Anthropic Gets A Warning, SpaceX Shares Fall on Fourth Day — Bloomberg Tech 6/17/2026

[00:00:00] Speaker 1: This is Bloomberg Tech. Coming up, Anthropic gets a warning from the U.S. Commerce Secretary saying it needs government permission to grant foreign nationals access to its most advanced AI models. Plus, after three days of soaring, SpaceX shares come back down to Earth and drop for the first time in the company's fourth trading day. And Bloomberg reports Apple's upcoming camera-equipped AirPods will launch in late 2027 as part of a flurry of new releases. There is a lot going on in the world of technology, and that's playing out in equity markets. The Nasdaq 100 has seen a lot of churn and chop. We're up modestly, a few tenths of a percent. Chip stocks are rebounding, but we had a really big drop in yesterday's session. But remember, as an index, as a bucket, chip stocks are up like 90% year-to-date. And SpaceX, that's a pretty reasonable pullback. Down to $193 per share, down 4%, but had added $930 billion of market cap in the first three days of its life as a public company. Let's get to the news stories and an extraordinary escalation in Washington's conflict with Amthropic. Bloomberg's obtained a copy of the letter from Commerce Secretary Howard Lutnick, dated last Friday, warning that Amthropic faces severe civil and criminal penalties if it grants foreign nationals access to its most advanced AI models. Bloomberg's editor in DC, Mike Shepard, is with us for more. So what do we learn from the letter? Let's start there. [00:01:45] Mike Shepard: Well, one of the things we actually didn't learn, Ed, was the rationale. What was the underlying security concern that drove Howard Lutnick and the U.S. government to take such a drastic step as to order and impose export controls on shipments and sharing of this sensitive technology? But what we did learn, Ed, was this, that the government is using authorities that are typically reserved for dual use technology. That is products that could be applied for both civilian and military purposes. And the government is invoking these authorities in a way to try to prevent foreign nationals perhaps from leaking either intentionally or unintentionally these wears, these products, the top two AI models from Amthropic to foreign adversaries of the U.S. And they are preparing to impose these export controls and have them in place until further notice. And they're warning, Ed, of criminal penalties and civil penalties if Amthropic fails to comply. [00:02:44] Speaker 1: Shep, just very quickly, let's recap the reporting of the show yesterday. Technical staff of Amthropic went to D.C. They met with the administration. What do we know? [00:02:55] Mike Shepard: Well, we haven't seen any further signs of progress in the negotiations between the two sides. We do, however, have today at the G7 summit in Evian, Dario Amadei, the Amthropic CEO, and President Donald Trump in the same room for an AI lunch hosted by French President Emmanuel Macron. Now, at the lunch, the two men did not sit anywhere near each other. And it's unclear whether they had any interactions or sidebars. But after his bilateral meeting with Indian Prime Minister Narendra Modi, President Donald Trump was asked, well, how are things going with Amthropic? And all he could say was going fine. We may hear more from the president on this, but clearly it's something that's on the mind of G7 leaders. And Macron himself is pushing this idea of trusted partners for sensitive technologies like Amthropic's AI models, in which these partners, allied countries, would be vetted by security authorities so that they too could perhaps regain access to sensitive tech like these two models. [00:03:55] Speaker 1: Bloomberg's Mike Shepard. Thank you very much. I want to go deeper on that story. The G7 is underway in Evian, France this week. And beyond geopolitical and economic debates, AI is in the spotlight. The chief executives of OpenAI, Amthropic, and Google all on hand, joining world leaders to discuss the global future of AI. The conversation comes as the U.S. government indicates its willingness to use emergency powers, forcing private developers into compliance over national security threats. The president of the United States also in Evian, and we expect to hear from him today. Bloomberg's Kayleigh Lyons, co-host of Balance of Power is with us. Such an interesting dynamic, right? Power players, world leaders, AI leaders. Recap. [00:04:39] Kayleigh Lyons: Yeah. Well, and Mike spoke to some of this incredibly well, Ed. Basically, what we have here is countries who are all calling for a regulatory regime around AI. The German Chancellor Friedrich Merz talked about the need for an intensive discussion around that, while also saying that all countries need to have access to this technology. And admitting, Friedrich Merz did, that Europe needs to play catch-up. And that's really what's at issue here, as we consider anthropic and these export controls being slapped on in particular, is that Europe doesn't have its own AI juggernauts such as those in the United States. This is somewhere, and we heard President Trump speak to this earlier as well, that the United States is leading in, and Europe wants to be able to trust that it can access the technology, thus the push to have Europe considered some of these trusted partners who would be able to access this technology versus adversaries like China, for example, when we're considering things like mythos and fable five. So this is obviously an act of discussion. We could hear more from President Trump about this, but it also feeds into another issue that was raised at the G7 summit, Ed. As we consider the technology questions that leaders are grappling with, it is also inputs into certain technologies like rare earths and permanent magnets. And we actually saw the G7 countries agreeing today to cap their imported supply from China to 60 percent of rare earths and permanent magnets by the year 2030. The ultimate goal, as soon as possible, is to get those imports down to 50 percent. And they're going to, by the end of this year, have a separate discussion around other critical minerals. But the idea is reducing the dependence on China, which has shown a willingness to restrict exports of this market in which they are dominant, not just in terms of mining these rare earths, but also in the refining of them. This is going to be a really ambitious target for these countries to meet, Ed, as they have to work on their own supply chains. Even officials of the G7 are admitting that may mean imposing quotas on certain industries like defense, for example. But this is actually one of the big tangible items that is emerging from the G7 summit. The 60 percent cap on rare earths from China again by 2030. [00:06:44] Speaker 1: The U.S. Kayleigh Lyons, thank you very much. And of course, we expect to hear from the president about the status of an agreement with Iran, another topic that has had direct impact on supply chains and tech, including semiconductors. Let's stick with AI. NVIDIA founder and CEO Jensen Wang sat down with the AP to discuss the need to build AI in a safe and responsible way. But beyond that, he says the U.S. needs to harness the new technology to stay ahead of the curve globally. Listen to this. We have to be very careful that [00:07:13] Speaker 4: artificial intelligence, this new next generation of technology revolution, that we're also enthusiastic about adopting it and make sure that we don't get left behind. Joining us now, Uday Chiroviu, a member of [00:07:24] Speaker 1: the portfolio management team for the International Equity Strategy desk at Harding-Lovner. You know, something that's come up actually very recently, yesterday's show, was what is the kind of black swan for this market where we are right now? And it might just be regulatory risk. Listening to what Jensen Wang had to say there, extrapolate on that a little bit. You know, we're still unresolved with China. Supply chains are impacted by the war in Iran. What's the top of mind for you right now? [00:07:54] Speaker 5: The top of mind for us is productivity gains, so the return on equity that can be generated from AI and how much people are embracing it and what purposes are they using and what is the adoption curve and the return on that adoption curve is what's on top of our mind. And we point out less regulatory risk might be less of an issue. The bigger issue is going to be people have to take time to digest what they're included, what AI they've started using in their workplaces and then understand the returns of it before they can generate additional demand. And to us, the timing of that is where the market might get most confused about. And to us, that's that black's fundament. Short-term risk might be extrapolated as long-term risk and that's where the market might [00:08:42] Speaker 1: overreact. Where do you look for the evidence of productivity's impact from AI? You know, very recently, and I think you're talking about in equity markets, right? We had a conversation 10 days ago with San Francisco Fed President, Mary Daly saying, "I don't see any evidence of productivity in this country being impacted by AI for better or worse." Yeah. So from that, I think there's a number of ways [00:09:06] Speaker 5: to look at it. The first thing we do is to talk to companies at the individual levels, how they're adopting AI and what impact that AI is having on either their cost management or the sales that they're generating or the activity that they're doing. So one clear evidence that we find is people are able to do a lot more with a lot less human resources or keeping human resources safe. What we're yet to see is that translate into profit growth for anyone either. And that to us is what the next step is going to be and trying to understand where that profit growth is coming from. Most immediate is that some costs can come down, but that cost is being reinvested into AI. And to us, that's where that [00:09:51] Speaker 1: discrepancy could come from. We started the program saying there's been some choppy trading and churn in equities in this morning's session at least. I remind the audience that we have a Fed decision and meeting today, the first of Kevin Walsher's chair. Is that a factor for technology investors [00:10:08] Speaker 5: right now? I don't think it's a factor for technology investors from a fundamental perspective. We still are focused on what's happening in the fundamental markets and the market is reacting to a lot of the fundamental news that's coming out. So the investors, in my mind, have become more and more sensitive to profit and what impact by AI is having on profit versus revenue. Over the last two years, revenues, backlogs, bookings was what the markets reacted to. It started to turn in the last two quarters where revenue growth is being given less of interest and more interest is given to profit growth. So I don't think we're there at the point where the interest rate changes itself will have a big impact on the market, but the longer run it will, because that sets the discount rate at which [00:10:54] Speaker 1: your NPV the future ends. Yes. It's been a long time since I've said this out loud on Bloomberg Tech, but higher rates discount the present value of future cash flows and continue to remind ourselves of that point. Look, I won't ask you about SpaceX specifically, but there's this idea that that presages a big IPO window. But more sophisticated than that, after a long era of buybacks and other shareholder-friendly policies, lots of new equity will hit the market maybe this year, how does that go for [00:11:26] Speaker 5: you? Sorry, go ahead. So that creates a technical risk in terms of where is it when anthropic and open AI come to the market, the equity that's raised for that and the funding that's raised for that, where is that going to come from? Which sectors of the market? Which parts of the market? And we've seen some of the reasons why the semiconductors have done so well is software has underperformed. So what other sectors of the markets are going to underperform? So there are going to be technical factors and they'll become even more important in the market over the next six months. [00:11:55] Speaker 1: Just very quickly, French President Macron saying in his own presser in Evian, it's good that the US realises the dangers of frontier AI models. We're showing live pictures of him speaking at the G7. This idea of government intervention and dictation on how technology can or can't be used, how is that [00:12:16] Speaker 5: playing out in the market mood right now? I think it's not as important to the market at this point in time because we're still trying to understand what are the capabilities of AI. And there's so much latent demand in even the previous generation of models that we're less concerned about. Is there going to be the next generation going to be restricted? Well, we still haven't used the full capacity of the previous generations. So let's do that and understand what the tam of that is before we're [00:12:43] Speaker 1: worried about the next generation. Like I said, there is a lot going on in the world of technology this morning. Harding Loebner, Portfolio Manager Udejiru, great to have you back on the show. Thank you very much. Coming up, we're going to hear from HPE CEO Antonio Neri, story about let's make networking great again in the world of AI. That's next. This is Bloomberg Tech. As AI infrastructure scales, attention is shifting beyond the chips to networking. HPE is betting that demand and its Juniper acquisition will be a major growth driver. I sat down with HPE President and CEO, Antonio Neri. Listen to this. [00:13:31] Speaker 6: We believe and has been for now three years plus that networking is the core foundation to deliver AI across many aspects of the market, not just enterprise, but obviously to train these models because networking is also the bottleneck to get the best productivities for all the GPUs we are deploying all over the world. And so Juniper has been a key thesis of that. We understood in 2023, when we started thinking about the acquisition of Juniper with Rami, which was the CEO of Juniper and now leading our combined networking business, that this is a massive opportunity. And the combination of Juniper and HPE, in particular with compute storage and our GreenLake platform, will allow us to differentiate the market. So clearly it's going to be the driving force as we go forward because enterprises need to put that foundation in place and then the cloud experience around it because AI is the [00:14:31] Speaker 1: definition of a hybrid world load. Antonio, that's pretend for a moment that the Bloomberg tech audience is not familiar with networking. They know all about the GPUs. They know all about the high bandwidth memory. But what is it that we're talking about materially here? [00:14:48] Speaker 6: Yeah, when we think about AI and we think about the amount of accelerating computing through the GPUs we're deploying today, you have to connect all of them together in a parallel process. And that connectivity is the networking fabric. And there are three aspects of that. One is in a rack, think about the cabinet. You bring together 72 or 144. That requires a lot of networking connectivity. That's what we call scale up. That's where Juniper is going to play a bigger role with the introduction of AMD Helios in the fall, where Juniper will be the de facto networking fabric inside that architecture. And then when you start putting cabinets together in large data centers, you need to connect these cabinets all together. And that's where we scale out. And that's also Juniper. And then finally, as we build these amazing data centers all over the world, you have to connect them with each other. And that's what we call scale across. And that's where our Juniper routing business excelled, because we have one of the best platforms in the world. So that's why networking is so relevant to train models and to [00:15:54] Speaker 1: influence models. That was HPE CEO Antonio Neri. There is a lot going on in the world of technology. So many news stories. Let's get out to New York and Bloomberg's. Yahaira Anand. Yahaira, hi. [00:16:04] Speaker 7: Hi, Ed. It's time now for Talking Tech. First up, AI creditors are taking over software company Medallia. It's an effort led by Blackstone after owner Tom Abravo said it would not inject fresh cash into the struggling SaaS company. Creditors will invest $150 million into the company to curb its debt and accelerate AI efforts. Plus, SoftBank says it's becoming harder to find Latin American startups that are ready for major investments. The company says fewer firms in the region are meeting the requirements for its preferred investments of $50 million or more. It's a big shift from just a few years ago when Latin American startups were attracting record amounts of venture capital. And we end on Lenovo. It is the latest company tapping the debt market looking to raise $2 billion through a convertible bond sale. The Beijing-based maker of PCs and AI servers is returning to the market just two years after securing a similar size investment from Saudi Arabia's sovereign wealth fund. It says the money will be used to refinance debt and buy back shares. Ed. [00:17:10] Speaker 1: Thank you, Yahaira. As we discussed earlier, the U.S. has told Anthropic not to give foreign nationals access to its most powerful models. But CEO Dario Amadei says the best defense against those risks is getting the technology to cyber defenders. He spoke with Bloomberg's Emily Chang for the circuit. [00:17:27] Speaker 8: If this helps defenders, it also helps attackers. Can we defend anything anymore? [00:17:34] Speaker 9: What I would say is that the reason that we're giving Mythos to defenders before we give it to attackers is to patch all the bugs. I don't know. As the models get better, there may be more and more bugs to be found, but there's only so many. They're finite, right? It's like you have this surface and there's only so many holes in it. You patch all the holes and the surface becomes very hard to attack as well as the code itself is written with the powerful models. So it's then becomes very hard to find flaws in or break into. So I think on the other side of this, hopefully six months or a year from now, we have a much more secure internet ecosystem than we had in the past. We're trying to get to that world and we're doing the best we can to open up Mythos to new cyber defenders. We've been talking to the government. We're very respectful of their recommendations. They're slowing the pace at which we open it up because they're worried about counterintelligence risk. I think that's sensible. I think all serious people here understand that there's real trade-offs here. We see a lot of sniping from people on Twitter and from other AI companies. You look at what they're saying and the inconsistency with what they're doing. They're not serious people. They're not seriously engaging with the serious trade-offs that we have here. Look, I have customers calling me up every day saying, I want access to Mythos. I have countries calling me up saying, I want access to Mythos. And I have the U.S. government and my security team saying, no, wait a minute, there's risk to it. You know, I'm not saying one side or the other is right. I think it's somewhere in between. Both sides have valid points. But there's a real challenge here. And we need to face it together as a society. [00:19:16] Speaker 1: That was Anthropic CEO, Dario Amadei, speaking with Bloomberg's Emily Chang. You can catch that full episode of The Circuit later today on Bloomberg TV and on Bloomberg.com. OK, coming up, SpaceX falls for the first time since its IPO run. OK, it's day four and we're only down five percent. We have more on that next. This is Bloomberg Tech. SpaceX shares come back down to Earth dropping for the first time. It's only the company's fourth trading day since that blockbuster IPO. It puts the stock on track to snap three days of gains where the shares have pushed nearly 50 percent above the IPO price and added almost a trillion dollars of market cap. Bloomberg Equities reporter Carmen Reinecke is trying to keep track of numbers that in real time keep changing. I mean, what is the what is happening right now? Right. There is scarcity. There is only four percent of this company available for trade on the flow. What's the big thing here? [00:20:26] Speaker 10: Yeah. So I think we're seeing some of the key note, you know, volatility that can be associated with new IPOs in their earliest days of trading right here today. I mean, shares are down six percent right now. We're a little bit off the lows of the day for SpaceX. But this is a pretty extreme drop. It was up about six percent earlier in the session is paired all of that to decline here. We know that there's low flow that can make swings in the stock price look extra large or bigger than they usually would. And that can also then have an impact on market cap. So yesterday we reported that SpaceX had jumped over Amazon in market value. And that's a little bit up in the air today. We'll see where it ends [00:21:06] Speaker 1: at today's close. There's also like the the listed listed options right. The options market came into play over the last 24 hours. Is that a factor in the volatility? It definitely can be. I mean, you know, [00:21:19] Speaker 10: traders are obviously now making bets on where SpaceX shares go next. You know, they're buying puts and calls. They're trying to additionally sort of game this system or see if they can. There's extra, you know, alpha here. One thing that we have heard or seen that is that options contracts are pretty expensive. You know, investor Michael Burry wrote in a sub stack note yesterday that he was looking at puts and they were just too expensive to buy right now. So I think we have a lot of traders looking there. We have more options trading actually coming online in the next few weeks from others. That's a that's a volatility [00:21:55] Speaker 1: premium. Just really, really quick. And I know this is unfair. Do we even look at price to sales ratio for [00:22:00] Speaker 10: SpaceX? What's the point? That's a great question. I mean, I think the point is just in comparing it to some of these other large companies where it now has a similar market value. Right. I mean, it brought in about 19 billion dollars in revenue last year. And if you compare that to Microsoft, that is, you know, right. It's 15 times higher. So it's good to keep that in mind sort of as we look [00:22:22] Speaker 1: at the value of this company. We hope to talk a lot more about the company a little later in the program. Bloomberg. Thank you very much indeed. Now, coming up, the next generation of Apple devices is taking shape. And that includes AI AirPods and a foldable phone. Bloomberg's Mark Gurman, the latest reporting and the power on. It's halftime. See you soon. This is Bloomberg Tech. Welcome back to Bloomberg Tech. I'm taking a look at chips and the semiconductor space. It's kind of interesting. So yesterday there was a big drop on the stocks, about six percent, only its biggest drop since the first week of June. And we're rebounding a little bit today. But what was interesting is there wasn't really a reason. There wasn't really a big catalyst other than for a long time we've been saying that the chip space has been right for a little bit of a correction. And it is an index, the Philadelphia Semiconductor Index, that is up around 90 percent so far year to date. In other words, like mega gains in just six months. So we're going to continue tracking it. Remember, there is a Fed meeting a decision today. The first for chair, Kevin Walsh. Maybe that's a factor in it. Our earlier guest said no. There's a lot of news from Bloomberg as well. Apple's investors are frustrated. The iPhone maker has been slow to deliver on AI promises, even with its latest Siri AI features. And after a disappointing performance at its annual worldwide developers conference last week, investors are becoming a little bit impatient. Bloomberg's Michael Reagan leads the team that's been writing about this. It's interesting, right? Actually, I was staring at the chart and being like, what are the shares telling me in the days since I was in Cupertino? But that's still the same story. Everyone wants to see a little bit more from Apple. [00:24:13] Speaker 11: Yeah, I think that's pretty much it, Ed. You know, as you're familiar with, it's not uncommon for there to be a lot of optimism ahead of the developers conference from Apple and then a little bit of a letdown when there's not really a big bang product announcement. So, yeah, the investors we've talked to are basically saying we're just hoping for a little bit more oomph from their AI offerings, their Apple intelligence offerings. There's a little bit of disappointment that Siri will only be in beta mode in September, only available in English, not available in Europe or Asia. So a bit of a disappointment there. But I would say, you know, Apple stock is still it's up 10% this year. It's right in line with the S&P 500. But I think it goes back to those chip makers you're just talking to. People are, you know, turning their head over there and seeing these triple digit gains. You know, there's something like a dozen chip makers in the Nasdaq 100 with with triple digit gains this year. And just saying, you know what, we wish Apple had been a little bit more ahead of the curve, a little bit more aggressive with AI. Now, that said, we do have these days in the market where everybody starts second guessing all the capex involved in AI. And that's when Apple sort of becomes almost your defensive tech stock play. So there is a little bit of disappointment. But, you know, at the end of the day, this is Apple with this fortress balance sheet, still looking at 15% revenue growth for this fiscal year. So, you know, I don't think we can make too much out of the disappointment [00:25:44] Speaker 1: from the from the developers conference. Just very quick, Mike, there was there was a part of our report on this that basically seems to indicate the market's more focused on the next hardware upgrade cycle. Just real quick. What are they talking about? Yeah, exactly. I mean, when you're looking at the [00:25:59] Speaker 11: iPhone that's coming out in September, you know, what is sort of baked into the assumptions about, you know, how much growth we can we can see because of people refreshing buying new iPhones, just because they're excited about the new one. Right. So far, you know, the revenue estimates are pretty much staying where they are. So it doesn't seem like they've really adding on to the estimates [00:26:22] Speaker 1: following the conference. Bloomberg's Michael Reagan with the Apple stock story. Thank you very much. Meanwhile, the company holds steady with its AI ambitions, planning camera equipped AirPods, smart glasses and new foldable iPhones as part of this like major product push, which will span the next two years. Bloomberg's Apple and consumer tech editor Mark Gurman is with us with the latest reporting. I just I read it a few times AirPods with camera and the new bit, I think, is the kind of timeline for launch. Go with that. Like, what was the important bit of what you reported, Mark? [00:26:56] Speaker 12: Yeah. So these new camera AirPods are going to launch at the end of 2027. It's going to be a pretty blockbuster fall of next year. That's the 20th anniversary of the iPhone. So they're planning a lot of stuff to go hand in hand with that. In addition to these AirPods with cameras, and just to give you some context, these are not for facial recognition. These are not for taking pictures in video. These are for being able to see the world around you and feed that information into Siri. So you can say things sort of like, what can I make with the food on this table? Right? What recipes can you concoct for me? Tell me more information about that object or that plant or what have you, or things for visual reminders or more specific turn-by-turn directions. In addition to that, you're going to see the second generation of the foldable iPhone, as well as the 20th anniversary iPhone, something I'm calling the iPhone 20 or the iPhone 20 Pro coming out at the end of next year. So it's going to be pretty jam-packed. And if you look at the more near term, this year is going to be pretty extensive as well. You have several new iPads, you have several new Macs, new Apple Watches, and of course you have the first foldable iPhone, as well as the iPhone 18 Pro, 18 Pro Max. And if you recall, Apple made that big design change with the 17 Pro and Pro Max last year. Typically, when Apple does a new design change, the first two years of that design are extremely successful in terms of upgraders. Some people don't do it in year one. They like to wait for year two. And so I think the 18 Pro, Pro Max are going to be [00:28:26] Speaker 1: pretty hot sellers this fall as well. Another big story that played out in markets, it played out on social media in a huge way in the last 24 hours is Snap and Snap Specs, augmented reality glasses. You know what the stock reaction was. You spoke to the CEO yourself. Summarize everything we [00:28:44] Speaker 12: need to know about this, please. I think the stock reaction is more in response to how much was spent on this product and investors and Wall Street knowing that this is not going to be a hit out the gate. This is a conceptual product right now. This is something like when the Vision Pro launched a couple of years ago. This is going to be for developers, for the biggest tech fans, for the earliest of early adopters. This is not something that's going to light the world on fire from the get-go. But this is the future of computing, I believe, as well. And this is what Evan Spiegel has been saying. These are augmented reality glasses that allow you to overlay applications, your work, notifications on top of the real world. And eventually, this is a product category that is going to click. But it needs to get a little bit lighter and it needs to get a lot less expensive. These are about $2,200 right now. Certainly not the priciest device in the category. The Vision Pro obviously being over $1,000 more expensive than that. But there is some work to be done. But this is another stepping stone towards that vision that Snap and several other companies have. So maybe I'm not optimistic about this first generation of the product from Snap. But in terms of the category, this is going to be the future. And Snap clearly is placing a foothold there. [00:30:00] Speaker 1: Thanks, Mark Gurman, who leads all of our coverage of consumer technology. Thank you very much. Did you see this one? Allbirds is now Smartbird, completing its pivot from footwear to AI infrastructure after selling off its shoe business and the brand's assets. The company tapped former DCAI CEO Nadia Carlston to lead the next chapter. Smartbird also doubled its convertible financing facility to $100 million and says it's already talking with potential customers as it builds out its first AI computing clusters. Wow. Okay, coming up, we're going to talk about the challenges facing space-based data centers, including satellite-to-satellite communication. The deep dive next. This is Bloomberg Tech. The US is the clear winner in the AI race. That's according to ASML CEO Christophe Fouquet. He sat down with Bloomberg Tech Europe's Tom McKenzie earlier today on the sidelines of the VivaTech conference in Paris. Here's what he had to say about his thoughts on the global AI race. [00:31:15] Speaker 13: If we compare how we do on the entire ecosystem versus the US and China, I think today the US is a clear winner. I mean, they are looking at champions across the entire AI semi-ecosystem. I think the one place they were missing a bit out was manufacturing, and I think they have been extremely aggressive in bringing some key company to manufacturing in the US. They can do that because [00:31:40] Speaker 1: they buy chips. AsML CEO Christophe Fouquet. Back to SpaceX. Despite today's decline, SpaceX has soared since its IPO four days ago. A big part of investor enthusiasm stems from Elon Musk's goal to put data centers in space starting in 2028. We've heard from those bullish on the prospect, including SpaceX investor Sean Maguire, who said this about orbital compute. [00:32:09] Speaker 14: "Every individual component here is fully proven by SpaceX, except for the compute side. But that is not, it's just not that hard." [00:32:23] Speaker 1: But space does present specific challenges for compute, including communication between satellites and with the ground. That's an issue Dan Roelker, former head of software engineering at SpaceX, set out to address when he co-founded Observable Space and Dan joins us now. I'm really grateful for you to come in with your expertise, right? When you read the SpaceX prospectus, it's like, okay, the big picture, get it. Space, data centers in space as early as 2028. One of the things we focus on the show is the ability for communication, RF, but also laser. [00:32:59] Speaker 15: Try and summarize the challenge for us. Yeah, you know, like everyone's known for a long time that laser communication is the next generation of communication for space. You saw that really as Starlink was getting built with the ISL links between satellites that are in orbit. It's, you know, because it's in space, there's no like atmospheric conditions or it challenges really establishing optical links. One of the big challenges you kind of have though, going from space to ground and ground to space is obviously the atmosphere. But everyone kind of knows that solving those technical challenges or what's going to provide the really like 10 to 100 times the bandwidth throughput that you're going to want and continue like low latency that we need for kind of expanding the space. That's what we're [00:33:42] Speaker 1: doing here. Yeah, okay. No, no, no, no. I'm sorry to interrupt you. I guess that the easiest way to help the tech audience understand is like, what are the current limitations of existing generations of [00:33:52] Speaker 15: hardware and software in that environment in space? Yeah. So, you know, I think from being able to deploy these ground stations for laser communications at scale on earth to communicate is definitely one of the challenges. There's a lot of technical challenges as well around being able to correct like the kind of waveform and wavefront for laser communication and around like adaptive optics as you're basically receiving the light on our ground stations. These are all like very active challenges, but we've actually made as a as a as a industry a lot of like advancements to make this feasible at this point. The data centers in space now is really going to require laser comm to be the primary mechanism of communication moving forward. So the timing works out really well. There's still challenges ahead, but it is, I'd say like a commercially viable at this point. And so that's why we're really excited about being able to get these deployed at scale over the next six to 12 months. Observable Space, the world's largest [00:34:57] Speaker 1: vertically integrated hardware and software company for space observation, but you are working on the laser component. That's right. Why is laser the best technology? There's a few different things that make it [00:35:12] Speaker 15: clearly the winner over RF. First, just pure speed. It's 10 to 100 times faster. The second is there's no spectrum management. You know, a lot of times with the RF bands, you have to work with different countries to be able to manage who owns the spectrum and how that's kind of being used and serviced from space. And then the other one that is really important, especially for kind of military defense purposes is the privacy of it. It's a very point to point communication. So it's extremely hard to eavesdrop on and be able to collect that communication. You have to be in a very specific area where the ground station is located at in order to collect any signals. So from a military perspective, this is like a clear winner and most future constellations for defense purposes are going [00:35:59] Speaker 1: to be optical first and moving forward. Dan, the basic idea of SpaceX going public is also that it will shine a light on commercial space. You know, SpaceX likes to do a lot of stuff itself, but do you see a kind of ripple effect or even an opportunity to work with SpaceX on those technologies? [00:36:15] Speaker 15: Yeah, totally. Like, you know, SpaceX, like, we love SpaceX. It's really the company that's defining the future of space. So, you know, partnerships with them, like, we work with them and we continue to, like, would love to expand our relationship with SpaceX moving forward. I do think you're going to see over this next phase of space technology development, you know, people starting to unstack, kind of pull out the different stacks in the space technology spectrum. It's kind of similar to, like, how this happened in the early days of the internet, when if you were going to be an internet company, you had to do all the things, you know, you were a data center, you were like a website, you were doing networking, you know, as the space economy and kind of space technology matures, you're going to see these different layers being pulled out. We're that communication, optical communication infrastructure layer that is now pulled out, and people can just buy our products off the shelf, turnkey solutions to establish optical cons from Earth to space and space to Earth. [00:37:20] Speaker 1: Lasers in space, no longer science fiction. Dan Roelker, CEO of Observable Space. Thank you so much. Now, coming up, we're going to keep the conversation on space-based data centers going. What other challenges does SpaceX's Orbital Compute Plan face? That's next. Also, take a look at some of these ETFs, ARK, SpaceX ETFs. There is volatility in SpaceX as a stock, but we've gone over how quickly some leveraged ETFs in particular, directly tied to SpaceX, got themselves set up. Actually, a lot of those are moving to the downside in a big way right now. We'll keep with that story. This is Bloomberg Tech. Blue Origin is rebuilding the launch site where its New Glenn rocket exploded last month, aiming to fly again later this year. New Glenn is a key part of Blue Origin's plans for space exploration, but also NASA's Artemis program. But it's years behind schedule. And as Blue Origin founder Jeff Bezos said earlier today, this is not a sector where it is easy to move fast. [00:38:28] Speaker 16: Traditional aerospace, I think, often suffers from slow decision-making speed. And I know how aerospace gets there. I mean, you know, you're building things that often have, you know, hazardous operations, life safety critical missions. And so, again, not every decision is like that, though. And so you end up making every decision as if it's life safety critical. [00:38:54] Speaker 1: Let's talk more about the challenges space presents the things we send up into it. Mackenzie Listrop joins us, a former director for the Goddard Space Flight Center, which is NASA's primary hub for building and operating scientific satellites, now principal consultant at Peridot Services and just an absolute expert in space systems. So there are different degrees of orbital compute, right? Different variations of a space-based data center. But at the extreme SpaceX is presenting, it is at hyperscale. It is running massive AI workloads, inference. Explain the distinction between those three grades of compute and why it's hard. [00:39:33] Speaker 8: Absolutely. So the first is really on orbit edge computing. So this is about taking, you know, imagery data, radar, weather, maritime, climate or defense data that's collected in space and then using, you know, AI, you know, edge computing, some kind of inference, some kind of processing in the space system itself before sending it down to Earth. So this is real. This is near term. So a satellite sends raw data to an orbital compute node. The node performs filtering or AI/ML inference and then only the higher value or smaller data products get sent to Earth. So this is a process that the company Axiom says that it deployed as a prototype unit on the International Space Station and has an orbital data center node as well. Now, the second is really resilient or sovereign storage and computing. So infrastructure that is physically separated from terrestrial threats. So disasters, political borders or ground based attacks. That also has some real defense continuity of government and critical infrastructure use cases, but still not entirely explored. Now, the third is hyperscale AI work in orbit. And this is really the moonshot. It's potentially enormous, but it is not merely about the launch cost or even the orbital downlink nodes. It really requires some challenges to be met in power, in heat rejection, in radiation tolerant computing, high bandwidth optical networking, orbital reliability and other issues. So it's a real [00:41:10] Speaker 1: challenge. So let's zero in on some of those challenges, right? Your job was to get systems ready to be in the environment of space. The sci-fi nerd in me thinks about the vast coldness of space. But in thermal, [00:41:23] Speaker 8: it's a very different challenge, right? It is. So, you know, space is not cold sometimes in the way that people think. And I would say that in orbit, a data center is a heat engine, first and foremost, and a compute platform, second. So a terrestrial data center rejects heat through chillers, through air, water or liquid cooling, through evaporative systems as well. Now, in orbit, there is no convective air heat sink. So heat has to be moved through the spacecraft and then radiated away. So this means that radiator area, radiator emissivity, the attitude control of the spacecraft, contamination, degradation, these are all effects that have now become first order economic variables. [00:42:12] Speaker 1: And then there is the power challenge. There's the belief that the sun is always on, infinite solar panel, as long as there's nothing blocking the path. But again, it's not as simple [00:42:23] Speaker 8: about that as that when it comes to access to that power. Correct. So, you know, the pitch for orbital data centers is often that, hey, sunlight is stronger and more continuous in the right orbit. Google's project suncatcher argues that solar panels in the right orbit can be up to eight times more productive than on Earth. So that is true. But just because you have space energy access doesn't mean that you have space energy infrastructure. So you're going to have to deal with some real constraints. And that means the close formation orbital dynamics, the radiation effects, the thermal management, the ground communications, all of this depends on a reliable access to power. And that means power generation, but also power storage distribution. Again, that thermal issue of heat rejection and also redundancy. Now, this has to be flight qualified. It also has to be able to be reliably operated in space and maintained. And maintenance is one of these hidden costs right now that we're not talking a lot about. [00:43:24] Speaker 1: Mackenzie, we have literally 10 seconds. Do you think SpaceX can do it? [00:43:29] Speaker 8: Mackenzie Listrop: Well, they're a vertically integrated company that has demonstrated many technical challenges before, so I wouldn't count them out. [00:43:36] Speaker 1: Mackenzie Listrop, principal consultant at Peridot Services, former director to NASA's Goddard Space Flight Center and a top expert in space-based systems. And get used to that. We're going to be talking about it a lot. That does it for this edition of Bloomberg Tech. Mackenzie Listrop: Really interesting mix of market news, political news, but the technology sector has a lot going on right now. Recap on the podcast, Apple, Spotify, iHeart, and all of the Bloomberg platforms. From San Francisco, this is Bloomberg Tech.