Anthropic IPO Valuation, Supercomputing Insights and Amazon Satellite Setback
The Anthropic IPO valuation leads the November 19 discussion on Techstrong Gang. Mike Vizard and Alan Shimel host Jon Swartz, Stephen Foskett, Jaime Hampton and Drew Gutstein. The panel also covers supercomputing news from SC26 and a rocket failure that set back Amazon’s satellite plans.
Anthropic Valuation
Anthropic’s IPO is expected to price at $70 to $78 a share. The stock would then trade on Nasdaq under the ticker ANT. Meanwhile, reported demand is about six times the shares on offer. Coverage asks whether Anthropic’s IPO could make tech history.
However, the numbers behind the offering pull in two directions. The panel weighs $17 billion in third-quarter revenue against $900.9 billion in disclosed purchase and lease commitments. As a result, that gap sits at the center of the Anthropic IPO valuation debate.
Supercomputing Insights
At SC26, Nvidia introduced the Rubin HPC GPU, which it plans for 2028. In addition, the company announced Nemotron for Science for scientific AI agents. Reporting details how Nvidia unveiled the Rubin HPC GPU and Nemotron for Science at SC26.
Hammerspace also used the event to unveil a data platform that isolates metadata to speed AI workloads. A related story explains how Hammerspace unveiled a data platform for the agentic AI era. Therefore, the panel looks at what both announcements mean for AI infrastructure.
Satellite Setback
A United Launch Alliance Vulcan rocket failed about 10 minutes into its November 15 launch. As a result, Amazon lost all 40 low-earth-orbit satellites on board. Coverage explains how a rocket failure dealt Amazon’s satellite ambitions a major setback.
Amazon plans a constellation of 3,232 satellites, and 398 are in orbit. The panel therefore discusses what the loss and a likely delay mean for that plan.
Together, these stories show how AI ambition meets financial and physical limits. First, the Anthropic IPO valuation tests how much risk public investors will accept. Second, the SC26 announcements and the launch failure show how long big infrastructure bets take to pay off. Watch the full conversation and catch new episodes every weekday at noon ET on Techstrong TV.
Transcript
Hey. Hello, everyone. What a crazy day.
Hey, mark your calendars at noon on the day Anthropic went public, you were watching Techstrong gang. Thanks. Thanks for that.
Thanks for the vote of confidence. Let me introduce you to our gang today, because I have a feeling it's going to be a lively discussion, as they say. Joining us fresh off his standout engagement at Supercomputing 26.
It's kind of like a Las Vegas-- What do they call those Las Vegas engagements when you're there a while? Anyway, it's- A residency. A residency ...
residency. Fresh off his residency at Supercomputing, Steven Foskett. Steven, good to have you.
My heart will go on. Yeah. Okay.
Are you going to break out in song for us? I might. I'll stop the show right now if you're going to sing.
I'm here listening. No. Okay.
No. That was enough. That's enough.
That was enough. From the Space Coast-- Well, she's not really at the Space Coast, she's more on the other coast, but she's close enough. Our own Jamie Hampton.
From New York City, Drew Gutstein. And of course, Mr. West Coast and Mr.
East Coast, Mike Vizard and John Swartz. I'll let you guess which coast is which. So guys, another interesting day for our Techstrong gang.
I don't know how we can just pick three stories, but the lead-off, I guess, Mike, has to be- Well, I would like to talk about the most consequential thing in IT that's happened in eons. "Grand Theft Auto VI" is available for download today. I heard that, and I heard there's all kinds of server congestion and everything.
Apparently, if you have the fastest router in the world, or switch, or whatever it is, or access point, you can download this thing in 30 minutes. Everybody else, you can figure out how long it's going to take you from there. It's only 210 gigabytes, and you can wipe out every piece of software on your console or system to run this thing.
But other than that, it is just awesome, and it's great that we teach everybody to be a criminal. 220 gigabytes. All right.
But I digress. Mike, I got a correction from the peanut gallery. It's actually 232 gigs.
Oh, great. All right. Awesome.
That's wonderful. So for those of you who had a 225-gig hard drive, tough. All right.
But this is the game that teaches everybody to be a criminal, so let's just segue right over to Anthropic and the IPO, and what's going on with that. Because, well, as far as I can tell, John, investors are high on Anthropic. So tell us what's going on here.
High is the operative word, Mike. So I'm just looking right now, just literally to coincide with our broadcast. It just started trading at about 105 a share.
It was priced at $83 per share last night. Before that, we were talking about a range of $70 to $78. 5 trillion.
It's going to challenge SpaceX for the biggest IPO. So it's trading under the ANT symbol. We are looking at a number of things, including the amount of money it's going to raise, which is about $70 billion at least.
And on top of that, we got another $26 billion from the strategic investors. That include AMD, Broadcom, Google, Nvidia, Microsoft, et cetera. I was in the green room earlier, and I want to bring Drew into this because he had some interesting insights about what this means in a bigger picture for AI.
And I thought it was really interesting when you talked about this direct impact versus going indirectly into the market. So I'm going to hand that off to you, but I'll monitor. It's going to be interesting to see where this goes, but it's over-performing in the early stages at least.
Yeah. I think it would be shocking if it didn't over-perform. It's kind of set up that way, I think.
And this isn't an ordinary IPO. This is a huge IPO. It involves pretty much every broker on the street.
There's a lot of activity that happens in the lead-up to something like this, and so many systems have to change because there's a new stock ticker. You'd think that would be something that was easy. But aside from that, what we're seeing is this massive sort of movement of investors' money.
Before today, your only way to get AI exposure is you've got to buy a Google or one of these big tech companies that has AI exposure. Now you can get it directly. And so I think a lot of what's been happening in sort of the more recent sell-offs is portfolio managers need to make room in their portfolio to bring the stock in.
And the question is: do we need more AI exposure? Right? So is this going to take a wind out of the sail for OpenAI when it goes network?
Is there So I've got some thoughts on this one. First of all, let's all give a shout-out to the seven co-founders of Anthropic, who each stand to make, I think I read about $18 billion. Let me say that again.
$18 billion each, and some couple of hundred million on the side for that, for starting this thing, including a journalist. So Mike, John, there's hope for us, right? We could still go start an AI company and become deca-billionaires, number one.
Number two, this is more than just a pure play AI stock coming to market. When we look at the models, the frontier models, the pride and joy of Donald Trump and the administration, there's really only two. There's OpenAI and Anthropic.
Yes, there's Grok. Yes, there's Meta. Yes, there's a few.
Google, Gemini. But when we think of sort of, I don't even want to use the term, it's beyond blue-chip AI stocks, it's Anthropic and OpenAI. And so I really think that this, when you think about what was the IPO that set the tone for the dot-com, and it may wind up the same way.
But what was the IPO that set the tone for the dot-com we spoke about yesterday? Netscape. Some would say Netscape.
Netscape. Others may say Google. But Netscape.
This is going to set the tone for this AI boom. Hey, Alan, you mentioned the founders. I'm going to mention two numbers that are kind of staggering based on some of the research.
So 600 non-founder employees are going to hold at least $100 million in equity after the IPO. And then 2,800 current and former Anthropic employees are probably going to hold at least a million of pre-tax vested equity. And so this is going to flood Silicon Valley- Exactly ...
which is where a lot of these people are. Exactly. John, I'm surprised you're not out there trying to grab your share.
Full disclosure, I turned down a job with Nvidia 10 years ago. Yeah. Really?
John's going to have to move. He can't afford to live there anymore because- Well, I don't want to- That's a very interesting point, Mike. That happened with Google, but we can talk about it on another show.
But it happened with Twitter, too. Let's look at it from another point of view, though. I think this sets the stage for OpenAI.
If you could do, I could do better. That's kind of the relationship between Dario and Sam. Whatever you could do, I could do better.
It's going to set the stage for OpenAI to have build-up on. $70 billion raised with another 26 in private placements. Wow, that should cover their losses for four or five months now.
That's the reality here, guys. You know, the other thing I was going to mention to you, and I should have mentioned it earlier, but these two numbers always strike me, and I know I'm being presumptuous about it. But okay, $17 billion in third quarter revenue.
That's great, but we talked about this yesterday, I believe. $900 million in disclosed purchase and lease commitments. That always- Kind of comes back to me.
And then if you're patient enough, the Futurum Group projects annual revenue from Anthropic is going to grow to between 450 billion and 750 billion by 2030. But that's a long- Will they be profitable? Is that how that works?
Well, this is actually really interesting because this came out, I think in their deal sheet a couple weeks ago, but of that $900 billion cost, 80% of that- Yes ... is for Anthropic's own usage of those environments. Internally.
So training their new models, maybe also using those models internally, I don't know, not clear. But of the other portion of that, they are making money on inference. There's actually a really fat margin there.
No, no, they do. That part of their business works, but it's the investment to get there. But I think the other thing that's missing here is we need to look at this in the broader sense of the market and the AI market, right?
You've got an investor community that is frankly up to here. The Jensen story, will he stay or will he go, or will he come back? He's already gone.
Will he come back or not? What does that mean for Nvidia? The recent election and what it means for data centers.
Trump doesn't want to upset Z, so we're not going to bring up the renting of our best Nvidia chips from Southeast Asian data centers. There's so much going on. The Neo clouds and how far pushed down their stock is.
The broader market is sour on AI right now, right? I know certain AI investors, some of whom all of us, most of us here know who are crying in their Huelia base right now. Their portfolios are getting hammered.
Now, is Anthropic the rising tide that's going to lift all of these boats? Or is it more like the Boca Raton real estate market where the houses that go for $30, $40 million get sold immediately? It's the regular houses that are on the market for nine and 10 months before someone will make an offer.
Right. So John, are the folks over at OpenAI suffering a little bit of Anthropic envy today? How long will it be before they get around to their IPO?
Yeah, so they were supposed to do theirs this year and they pushed it back. And I always wonder, the timing, Anthropic was doing well in the enterprise. OpenAI really not as much.
But I think there's probably, if I were them, or I'm not going to put myself in that position, but I think they're going to try to boot this up as fast as possible because as Alan mentioned, all these different factors, geopolitical, legislative perhaps in the US, the sentiment against these companies, they better push it faster, maybe as early as they possibly can next year. So are investors, Alan, I'll ask you this, are investors driving into this stock because they're true believers or they just don't seem to have anywhere else to go because, well, the rest of the market is shaky, or what's driving this? So this is a unique opportunity.
Shelly Levine here. Shelly "The Selling Machine" Levine from Glenlarry Glenross. This is a unique opportunity to own some oceanfront Florida real estate on the swamp.
This is a unique opportunity to own one of the two, as I said before, blue chips in the AI space. I think it's independent of the broader market malaise that we're seeing. You like that word, John, malaise?
That's a nice word. Malaise is a very good word. It sounds like a word from an author, but- Yeah.
We should stop playing that game. You guys have been throwing around $2 words all week. Yeah.
But, I think it exists separately from the broader malaise in the market. Yeah, this is an outlier. This is a unique opportunity.
I hate to sound like a shill for Glenlarry Glenross. For Shelly Levine from, yeah. Like Glenlarry Glenross.
No, Matt had it right. But yeah, this is one of those generational things. But, Drew pointed out, again, we talked earlier, it's going to be interesting where this stock is in a couple of days or a week.
We'll see how- Well, that's what I wrote in my story too. I say six months because at six months the handcuffs come off. Because a lot of these people, you're saying are billionaires, they can't sell today, right?
They're locked. And in six months when those handcuffs come off and it starts trading a little bit more freely, and we'll see where the market is and has open AI come out by then. All right.
So we can tell all those pump and dump shops in Florida to stand down, they got to wait a few months? They never stand down. That's all they do.
All right. Glenlarry Glenross leads. Let's shift the gear here and talk about some real data science and IT and AI and stuff that we love and might arguably matter more in my mind.
Anyway, Supercomputing Conference was this week in Chicago and our good friend Stephen Foskett was there, and Jamie Hampton has a story about what Nvidia was up to at the show. And I think I covered at least one announcement from Hammer Space talking about what they're doing with storage. But Stephen, you were there.
Give us the vibe for what's going on, because I sometimes feel like all this AI stuff is derivative of what we used to call high performance computing and supercomputing, and it's just like we're monetizing and commercializing stuff that scientists did many, many moons ago. Well, yes, I was there and I really appreciate the chance to go to Supercomputing as media, which gave me an opportunity to go behind the scenes a bit and get access to some of the more academic aspects of this industry. It was really, really exciting and interesting.
As you say, one of the academics at the press conference at the beginning of Supercomputing literally said the same thing you said, that AI is HPC, but scientific computing is also HPC, and AI is not scientific computing. And so we have a strange situation. If I can, I guess, focus more on the academic aspects of the conference.
We have a strange situation now where the top 500 list has never seen so few submissions. The last few have had less than 10 submissions from the graphs that they showed at the event. And yet everyone knows that there are bigger, badder, more powerful supercomputers out there that just aren't being tested because they're either in China.
China's only ever submitted one computer, and it's the number one since 2019. And it uses a homegrown processor, and it was submitted for basically national pride reasons. But they obviously have more powerful computers that they just haven't tested and haven't submitted.
There's also commercial supercomputers that don't get submitted because they're too busy making money. And then there's all the AI clusters out there that are obviously way bigger and way more powerful than anything we've seen. They aren't being tested because you think Anthropic or CoreWeave or Microsoft is going to shut down their entire AI business to run some benchmarks on the system?
No way, no how. So it's a really funny situation for that reason, but also another funny situation is that AI, well, technically speaking, AI relies on quantization down to from FP64, which is what the top 500 list tests down to FP8, FP6, FP4, even FP2 math. And the hardware is increasingly being optimized.
When I say increasingly, basically almost everything, with an important asterisk, is being optimized for these lower precision floating point operations to support AI workloads rather than the traditional scientific computing workloads. So we really have a divergence now where AI is HPC, but maybe AI was HPC and isn't anymore because it is increasingly a different kind of hardware. So I don't want to spend too much time on that, but I could really go in deep on that one.
It was really interesting conversations. ai. What's your take on what Steven's saying?
What's going on here in that HPC space, and are they living in some isolated reality but maybe we're the ones who are delusional in the AI world? I don't know. Well, with the introduction of AI agents into the workflow, and with some of these heavier scientific processes, I think that it's interesting.
One of the announcements that Nvidia made was they announced the Rubin HPC GPU, so a specialized GPU that has, I think, 50,000 CUDA cores, like three times what Blackwell has. So I think they're returning their attention to FP64, to that double precision computing, which is fundamental. And over the past few generations of GPUs, like Steven was saying, they've really focused on bringing that math down, lower precision.
And so I think a lot of HPC customers, they wanted that FP64 performance. The kind of like what AMD is offering, with their newer things. So this is interesting.
This is them, I think, getting back into the HPC game a bit. Maybe. So Steven, are we hinting at...
I don't know, I'm just going to ask it because I'm not sure, but are we saying that our obsession with AI is harming or hurting our research efforts in other scientific environments, and these things come at a cost there? What exactly is the relationship here? Well, I wouldn't say that it's harming because I would say that we're basically building a parallel universe of HPC that's focused on quantized processors and a different kind of work.
Again, another scientific paper that I was really interested in was basically the importance of data movement and memory movement as opposed to processor speed. You would think that it would be all about the CPUs or the GPUs or the NPUs or the special purpose accelerators that are doing all this math. But there was a really interesting paper that I was looking at, I don't recall the author right now, at the show that was basically pointing out that it really just comes down to memory bandwidth, which is really kind of funny.
Basically, the number one supercomputer on the list is all CPU with no accelerators. And so there's a real good, big question there of, have we been doing all the wrong things? But that being said, you look at these AI clusters, which are essentially HPC supercomputers, and they're all optimized for this AI workload.
I really appreciate Jamie pointing out the Nvidia announcement. That was really making a lot of buzz, along with a new company that I want to introduce to my friends here on Techstrong gang. So one of the big banners and one of the companies that made a big splash at the show was a company called Bolt Graphics.
Now, you and I probably made sourdough or worked on crafts during the pandemic, but the guy from Bolt Graphics and his wife decided to develop their own supercomputer GPU from scratch, from zero, with no background knowledge of chip design or electric engineering, electrical engineering, graphics, anything like that. And they did. And they're actually bringing it to market in a couple of years.
They've got an FPGA simulation. They are getting ready to tape it out. And everybody was buzzing over this thing because it seems really out of touch with where we are today.
But it is specifically optimized for HPC and FP64. Because I talked to him for quite a while, and he was saying that essentially the whole industry, including Nvidia and AMD and Intel, have really abandoned the HPC market because they're so focused on AI. So they decided to make a GPU focus that doesn't work for AI, which is just wild.
And the day they announced it, Nvidia announced a special version focused on the HPC market, which is not wild. They would love to take credit for that. I don't think they care about some guy and his wife making a GPU.
I think what they care about is they want to make sure they have an entry in every market, and they don't want to be bumped off these lists because we did see the first Blackwell- Mm-hmm ... supercomputer on the list. And so, yes, we're seeing a divergence, but I think that there's still enough attention on the scientific computing side that it still matters.
Maybe Intel and AMD will care. Drew, you've been chuckling all along here. What's your take on what's going on?
Yeah, this is an area that I'm not super familiar with in terms of academia and research computers. So I think it's fascinating. I was really struck by what Jamie said about sort of having this be a second or an additional ladder for Nvidia, right?
AI is one part. This is definitely something else they can do. Maybe at this time of turmoil, they want to be announcing all the good things that they're up to.
Mm-hmm. But what I'm curious about, maybe Steven, I don't know if you heard any of this at the conference, but we're talking about AI versus HPC as basically two different models for compute. I think there is a scientific interest in the AI approach.
Not necessarily LLMs, of course, but probabilistic, large language, large models. Are there certain kinds of problems that work for that AI approach that don't work for, say, the traditional HPC approach? Yeah.
I think that the biggest thing that people are focused on is this quantization issue. Most traditional floating point units are FP64 or bigger because you need precision in scientific computing. Most AI models, in fact, I think all AI models that we've got now, especially the popular ones, don't need any kind of precision and be quantized all the way down to literally FP2, and they run great and they're much, much faster, order of magnitude faster than they would if they were running with that kind of precision, probably multiple orders of magnitude.
And so that's a big one. But the other thing is, what I was talking about before with the memory bandwidth, with the overall construction of the system is completely different. And we're also seeing a big change in the construction of AI systems.
So just like AI systems diverge from traditional academic and scientific supercomputers, we're seeing AI systems diverge as well as companies are developing systems designed for inferencing, for running models as opposed to training models. And those systems are wildly, wildly different. So instead of having a rack scale or data center scale supercomputer, they're actually rolling out inferencing systems that have multiple GPUs, a lot of memory, a lot of flash storage and so on, in pizza boxes.
2U, 4U boxes. Because with completely different interconnects, completely different architecture, different balance between CPU and GPU, much more CPU heavy, which is wild. Again, that's not been a part of the AI industry until now.
We even saw some companies, another company, Tenstorrent, was there with a RISC-V based, it's RISC-V plus accelerator, on cards that can do inferencing. And that would've been just laughably, ludicrously slow in the training world. " Alan, what's your take on this?
Because I can't help but wonder if we're so overly invested in AI that we're starving the rest of these scientific applications for resources, and these are the things that might, I don't know, actually make a difference in curing cancer and all these other things we want to do. I don't think it's mutually exclusive. Let me first of all say, I'm bummed.
I actually, like Steven, had a media pass to Supercomputing, but after getting sick in KubeCon, I passed it over to another one of our colleagues, Lee Sellers, who was there, and I've spoken to Lee a bit. It was a great show. Look, Supercomputing was not the sexy show it was just four or five years ago Right?
For some people, it was always the event of the year, but it wasn't this big, glamorous event that Steven went to this year. And I think a good chunk of that is because of AI. " It's put the spotlight on this.
But here's the good news. Just as China hasn't submitted a thing, a new machine since 2019, there are a lot of machines that I think are running over in places like Fort Meade and Langley and those kinds of places that are pretty damn good, too. And I'm not sure if they're using floating point two or floating point 24 or what have you, but they're pretty powerful.
Well, I will mention to that note, I actually overheard some of those folks-- I'm not supposed to say this. This is just between us. Don't.
Well, let's not get us all in trouble, Steven. I did overhear a conversation relating to a supercomputer that was way faster than anything on the list that the US government has. Yeah.
And I'll throw something else out at you. Yeah. What makes you think there's not some quantum stuff going on at some of these places?
Well, I had some great conversations about quantum I'd love to talk about, too. " Everybody in the quantum industry, and I spent a lot of time, most of yesterday, focused on quantum because I was just interested in that question, like, what the heck's going on in this industry? Everyone believes that quantum actually is here, but nobody wants to talk about it because it's giving them a unique competitive advantage in their weird little field.
Mm-hmm. So, if you're doing chemistry or network routing or resource planning or something, quantum actually absolutely works, but nobody's talking about it. Yeah.
It's not like it cracked RSA. It's like it delivered- No, but I would not be surprised if in some labs or somewhere in the government, this stuff is being worked with. Right.
Now, Mike- And these people are not telling you about aliens in the next breath, right? They're serious scientists? Yeah.
These are serious business people. Are you kidding me, Steven? You think Steven was played like that?
No. I don't know, but I got to say, I never thought I'd live to see the day where the phrase "supercomputing" and "sexy" were in the same sentence, but here we are. Well, no, but it is.
Well, it's working like that. But I got to tell you, this has become a marquee event for the industry. And so, Mike, to answer your question, this town is big enough for AI and academia and quantum and all of the many flavors that we're seeing in high-performance computing.
I look forward to us spending more time on these issues. Steven, final word. So, I just want to bring up a couple of stupid things from the show because we're all friends here.
" I actually saw one of those, and the company had put the literature over the Jensen Huang signature so that nobody could see it. It was kind of- Steven, that's becoming a thing, right? Absolutely.
Crescen even mentioned him by name during their earnings call. It's pretty hilarious. And then the other thing is that the Department of Energy had a huge booth at Supercomputing, and it was so facepalmingly bad because everything said SI or super intelligence instead of AI.
Oh. And literally nothing else at the conference did except the stupid US government booth. " But it's- Just know the clock is running.
The clock is running, and the cavalry is on the way. But they absolutely did. " And then he continued, and we talked about AI, and he used a different word than I did because these people are absolutely not going to lose their jobs over talking to me at a conference.
Oh, that's disgusting. It's crazy that's happened to that. Crazy.
So let's shift a gear towards satellites and Amazon, and I think, is it Blue Origin? Yeah. Blue Origin is Jeff Bezos' company.
Yeah. Yeah. So Blue Origin was trying to take a bunch of Amazon satellites up into space, and well- No, it was not Blue Origin taking them up.
It was the Space Launch Consortium or whatever it is. All right. The United Launch Alliance.
United Launch Alliance. United Launch Alliance. ULA.
All right. Blue Origin engines. We stand corrected.
All right, Blue Origin engines. So everybody was involved in this little adventure. So Jamie, I lived in Florida for a few years, hanging out in Boca, and people go up to Cocoa Beach in the land of "I Dream of Jeannie," and they pull out their little chairs, and they watch these rockets take off, and every now and again, one of them goes boom.
So, from your perspective, is this kind of where you sit? Is this becoming a regular thing? Like, I don't know, does one out of every 25 of these things blow up, and is this just what we're living with?
And what is life down there with all these rocket launches? Well, I live on the opposite coast, but we can still see these rocket launches. " It's like nobody...
It should be business as usual by this point. Mm-hmm. But I think this launch was kind of a special failure, though.
It destroyed 40 of their Amazon Leo satellites. This is a major setback. Because it only has about 400 satellites in orbit right now compared to SpaceX.
I don't know the exact number they have, but Amazon has 3,200 of these planned. So, what is this failure going to do for their launch schedule? That's my question.
I mean, I agree. Oh, could I ask Jamie a quick question? I'm sorry, Alan.
I didn't mean to cut you off. No, go ahead. So I want to ask you about that because you mentioned the 3,200 satellites they have.
Don't they have- Well, 3,200 they're planning. But they've explained- They only have 400 in the air. But they have yet to deploy almost 90% of them ahead of this 2029 FCC deadline.
Is that- Right. That's the time crunch, John. They got to get these up before then.
They have 38 planned Vulcan launches. Mm-hmm. But they've done an investigation, and the preliminary results suggest that the rocket had insufficient first-stage booster thrust getting it up there, because there was inadequate- It failed ...
fuel supply, I guess, to the Blue Origin engines, so. Right. You talk about this United Launch Alliance, and it sounds like one of those government kind of acronyms where in the end, it just slows stuff- Well, it's the combination of Boeing and Lockheed Martin.
They've merged. Right. This is not the minor leagues.
These are major players. No, this is the New York Yankees and the Dodgers getting together to play baseball. Yeah.
So, yeah, but- But this happens in rockets. Right. It's part of the- But there's another team over in Texas.
Are they laughing about this right now? Why? They've been perfect?
No. Well, they've been pretty good at this. They're- They've been pretty good lately, but how many do they get scrubbed?
How many... They've had their share. SpaceX has the best launch record in the history of space, with the Falcon rocket.
With the Falcon, yes, but with the new one- And that's what this is. This is a competitor for that. And so- But SpaceX is phasing out Falcon.
Exactly. Right. Because they got one that's better, and it's vertical?
No, it's bigger, and he wants to go to Mars. Mm-hmm. They've got a plan to make a larger and 100% reusable rocket.
I'm sorry to inject logic and normalcy into this conversation- No, I just think- ... instead of just dissing on Elon here, but- Well, Elon needs to be dissed on. Yeah.
Maybe. But that being said, look, we need a competitor to Starlink. Absolutely, positively.
Our country's national security shouldn't depend on what hair curls up his a*s and whether he wants to shut someone's access down or not. Yep. That's wrong.
We need a competitor. We need options. This Leo system seems to be our best bet right now until the Chinese, the Indians, and every other country starts doing their own thing because they're not going to march to Elon's beat either.
Mm-hmm. And so it's important ... that this United Launch Alliance gets their act together.
It's important that we have a Starlink competitor. I- So stay right there ... I agree with that a wholehearted.
I think this is super important for the commercial space industry. Yep. What SpaceX has done is incredible.
They've created just a miracle. No one thought they could do it. And we're seeing a lot of companies trying to replicate some of that success.
But I think the model is different, and that might be part of the problem. Is SpaceX more of like, hey, they're going to create the whole thing, like Apple, just top to bottom, they own all the hardware. And ULA is sort of like the PC market, right?
Like, I'll use these engines, and then that fuel tank or whatever. It seems like a much harder way to do space, and I think space is already incredibly hard. And honestly, I don't think Boeing is up to the task.
Come on, they can't make planes. This isn't Boeing of-- You're right. Yeah, I think you're actually right about that, Drew.
That's a great metaphor. United Launch Alliance is IBM. They are the best.
They know they're the best. They've had success in the past, and now they're absolutely unable to compete with Apple of space that is trying to verticalize and integrate everything. The problem is- But that Apple is also trying to build robots, cars, right?
Because he wants to combine Tesla into it. He does Twitter. They do their AI.
Are they being defocused? No, they're actually incredibly focused. They're just not focused on this anymore.
People, I think, are misunderstanding why Tesla or why SpaceX is getting out of the rocket game and why they're building Starship. And they don't care about this market. They want to make boatloads of money by replacing all telecom on the planet with Starlink.
If you look at the Starship, yes, it's failed, but it hasn't really been in production. They're still ramping up, but the reason that they're building it is so that they can launch tons of giant satellites for themselves. Their thing isn't even compatible with launching regular satellites.
They're not even interested in the market anymore, and that's why they're retiring the Falcon, they're retiring the Heavy because they just don't care about anything except themselves. And so if you want to diss Elon, diss Elon for basically cornering- Being a megalomaniac ... the SpaceX market and then walking away for a selfish other reason, which is to corner all telecommunications on Earth.
Mm-hmm. Well, it's the ultimate shiny trinket. But aren't those satellites that he's walking away from a subset of that telecom network, so won't they all connect?
Yeah, but no, they're not heavy-duty supercomputing enough to do the kind of heavy lifting that they want to be able to do. The ones they're going to launch off Starship are powerful enough, and SpaceX has just licensed the spectrum to do direct satellite-to-phone so that they can compete with Verizon and AT&T and T-Mobile and Deutsche Telekom and every other- Cellular ... access provider on Earth.
They're not going after Spectrum or Comcast. They're going after every telecom provider on Earth, and the way to do it is to get piles of these giant satellites in low Earth orbit, and the way to do that is to make Starship work fully reusable. Oh, we've recreated Ma Bell.
Oh-no. But that road is not paved. It's a slippery road.
Yeah. There's no guarantee that gets done. And I mentioned China, and I mentioned India.
Both of those countries have huge projects underway. Yep. I did a special- In fact, theirs are running.
China has gotten- Yes. But it's not at scale yet because they're good at scaling, too. But I did a report on this on Techstrong a couple of months ago.
If you look at how many satellites Bezos wants to put up and how many SpaceX already has up and what they want to do, and how many India wants to put up and how many China wants to put up, that within 10 years, we're going to have something like 10 million satellites orbiting the globe. And each of these satellites has to have its own Star Trek shield to divert it if it gets too close to another satellite. It's way worse than that, Alan.
All of these satellites have to be in a very narrow orbit to stay in the sun. Absolutely. So, it has to happen eventually.
And now you're going to start having collisions and anti-satellite technology, and- Yep. The whole thing is a giant s**t show. Sorry, I had to say it.
Jamie. And they're abandoning- Yeah. Yeah.
I'm sorry. I was just going to say, there's a company that was founded by Steve Wozniak, that's out there. They're trying to come up with ways to go out and collect all of the junk that's out in space because there's so much garbage in space.
Well, and that's the other thing. It's going to be a problem for deploying more satellites. As the satellites age out, do we have recycling available?
Do I throw that at the curb? So won't somebody just put a satellite up there and set it on self-destruct and knock out about 300 of these things in one go? Is that the danger?
You know what? Is that warfare in the rest of this century? It actually has happened a few years back.
China- Yes ... blew up one of their satellites as a test. So yeah, that's real.
All right, John, another thing to keep you up at night, see? There you go. Well, I was looking at the-- Thanks a lot, Mike.
I'm looking at the Anthropic stock. 92. I've just been monitoring it.
How's the broader market, though? Yeah, I haven't looked. I'm looking at just specifically the stock, but I can look at that, too.
Can I weigh in on that first story? I didn't get to say a single word about it. Sure, yeah.
Here we go. Ready? Don't buy.
That's my financial advice. Congratulations. Thanks, Steven.
Yeah. Someone's going to prove you wrong, Steven. You shouldn't have said that.
All right. I'll take it one step further. Wait a minute, then buy low.
Sell high. Someone said sell high just now. Nasdaq is down 1%, by the way.
Oh, okay. So it's kind of treading water and- It's treading water, basically. That's a good way to put it, yes.
Thanks, John. Thank you. All right.
Hey, we got to wrap up, though, guys. What a great time to be doing a show like the "Techstrong Gang," where we can come on here and talk about these super interesting topics and weigh in and have our opinions out there. Steven, thank you for your yeoman's work today.
You did a lot of the heavy lifting, and I don't mean from Starship. But- Yeah, I'm kind of like a super heavy rocket, yeah. You're a supercomputer, Steve, but we don't put you in the top 500 yet.
You're our secret weapon. Jamie, thanks for the reporting and timely input as always. Thank you.
Drew, it's always a pleasure having your bright star here. And then John and Mike, what can I say? Another great day at the Gang.
Mm-hmm. I hope you've enjoyed this. You know we do this every Monday to Friday, noon live Eastern time on our Techstrong network.
tv, Techstrong TV YouTube channel, and our OTT app for Android and Apple, and Apple TV and Roku and Amazon Fire, and wherever you've got a screen, you can watch the Gang. And I'm glad we did this at noon because the kids will be home from school around 3:00, and network bandwidth will be at a premium as they start downloading There you go. I heard some of the bigger kids, like some of our own people here at Techstrong, were up at midnight because it was released at midnight to play these things.
But I digress. Okay. Look, we all have our toys, right?
For Steven, it's rockets and supercomputers. For others, it's video games. We'll be back tomorrow with the Gang.
Mike, you'll be driving the bus. All right. See you guys tomorrow.
Have a great day, everyone.
