Elon Paid $8 Billion For Potentially Dead Airwaves | The Brainstorm 153

0:00 If the world is going to get hacked, it's actually more likely to get hacked by openweight models than closed. >> What about quantum encryption? Is that crackable with AI? All right, SpaceX coming for the jugular of AT&T, T-Mobile, and Verizon picking up Nexttel's old Spectrum and according to Musk, the last piece of the puzzle for direct to sell. And then OpenAI absolutely demolishing mathematicians worldwide with a massive release of proofs, 722 proofs out there. But I think the more interesting question that we'll dive into is whether or not this is a risk for crypto and how this is kind of butdding up against the fear for quantum. Right? Maybe we'll start there and in all of these proofs, one, how significant do you think they were? Two, notably nothing on cryptography.

0:58 Is that because if they release it into the wild, it is too dangerous and everyone's, you know, Bitcoin wallets then immediately go to zero? or is there something else at play here? I think I think there's one it's definitely a demonstration of capability in that you know these are the sorts of proofs that working mathematicians work towards as you know multi-year projects that they then publish on and get recognized for. And so the mathematical community is some excited, some very upset that OpenAI is speedr runninging all of mathematics and seems to be able to do so with you know the people they have on staff and some number of hours of compute time with their super powerful models. and I think that everybody noticed that across all of these proofs, many of which are, hey, these are interesting and could lead us to other kind of discoveries in formal logic, some of which have, you know, potential commercial applicability. there was this notable lack of of of advancement in cryptography. One way to interpret that is that they recognize these models are very powerful and you know probably in collaboration with the US government have said hey we're going to withhold any advances we have in this particular area so that we the US and US enterprises can harden our defenses against the crack in cryp cryptography. Another interpretation would be well there's so much money at stake with you know you know crypto cryptographic architecture that these are areas that there's already been in commercial incentive for people to go after and try to probe and try to advance the frontier and so it's not as easy pickings for an AI model to unpack it. and so the I I think it's you don't necessarily know which is the truth. I do think it's accurate to assume that you know formal logical mathematics is going to just fall to this super intelligent call it beast that we're building. and that does at some point begin to put you know some of these advanced cryptographic techniques under risk. And I would assert actually there's greater risk from AI than there is from quantum computing to kind of like certainly penetrating kind of cryptocurrencies broadly. but the risk is not necessarily in the cryptographic primitives but in everything else that tries to secure your blockchain and that all these organizations have multiple risk vectors and and this is just an indication of power and how risky it all could get.

3:46 Isn't the risk actually that it's not OpenAI that releases something that is able to crack cryptography, but that the rest of the field catches up to where Open AI is and the frontier is in 6 to 12 months. >> Yeah. I mean, I think that that you can you can see what is there is what's publicly available now which across anthropic and open AI you can pay you know nosebleleed prices for best in the world intelligence. then you know that behind the scenes they have something more powerful that they're able to work on and that you know they're driving forward to kind of uncover these vulnerabilities and then you have what you could run you know in your own cloud implementation and do whatever you want with without anybody being able to stop you and that trails kind of like the frontier by you know six or 12 months and I think it's reasonable to assert that you know Maybe they the industry as a whole lags the frontier labs. Maybe it catches up.

4:50 I would tend towards lags, but this capability will be available. And so the open AI and and anthropics approach is one. We're going to be careful about what we release to the public because we're going to selectively provide you know infrastructure providers with this protection service. so that when kind of the nefarious actors have access to this intelligence, we have a more hardened infrastructure. and yes, I think if the the world is going to get hacked, it's actually more likely to get hacked by openweight models than closed.

5:22 because you know, those were will be where the nefarious actors will operate. >> And great, as we said in previous episodes, for the Frontier Labs to step in and say, well, great, here's our cyber security package pay up. but we've also seen a lot I don't I don't know if I can say a lot. We've seen some early Bitcoin holders moving coins. Do you think that's in direct reaction to these fears or something else at play?

5:54 >> Well, I think there's been a lot of kind of the the the danger of self-custodying seems to be going up just broadly. There was you know, there's various like self-custody hardware wallets that have had you know, bugs in their code that have been successfully exploited to steal coins. there was this crazy supply chain attack that that plagued I think it was cold card no sorry Ledger where you know they they inserted a chip that like just broadcast your seed phrase out to them as soon as you plugged it in. and so I think people generally feel like the power of kind of combination of these tools and the sophistication of the people trying to go after kind of their self-custodied coins means that the big infrastructure providers who can pay more for security just might be a safer spot to keep your coins. Now, that also means there's these large honeypotss out there, and you could imagine rather than all of cryptocurrency being called into threat, some exchange actually does get severely hacked in a way that is very disruptive. and so I don't think the entire cryptocurrency architecture gets pulled down by its cryptographic primitives, but instead you have a really really large kind of hacking type event. that's potentially inbound here depending upon you know how security conscious and how aggressively these you know cryptocurrency exchanges are in spending on frontier level intelligence to protect themselves.

7:35 and so >> does this is this one of the fundamental I feel like okay Bitcoin there were some fundamental theories about its value and it's like unit of exchange out the window right it's like >> I feel like >> account you mean but >> yeah unit of account and then is like self-custody or the distributed nature of it and the fact that it is like a bearer asset that you can take with you in your brain and not have to Scrooge McDuck it. It's like, is that going away? And if that goes away, is that a fundamental problem or have we crossed the chasm where it no longer really matters?

8:18 >> Well, as long as you still have the option value of doing that, I still think it retains that useful characteristic. and kind of like the central infrastructure providers have have taken advantage of some of those characteristics to provide like lots of other benefits on top of it. The the ability to you know what's happening in crypto broadly is all of these traditional financial instruments are getting converted over. people are converting it over not because of like philosophical conceits but because hey this makes it like easier for customers to borrow against these assets and to transfer these assets and so it becomes actually more practically useful rather than useful as this insurance system against kind of like everything goes tapioca and by the way you know I I think that cryptocurrency isn't the only set of financial institutions that are at risk from these frontier models getting really powerful and it's in performant. you know, you could imagine some European bank wakes up and its accounts have been totally scrambled. Now, the the hacker couldn't couldn't take and run with the dollars in the system, but they could certainly cause a lot of damage. And you know, I don't know how we would kind of recompile those accounts if if like there was a real fundamental attack on a systemic financial institution which had underinvested and that that would cause in some ways like a much greater disruption because everybody would need to freeze and try to figure out well like where was this money? I don't know what my account balance was.

9:50 Do you know? and so that would be like very very very meaningfully disruptive at least with cryptocurrency. If somebody steals the coins, you know where the coins are. They're just happen to be in the wrong person's hands. and so I I just think that there's there's there's a bigger nest of potential kind of disruption out here in financial land. >> Cash is king. >> Cash is king. >> Cash is king. You want to be losing losing half your the value of your money every seven years.

10:20 Well, that's this is the argument for like real assets like if you know presumably the property record is not wipeoutable. and so if you have proof of ownership of a piece of land, you can at least point to the piece of land and be like, "That's mine." whereas if your bank account comes back different, you know, next week versus the prior week, what proof do you have that it's been changed? it's a little harder. >> Well, you got to be able to defend your land, >> but that's a that's if things really go tapioca, I I've never heard that phrase before, but >> yes. Yes. Everybody loves Boba Tea.

10:54 >> all right. Any any other comments here before we move on to SpaceX? I think it's just widely I I think we can't really tell whether or not this stuff is going to be crackable. But this is another example in which the acceleration in AI is actually stealing some potential commercial opportunity from quantum computing which I think is interesting where it's kind of people are like well at least quantum computing will be useful because we'll be able to crack crack these cryptographic you know codes but if AI gets there ahead of it you know why bother funding the quantum computer when you can grab all the Bitcoin beforehand using super intelligent AI.

11:32 >> What about quantum encryption? Is that crackable with AI? >> Some say it's easier to crack >> with AI. >> Yeah, the the the quantum the the the quantum safe encryption schemes are more complex and less tested. And so if if anything, they're like a little more tenuous here because they probably have more vulnerabilities that haven't been discovered that AI would then crack. And so even the the life raft people were hoping to plunge into because they're so worried about quantum computing, maybe it has more holes than they think.

12:04 And you know, I'm at least publicly and and personally skeptical on the timelines on quantum computing. I think we're at least a decade out, if not more. and this to me seems like a much more realistic risk vector against, you know, the entire industry, but not just because of cryptography, but because of all of the other things people are doing along the way to try to protect their coins where there's plenty of vulnerability as well. All right. SpaceX eight rumored $8 billion for this spectrum. roughly 800 what is it? 800 meg meghertz. No, that's the wrong >> Yes. Meghertz, right? Okay.

12:48 >> I think so. Yes. And so this unlike their higher spectrum bandwidth is low throughput but strong penetration capability. So where they couldn't operate before indoors behind thick panes of glass etc. weather this theoretically solves that problem. And they were also approved for 15,000 directto to cell satellites from the FCC.

13:21 How much of a problem is this for the big three telos in the US? >> And Nick, are you going to sign up for the SpaceX Starlink cellular plan when it's available? >> Maybe. I mean, I don't get great cell service as soon as I leave my house for about 5 minutes in any direction. So, if it can solve that, that would be nice. But I don't I my my question is how much time do they need to get to full coverage equivalent to what AT&T and Verizon have?

13:56 Yeah, I think it's a there's the SpaceX plan is provocative and and really there's this interlinking between their call it broadband business and their perspective cellular business that is what makes it potentially possible and super disruptive to the Telos and is the reason that the you know AT&T and T-Mobile are out there being like this plan doesn't make any sense. We tried this before. It's not going to work because their idea is with this spectrum and with all of the dishes that are popping up on everybody's houses like mushrooms, they can you know provide downlink service from their satellites to the people's dishes on their houses and then transmit it from the dishes to the individual cell phones you know throughout the a any particular city.

14:48 And so your coverage will be dependent upon how successful their broadband business in the city that you live in. and so kind of like selling >> Yeah. >> Yes. They have to sell in the broadband in order to also kind of like have the cellular service work in a consistent way. I do think they can work around that by also basically subleasasing kind of bandwidth or or basically network throughput from cellular providers in so far as those cellular providers would be willing to give them the time of day to but otherwise there is this kind of startup problem of they have to sell in a bunch of broadband in order to make the cellular thing work and you do have to have good coverage on the cellular side to get people to rely on it as like the first the only provider that they're going to rely upon.

15:43 >> But and in terms of timing, you know, I think this is a four year four to five year five is a long time in in the world today, but four years it's like you look at the progress with Starship, how quickly they can send stuff up there. I think, you know, this could be knocking on people's doors faster than AT&T, Verizon or T-Mobile really wants. Obviously, they're not standing still either. You know, they're partnered largely with EST Space Mobile.

16:21 but it feels >> like it'll be a fundamentally different offering and I and it, you know, over the past three weeks. Yeah. Go ahead. No, we I mean we originally approached thinking or you Tam originally approached thinking about this market of saying kind of hey this will be SpaceX will be able to clip like a dollar off of other cellular providers plans by providing kind of like the emergency you always have connectivity albeit higher bandwidth connectivity than exists for like the Apple product and stuff. and so they just are a fill-in service whereas now you know and and that's kind of how T-Mobile was comfortable with partnering with them but now they are because they have this physical infrastructure they're going to install they can go after you know the full you know monthly package you know whatever you pay SpaceX is going after all of that rather than a buck or two per month from the subscribers so it's a big change in kind of their positioning and strategy and you know materially increases the addressable market they're going after.

17:28 If you look in the US, you know the the broadband market is 170 billion annually. whereas the cellular market or this is what we think it'll be in 2030. The cellular plus broadband is in excess of 500 billion annually. So kind of they're really expanding their TAM here in a way that could be material to the story. here's >> owning the customer relationship. It's like there's that's obviously part of capturing that dollar value, but the strategic difference there is pretty fundamental as well. Have we and maybe this is a dumb question, but have we done the math on what I don't know if you'd call it dish density you need to be able to service like a relatively populated area? like how many Starlink dishes do you need to actually provide relatively good coverage in an area of like say 100,000 people >> that we have not done the math >> they have not >> next week you provide the numbers >> yeah yeah >> homework I can say that >> but but their their ability to sell in broadband is partly as in the service they can offer is going to improve over time per cost in a in a very material way And so, you know, right now they they one on the broadband side almost nobody uses all of their spectrum or all of their bandwidth as in people get one gigabit per second. You know, they they get one gigabit per second but are never using that really they are buying as a proxy for I never want to run into this situation where suddenly like it's the internet that's the problem with me communicating. and so, you know, SpaceX is selling in a service that, you know, is plenty for everybody, but it might, optically look worse, than than other services for the same price. but the thing that people really want is reliability. And if you have more satellites up and you have more bandwidth up, we think their bandwidth is going to increase from like 900 terabits per second by end of this year to on the order of 24,000 terabits per second by 2028 year end.

19:46 so you're going to have this huge increase in kind of like the throughput of their pipes. and so that that should mean either they're able to advertise to people, hey, you can get the same speed guarantee at a lower price on the broadband side and then cause that to, you know, people to go direct to them as opposed to using it as like backup service. and and then that should then seed the market to allow the cellular business to work.

20:14 what does that throughput increase translate to in terms of customers like for you know does that mean they can service 10 million 20 million 30 million 40 million how many households? Well, so you can think it's like the >> I mean what I'll go to fast.com right now quickly and see what I'm getting. I'm getting you know 8 850 megabits per second. But it's like how much am I actually using right now doing this recording?

20:49 15 probably. And it's like what else is happening in my house? I don't know. Some thermostats etc. It's like I'd be shocked if I'm using more than 40 right now. So that leaves an extra, you know, 20 people could be on my internet doing exactly what I'm doing. >> Mhm. >> Yeah. and and it in in their filings and actually a lot of people model the company by saying well you have this many customers this many subscribers and this is how much they pay per subscriber and that's coming down but their subscribers are growing and this is how we triangulate to revenue. I think that's actually too intricate and complicated and as it turns out it's not the right way to forecast the business because you know how do you count the customer who's in an airplane who is paying you know $50,000 or $100,000 a year for the dish on the on attached to the single airplane. You know that's one customer but you know so they're not like for like comparable. So the way to think about the business is they have bandwidth up there and they can monetize the bandwidth at a consistent but declining rate over time.

21:58 With that we get to kind of like and that will include the cellular business and the broadband business. And then enterprise you know yeah enterprise government government will pay up because they want always on connectivity through Starshield. net we get to them getting on the connectivity business to $500 billion in revenue by 2032 and around 400 billion by 2030. so it's they lay out this total market that they can address of 1.6 trillion. This would suggest they're 25% through what they see as their total addressable market as they kind of saturate the space. you know 500 billion is approaching saturation where marginal satellites up are just maintaining the bandwidth they have in play. and there's huge execution that has to happen between here and there to do that. They have to get the broadband business working. Then they have to provide this cellular business and then they have to begin selling direct to to the cellular customers and I can throw away my my Verizon bill and actually get decent coverage in the second largest city in the US which is insane that I can't and and they'll almost certainly bundle intelligence into the package and and that's actually you know so you get your connectivity and your Grock bot subscription allin one it will be an interesting >> it's finally the distribution that Nick needs to belief.

23:24 >> Except for Grock is horrific. >> But Grock, >> outside of that, >> compared to Muse, come on. >> Oh boy, >> you guys are under the Elon spell. I can, you know, go in and out of the spell. >> Muse Muse is so dysfunctional for me, but who knows? >> They just showed there was numbers that came out over the the weekend. it was sensor tower data on MW's retention figures.

23:54 Very strong. >> Yeah, I saw is is comparable to chat GBT. >> Yeah. >> Stronger earlier and then tailing off slightly worse by like day 14 or 24 or something like that. >> Yeah. But like barely, but better than every other chatbot >> agent service out there. Also, Mark Zuckerberg was hinting in his videos over Watermelon, the new model that they plan to release soon, as he put it. >> He said it's very big. That's what he used to describe. That was his term. I I do think it is interesting that on connectivity like the cost of not being connected is now so much higher than it used to be because like suddenly you're cut off from all of your super intelligence tools if you're not connected. so you know, do not fly Delta Airlines over the Pacific cuz they will not give you connectivity once you get offshore the US, which is really frustrating because, you know, there's Starlinks like whizzing over your head that you could be connecting to. but they have a geostationary satellite they're relying upon that clearly just hovers over North America. And so suddenly you're like, "Oh, this flight has Wi-Fi." It's like, well, no, it if you're flying to Australia, it has like 3 hours of Wi-Fi and 12 hours of no Wi-Fi. and so that's >> Well, yeah. So, maybe on that point, I have a question. Okay, so you go back you go back two years, time to power is, you know, around $12 million per megawatt. this year obviously jumped up, you know, some numbers in like the 50ish million dollar per megawatt.

25:40 Does that keep going up? Does that come down? It's like the two factors at play simplistically in my mind is you know supply of compute and capability of the models, right? Right? It's like as the models become more capable, they should be worth more. And if there's still a shortage, that would go up. If there's increase in supply of chips, then it comes down. Models are improving at 99.99% annually.

26:14 >> Doubling every month is a way to >> right. And chip production is not doubling every two years. So >> yes, >> does that price keep going up? Does it come down? What's the equilibrium in your mind? >> I can tell you how we model it and I think it's like a reasonable guess which is that the the basically if you're renting at the margin today you have to pay $40 billion per g gawatt. If you're renting at the margin by 2030 you're going to have to pay 50 plus. And that's just increasing it at the rate of nominal GDP. I think there's plenty of things that could drive it higher, but that seems reasonably conservative.

26:55 What's interesting about SpaceX though is because they're all renting at the margin. Like SpaceX builds and is renting short-term, right? And so they get to collect like the the premium that those new chips have, whereas like you hear that 40 and 50 billion number 50 billion is what Google paid to them. You say, how are they getting that? Cuz like Cororeweave and these other Neoclouds are only getting like, you know, 10, a little over 10. But it's because core and the other neoclouds are having to sign fiveyear deals where they get the premium for the first year where the chip is really valuable and then the chip falls off in value over time.

27:30 and so but they have to do that because they need kind of that assurance in order to finance the data center build. Basically they're trying to like get money from customers guaranteed so they can get capex that their financeers are willing to give them. as a result, you could see these Neoclouds continuing to get, you know, around 20 billion dollars per gigawatt even as like the newest build stuff monetizes at 50. But SpaceX sign shining short-term deals could collect between 30 and 40 across their fleet.

28:01 >> but short-term deals also means they'll be at >> it's like, you know, you're you're at the long end of the poll for fluctuations. they will be on the hook if if suddenly there is over supply and so the reason they can do that though is because that one they have strong balance sheet and Elon has you know the banks you know wrapped around his finger call it like he can get financing in very very large size without having to sign these long-term contracts and SpaceX is sitting there and saying listen if there's if there's supply over supply that's fine we'll use it for our own internal models and we'll do you we'll serve Grockbot better, we'll do more R&D to to lead to the next stage.

28:45 So, he's taking on the risk of kind of like fluctuations in the demand market because he has, you know, internal use for it just like with launch, you know, they they they basically soak up all the demand with with Starlink and then kind of like filter stuff at the margin out externally. And they'll do the same with Starship, I assume. And I think they're going to do the same with compute here. And it gives them really a superior kind of monetization position with the risk of like hey suddenly you know I don't know >> the next hundred billion dollar per gigawatt isn't coming to this corner.

29:27 >> What what percent of SpaceX launches are used for Starling and SpaceX's own purposes? I feel like I've asked this question but I think >> it's it's like north of 80%. Oh, really? Yeah. Yeah. Yeah. It's like I think the last >> Well, that was like ride share missions. They only have like singledigit of those, but then they have more that are specifically contracted out to non SpaceX companies. >> but the vast majority, >> it's like the government stuff. NASA is is like, you know, those are missions that not only are they they're doing like additional work to facilitate those missions that where they're devoting engineering talent to it. But on Starship, they, you know, they they are trying to force the industry to get into like this. It squeezes out a PEZ dispenser door stance. and so that you know for now they they on a strict narrow economic sense it's unclear that they should give anybody any capacity on that other than their own because they have such great ROI on just launching their own stuff right now.

30:35 >> Yeah. All right. Just over 30 minutes. Thank you everyone for joining another exciting week in next week we'll talk about flying cars. That's like >> go touch grass. Yes. Go touch grass while you still can. >> Go touch grass in a in a low connectivity area. Get off your phones. >> That's right. Low connectivities are the best connectivity areas. >> Yeah. >> And of course, yes, Brett teased it next week. likely reporting back on the newest Roadster and whether or not it is joining the other Tesla Roadster in outer space, but this time self-propelled. ARC Investment Management LLC is an SEC registered investment advisor. ARK, Wolf Financial, and Public are separate, unaffiliated entities and do not have a relationship with respect to any company marketing or selling the products or services of the others. ARK, Wolf Financial, and Public make no representation or warranties regarding and expressly disclaim any responsibility or liability for any acts, omissions, losses, or damages of the other entities or their affiliates, officers, employees, or agents. No company shall be liable for any indirect, incidental, consequential, or special damages arising out of, or in connection with this podcast or its contents. The information provided in this show is forformational purposes only and should not be used as the basis for any investment decision and is subject to change without notice. It does not constitute either explicitly or implicitly any provision of services or products by ARC. An investor should determine for themselves whether a particular investment management service is suitable for their investment needs.

32:14 All statements made regarding companies or securities are strictly beliefs and points of view held by ARK and/or show guests and are not endorsements by ARC of any company or security or recommendations by ARC to buy, sell, or hold any security. Historical results are not indications of future results. Certain of the statements contained in this show may be statements of future expectations and other forward-looking statements that are based on ARK's current views and assumptions and involve known and unknown risks and uncertainties that could cause actual results, performance, or events to differ materially from those expressed or implied in such statement. ARK assumes no obligation to update any forward-looking information.

32:56 ARK and its clients as well as its related persons may but do not necessarily have financial interests in securities or issuers that are discussed. Certain information was obtained from sources that ARC believes to be reliable. However, ARC does not guarantee the accuracy or completeness of any information obtained from any third party.