-brothers/id1772360235https://www.youtube.com/@TBPNLive

SpaceX IPO under the Christmas tree, $1.5 trillion for the big man, for Elon Musk. He's going to raise $30 billion, something like that. That's an incredibly small amount of dilution if he actually goes out at $1.5 trillion. I thought he hated running a public company. Just $1.5. Just $1.5 trillion. He said the biggest number because you know that there was the whole leak, I think it was in Reuters, open AI going out at $1 trillion.

You've got to go a little bit higher. What's just a little bit higher? He should have done, he should have done, he should have in 1.1. I'm going to IPO at 1.1 truly. Really stick it to Sam. But he didn't. He says he's going out at 1.5. He's got to lose some room for the, for the first day pump. Yes. Get it up to two. Get it up to three. Yeah. I mean, in terms of companies that should benefit from the space economy. Like, is this not the Nvidia of space? Like what other company is there? It's the space company. It's SpaceX. 30 billion coming in. Company was doing great. Launch, launch product is. fantastic. The rockets go up. They come down. They're putting sonic booms over Montecito. Apparently. But I think that the actual launch market for anything but satellites, but for anything but data centers has just been a little small. And so Starlink has been the big unlock, of course, for SpaceX. Massive market there. Putting the screws to Verizon, AT&T, and the other telcos.

The other telcos, of course, have been noodling on working with a rival, putting up smaller satellites in a sparser constellation with more bandwidth per... Somebody was talking earlier that to get the same amount of compute as a one-gigawatt data center, you would need 10,000 satellites. That's not that many, though. Starlink has over 10,000 up there already, I believe. So that's actually not that crazy. That's actually like extremely bullish.

Well, there is a pathway to doing it with far fewer, which would just be like ultra-ultra long. I don't think they want to do big. I think the whole thing is Constellation. That was the beauty of Starlink was that you could go up there. And because it gave them the ability to have all this residual capability, as Gwynne-Shottwell puts it, which is like if the CIA shows up and they want to put a spy satellite on there, or some company shows up and says like, hey, we want to put something really serious in space,

but we only need 60% of the faring, or we only need 40% of the fairing, or we only need 80% of the faring. It's like, okay, great, we'll fill the rest of the space with two starlings, five starlings, 20 starlings. So in terms of the space thing, I mean, we keep going back and forth on this. Is it just a pump? Is it just a new narrative? Is it a pivot to AI?

I don't really care. I think it all comes down to this just idea of like Elon math. You know, the Elon math is always crazy. He's always putting up sort of wild predictions that then may or may not come true. A lot of them are way off. But at least he's telling a story that's optimistic about the future.

And at least it feels like that Trace Stevens, like, good quest thing is like, get to Mars is good. If he misses it by five years, at least he's still the first, he's the only person that cares about it, which is great. In true, like, hit piece fashion, I pulled up all the different, all the different times he's like made a prediction and then like not landed it

or not delivered on time. And it's crazy. The guy loves to rip predictions. He said he was going to put early SpaceX concept, when they started SpaceX, was to put a small greenhouse with plants on Mars as a PR demo before they would even start the company. That, of course, did not happen. Before starting the company? Yeah, this was like the original thesis was like he went to some,

Elon went to some kind of space nerd meetup. It was like, we need to put a plant, a physical plant with a webcam on it on Mars. And if we land that there and we prove that life can live on Mars, we bring life to Mars, it will inspire everyone and marshal all the resources and all the capital and all the excitement to actually go put boots on the ground, which also he predicted. He predicted humans on Mars by 2024, of course, hasn't happened that we know of. It's possible. It's not one up there.

They keep saying that the star, they keep saying the starships blow up, but maybe one of them didn't. Maybe one of them sneaky went off and put a person on Mars. It's just the shell of the real ship. I like the inverse, the inverse conspiracy theory. You know, like, oh, yeah, we never went to Mars or we never went to the moon. No, the inverse conspiracy theory, we're going all the time, secretly. But you're not cool enough to get in on it. You don't have a ticket.

Everyone else does, except for you. Because I was just on the moon yesterday, and I'm not telling you about it. First crude Mars landing as early as 2029, a little too early to say. But that one feels aggressive. We're just four years out. You've got to put a whole crew down there. also said tourist trips were going to be happening around the moon by 2018.

Didn't hit that. Yeah. Self-sustaining city on Mars of 1 million people, 2050. Feels a little early, but I don't know. 2050 is a long way away. Any time you're getting more than a couple decades. A couple decades out. Anything's possible.

I agree, I agree. But then, of course, there's a bunch that he succeeded it. He said that he was going to deliver reusable orbital rockets. He did it. He's also predicted fully reusable starship with rapid turnaround. It's a little too soon to tell. He's behind some of the early dates, but in general, that project seems like it's real.

It's going to happen. It's taken a little bit. But it is interesting how the SpaceX narrative has shifted from first, we're going to Mars, then there was the whole, do you remember SpaceX point to point? Are you familiar with this? So there was a pitch for a while. Like a commercial, is this like, this is a replacement

for an airliner? You'll just go up and then come down and land. Exactly. Exactly. So it was like New York to Tokyo in like 30 minutes. That would go so hard. Insane, right? Insane.

So the idea was like, you're in men. Imagine being able to commute to Tokyo for some obscure job. It would be amazing. It would actually be remarkable. And there's a whole bunch of economic research that shows that when you increase transportation, reduce transportation times, you really do get an economic boom

because people can do business in areas where they couldn't before, even just reducing the time from San Francisco to LA, from a six-hour drive to a one-hour flight, all of a sudden, people can go up and do business, come back the same day. They just get there more often. I had a hot take for a while that if I was the president, I would recommission the SR-71 Blackbird,

which of course goes, I think Mach 3. So you can actually get from DC to London in like two hours, because this is a supersonic jet. And so imagine that, imagine the presidential SR-71 would go also very hard. Remarkably hard, right? Elon is always the king. of opening up new markets and just kind of saying like, well, what if, what if you was also

replacing commercial air travel? Yeah. And people are like, wait, that's a huge, huge market if you can do that. And of course, if you could make it more economical, who would ever pay for a ticket on a 747 when you could ride a rocket and get there literally 90% faster? Now, you have to go from Manhattan on a boat out to a launch pad. You can't take off from JFK. like there's a whole bunch of wrinkles and yeah and that's why this is all

practically probably like 20 years in the future but it was it was one of those famous examples of an Elon project where if it doesn't violate the laws of physics it eventually will happen and that's a lot of what's happening with with the data centers in space I think that the space data center thing it's people are still having the debate on like is it possible it's like obviously it's possible the question is is over under a gigawatt in by 2027, which is when the other big clusters come online. Will it be competitive in the short term, in the near term?

And then on the flip side, like, you know, if you're a bear, are you saying that it's not going to happen in 20 years? And then, I mean, the big question is, there's one more Elon Gambit that he could run with. One more like, oh, you wanted a fifth act? Just one? No, I mean, there's tons. But there's one that's really wild, which would be if he straight up said, we're building the Dyson sphere.

Like, we are going to launch. Tyler's nodding. He's pumped. He's pumped. Isn't he the obvious? Isn't he like the only real candidate? Well, he's the only candidate for everything in space because he has the leading space company by like, it's two orders of magnitude.

Tyler, the chat wants to know, did we ever find out if this kid, I'm assuming they're talking about you is related to Elon? Why would he be related to Elon? I mean, he's got a lot of sons. He's got a lot of sons. Oh, okay. Yeah, could be anyway. After years of resisting at SpaceX, now plans to go public Y, and Elon actually replied to Eric and said, as usual, Eric is accurate.

So I thought it was worth reading through this. Eric says, SpaceX is planning to raise tens of billions of dollars through an initial public offering next year. Multiple outlets have reported, and ours can confirm. This represents a major change in thinking from the world's leading space company and its founder, Elon Musk. Wall Street Journal and the information first reported about a possible IPO last Friday, and Bloomberg followed that up on Tuesday evening, with a report suggesting a $1.5 trillion target valuation. This is an enormous amount of funding. The largest IP owned history occurred in 2019

when the state-owned Saudi Arabian oil company began publicly trading as a Ramco and raised $29 billion. In terms of revenue, a Ramco is a top five company in the world. Now, SpaceX is poised to potentially match or exceed this value that SpaceX would be attractive to public investors is not a surprise. It's the world's dominant space company and launch, space-based communications, and much more.

For investors seeking unlimited growth, space is the final frontier. But why would Musk take SpaceX public now at a time when the company's revenues are surging thanks to the growth of the Starlink internet constellation? The decision is surprising because Musk has for so long resisted going public with SpaceX. He has not enjoyed the public scrutiny of Tesla and feared that shareholders' desires for financial return were not consistent with his ultimate goal of settling Mars. A significant shift in recent years has been the rise of AI, which Musk has been involved in since 2015 when he co-founded OpenAI. He later had a falling out with his co-founders and started his own company XAI in 2023.

At Tesla, he has been pushing smart driving technology forward and more recently focused on robotics. Must sees the convergence of these technologies in the near future, which he believes will profoundly change civilization. Raising large amounts of money in the next 18 months would allow Musk to have significant capital to deploy at SpaceX as he influences and partakes in the future. in this convergence of technology. How can SpaceX play in space? In the near term, the company plans to develop a modified version of the Star-like satellite to serve as a foundation for building data centers in space.

There you go, John. Has SpaceX been capital constrained, or are the capability constrained? Clearly, there's been, like, infinite demand for SpaceX equity for a very long time, so much so that people are, you know, investing in triple-layered SPVs just to get, you know, and enduring massive fees, just to get a taste.

Does this massive capital infusion actually accelerate the business that much? Or is it just the right moment in time for a number of reasons? I totally get that SpaceX was never crazy, crazy capital constrained. Even though, like, building a massive rocket, that feels so expensive. But at the same time, they've just been able to raise the money,

and it just hasn't been a problem. And of course, like SpaceX, as a business generates a lot of revenue. Like some of the contracts with the government are in the billions. And then Starlink is just throwing off, it's telecom, so it's just throwing off tons and tons of cash.

Yeah, and you just look at the growth opportunity and effectively telecom. You have T-Mobles, a $217 billion business that is threatened by Starlink. AT&T is $172 billion business. And Verizon is $170 billion business. But I think the reason you need $30 billion, like sort of paradoxically or counterintuitively might not be to build more rockets.

It might be to actually just buy the GPUs. SpaceX doesn't make those GPUs. They don't just pull them out of thin air. They got to buy them from somewhere. And if they're buying them from Nvidia, those are expensive. And you got to buy a lot of them. And for, to make a dent in, like, let's assume all the physics works. Let's assume that the launch costs are cheap.

All the math works out. Well, if you're putting a gigawatt of compute in space, you probably have like a billion dollars of hard cost just on the chips or something like that. I don't know the exact number. But you can imagine that you raise $30 billion. The business is humming, but you still have to outlay a ton just for the GPUs. You think that XAI gets rolled into SpaceX or Tesla? That's the hard part.

I thought it was going to be Tesla for sure, because Tesla has the lineage. They have their own chip. They have massive data centers to train AI models for full-sum. driving. It felt like such a logical place for X-A-I to land, but now maybe it lands at SpaceX. I don't know. Well, we should read through some of the history about Founders Fund's investment in SpaceX because it might be one of the greatest, it might be, it might become the greatest venture capital investment of all time. Because Founders Fund investing 20 million out of a 20, out of a $220 million fund too, which also

invested in Palantir and Spotify early is just nasty. And Dan Pramak says, flashback to chatting with Founders Fund, Luke Nosek, when he led the first investment in SpaceX. I spent some time on the phone earlier this week with Luke Nosek to discuss his firm's $20 million investment in private space launch services provider, SpaceX. The company is run by Elon Musk, who co-founded PayPal along the Founders Fund partners, a few questions and answers. Dan, Founders Fund usually invests no more than a few million dollars in a company.

Why invest $20 million when your fund is only $220 million? We obviously have internal caps that we follow. What we do is look at how much a company needs and invest that much. For example, we only invested $500,000 in Facebook because that's how much it needed at the time. A lot of firms were offering more. Historically, like the greatest venture investment ever, in my view, was Mossa's $20 million investment in Alibaba. That's true. Turned it into $70 billion at the time of the IPO.

Hard to, I wonder if we can figure out how much. FFone's about 10% of space. But I don't know how many, they've put in billion. Oh, yeah, yeah. To maintain that stake, they put in time. So, again, it's still, like, going to be a $150-ish billion dollar position, just the carry check alone will be nasty, as Nico said. Larry Allison is addressing, he's sitting on stage here in this Wall Street Journal article

on the news that Oracle shares have tumbled as AI spending outruns returns. So Oracle is facing mounting anxiety from investors about how much it's spending to build out data centers for the artificial intelligence industry. The cloud computing company's revenue and operating income for the most recent financial quarters fell slightly short of analyst's expectations while the company raised its spending forecast,

adding fuel to concerns over the timeline for turning the AI industry's ravenous demand for computing capacity into profits. I think investors didn't realize that like when you do the one of these crazy AI deals, you have to go build the data center and then you have to wait and then you have to get power and you have to buy the chips,

and the chips, and then you've got to turn it on, then you've got to test it. And then you can generate some tokens and hopefully sell them. All well, enterprise AI adoption might be stacking. We'll see. There was some crazy data out of ramp today. Ara Karazian, the economist over there.

At least some businesses might be saying, we've got enough AI. We like it, but we have enough. Apparently 55% of businesses are just, like, good without paying for AI? I mean, the paying thing is- At least directly. The paying thing here is, so basically, and just the examples, like a lot of businesses don't pay for cloud.

Yes. But they effectively pay for cloud because they pay companies that leverage cloud to deliver their products. But it's an interesting chart to see where Enterprise AI adoption is going. Ramps AI Index shows AI adoption held flat at 45% of businesses in November, driven by slight declines in finance and technology sectors by the model. Open AI went down negative 1%, Anthropic up 0.8%, and Google went up 0.7%. And so you can see that this chart in 2025, Open AI had a big, big jump. And then just has been flat for the last couple months, a little bit down.

Anthropics been on a much smoother, it seems like they didn't exist in 23, basically, or maybe they didn't have a paid plan that really appealed to businesses. Then 2024, you see some slight growth here. And then 2025, a much smoother, like, okay, uptick, growing more, growing, more, growing more, kind of all keeps up with this, you know, they've been 10xing every year, like an Instagram hustle account. It seems to be pretty obvious at this point that the model is just getting smarter is not going to, by default, make adoption, like, accelerate massively again. They need to become more useful.

They need to be able to, you know, do our kesh, has talked about this. Like, they need to be able to learn on the job, basically, in the way that a human does. And I think, I think until we have that, there's some thing, you know, the forward-deployed engineer movement of, like, having people out embedded in companies, like, trying to unlock the potential of the models. That's great. But I think the models actually just need to become generally more useful. 45% of businesses don't pay for AI, don't use AI. And the question is, like, are they just not? not using AI at all or are they just good with free?

Because if you told me like 55% of companies are in sort of like a private equity, scarcity, wartime mode where they're watching every dollar that goes out of the door, even a $20 subscription is something that will be scrutinized because you know what? Like the son, we've been running this business for a generation and we're not gonna spend frivolously. Every dime that goes through our business is precious.

And so we're not just gonna go sign up for whatever. No, you don't just spend, spend, spend. I do think there is a bit of a slowdown here. It's just getting harder and harder to justify, okay, it's a no brainer to bring it in, it's a no brainer to pay for it. If you're a business that's not doing code gen, you're not generating a bunch of assets, et cetera,

and somebody says, hey, for $20 a month or $100 a month, you can get somebody that's PhD level and math, but extremely low agency. You have to tap, you have to tap the person on the shoulder. I like the laundromat. The laundromat's like, actually, I don't need somebody who's PhD level in math. You know what? Sir, this is a laundryman.

We've been running this laundromat and this family for 45 years. And I've never run into an IMO gold medal level problem ever. Disney, which is investing $1 billion in Open AI, and is going to license their characters for use in SORA. She says, look at this hype, isn't it neat? Wouldn't you think my bubble's complete? So Disney's making a $1 billion investment in Open AI and will allow the AI platform to use its characters and properties to generate short user-prompted social videos. Disney's three-year licensing deal will let users generate videos using SORA of more than 200 Disney Marvel, Star Wars, and Pixar characters.

A curated selection of these short videos will be available to stream on Disney Plus. Whoa, that's the bombshell there. I didn't realize that. For a kid who loves, you know, a Pixar character, they want to see Wally join the Avengers and go fight Thanos. Like, that could be a delightful, delightful experience for a young child who enjoys Disney's IP. And yet it could make no sense for Disney to actually create that. I just think the idea that this is going Disney Plus is crazy. That's the crazy thing. Is there real demand for SORA level quality content?

in the Disney Plus app. To me, what's exciting about this is personalized entertainment, and which is why I think that Disney would be interested in doing this. One thing that's notable, this went out this morning. Disney accuses Google of using AI to engage in copyright infringement on massive scale. As Disney has gone into business with OpenAI, the mouse house is accusing Google of copyright infringement on massive scale,

using AI models and services to commercially exploit and distribute, infringing images and videos. So the question, so Disney's not just allowing OpenAI to generate these assets, but they've actually invested. And so the immediate question is, will Disney do any of these licensing, IP licensing with other platforms? I don't know. It does feel like Disney's sort of like picking a winner here in AI video. And they're saying, like, we like Open AI and we don't like Google. which is sort of a bold move. You would think that they would be a little bit more platform agnostic. I'm fascinated by this.

Then again, Thrive Capital, bet the fund in some ways on Open AI. Who's co-owner of Thrive? Bob Eiger. Get out the red string. It's all connected. Nissan says GVT 5.2 Pro. It has ridiculous pricing, once again, $20 and $168. Tyler, do you know if GPD 5.2 Pro is more expensive? I guess the raising prices. Yeah, I mean, so this is the pro mona. These are always like the very expensive ones. Yeah, I think it's pretty comparable. I guess 5.1 was one $1.75 and $1.14 and $14. So it's a pretty significant. 5.1 was a $1.25 on the input. It's like, you know, marginally more expensive.

We're eating good. Future pipelines may need to unify pre-mid post training, in checking. reasoning data earlier and more continuously, to anyone wondering if 5.2 is a new pre-trained, I'm letting you connect the dots between this, says Alexander Doria. This is the new main catch of deep seek 3.2. This isn't a new model card, new ELO update. It's an architectural update. GPD 5.2 thinking is at 52.9% up from 17.6%. That's a huge leap. Wait, we got to run our joke benchmark. Yes, let's run some of the TVPN benchmarks. Pull it up.

Pull it up, Tyler. Try to make us laugh. So, yes, this is my benchmark where I ask. Should we create the joke about the shrimp fried rice? Yes. So I'll just go through some of these and you guys tell me how good they are. Yes. You're telling me a crab rang this bell?

Crab rang bell? This is 5.2 thinking, right? This is 5.2 thinking. This is the frontier of AI. I kind of like, you're telling me a crab rang this goon? That's kind of funny. Yeah, I think that's what I would have thought. But that's not it.

You're telling me a bear built this market. Oh, because markets can be bullish or bearish. A bear built market. You're telling me a ham. You're telling me a ham wrote this lettuce? A ham wrote this lettuce? Ham wrote lettuce? Wait, wait, this is a good one.

Terrible has happened. You're telling me a ghost wrote this script? Ghost wrote script. Ghost written? Ghost written? Okay, yeah, yeah. You're telling me an owl delivered this mail? Yeah.

So I think, I think Gemini is still the frontier here. In terms of, in terms of shrimp fried Reich's bench. Yes. Yes, because it were some good ones there. But, but again, to understand, it seems like Gemini understood the assignment, but still ultimately leaned on Reddit data and broad Internet data and did not come up with any unique joke that had never existed on the Internet before. Yes.

But it did find the good jokes on the Internet. good ones that it gave me I could find elsewhere. Alright let's try at least one more. I wonder if this benchmark is like saturated in the sense that like every possible shrimp fried rice joke has been said and written down on the internet. Is there

any new territory? Like if I put you on this job for a month, could you come up with a new joke that's as funny as shrimp fried rice? Yes. You think you could? I think so. Okay, maybe that's a challenge. Do it. You're telling me a chicken actually tendered these? You're telling me a chef, boyard, eat these ravioli?

Okay, yeah, this is really, really rough. All right, you just bombed, Tyler. Google names a new chief to head up its $100 billion a year AI infrastructure build-out. Amin Vadat, who will report to Sundar directly. This is very exciting. Imagine having a $100 billion budget for CAP-X. That's like Ben's dream. Four times NASA's annual budget.

Yeah, Ben. So Ben's a CapEx guy here, almost every time he walks up. To me, he wants another $30,000 for a new camera, this light, this. It's going up next to you. One more. Meanwhile, we are actively on the show asking for goalpost that will probably wind up costing 30K. We're part of the problem. We are.

We are. I want to watch this video of a hillside in China that has been covered with solar panels. When you see this video, do you react with admiration or with disgust? Tyler's nodding. This is sick. This is sick? Yeah, for some reason, solar panels over mountains do you give me a little bit of disgust, whereas I get, I'm less.

Yeah, mountains are beautiful. Deserts, deserts. I'm like, blanket it. It's sort of already. Blanketed. This looks like it would be a really enjoyable mountain range to just go on a nice little hike. And so I'm feeling the disgust, but also impressed. I mean, at least the mountains won't catch on fire.

You know, that's often a problem. Yeah, I'm curious around the efficiency. because just given the movement of the sun, aren't you gonna have, like, it's actually blanketing both sides of these peaks? Trump says doesn't see why we can't have 20% or 25% GDP growth. It says the market should continue to go up with great results.

Don't see why we can't have 20% GDP growth because it's like never happened in history. Maybe. He is AGI-I-pilled. That's extremely AGI-I-pilled. Borkech has talked about this. Dracch literally has talked about this. 25% GDP growth, the first AGI-pilled president, I suppose. Apparently, China hit 19.3% in 1970.

Okay. Yeah, is that the greatly forward or something? More recently, more recently 14% in 2007, so I think Trump's looking at the data. Greatly forward was 1960-ish. First Lady Melania Trump is introducing the presidential AI challenge. artificial intelligence is America's next competitive edge, driving advancements in your career, supporting your family, and strengthening your community. This is why I launched the Presidential AI Challenge a nationwide call to students and educators

to shape America's future. Teams from all 50 states have already registered. Shape the future. Be part of the Presidential AI Challenge today. What is the Presidential AI Challenge? It's like a riddle. This is a riddle. It's some sort of like, you know, it's shrouded in mystery. That's the whole goal to test the American people. Can they understand what it is? Okay. I think I figured out. So there's elementary school, middle school, high school.

And there's challenge projects that groups of students will be able to participate in, as well as awards and prizes. Hopefully, solid gold, solid gold trophies. Massive trophies. There actually is a cash prize. Sick. $10,000 per team member. Wait, is it Peyton Melania coin or Trump coin or U.S. dollars? I think the, uh, I think you mean from meme.

Trump meme. That's right. It's not Trump coin. Get it right. It's Trump mean. Okay, here's some questions. You also win cloud credits. That's so. You do? Okay. That's actually amazing. I love that. A presidential award certificate. So boom certificate. You're getting cloud credits. You're getting 10,000 for your school, homeschool, or community group. And you're getting $10,000 per team member. in the middle school category and the high school category and the educator category. Tyler, you, I guess there's no college tier. Too old.

So you have to go back to school and participate and win. Get a fake ID that says you're 13. Just to go back and dominate the presidential AI challenge. Leave us five stars on Apple Podcasts and Spotify. Please do. Have a great rest of your day. And Merry Christmas. Goodbye.

Thank you.