Amazon to Buy Globalstar for $11.6 Billion

23:00Bombas Sport is pro-level socks from the pros of socks. For another pro, you. Go to bombas.com slash audio and use code audio for 20 % off your first purchase. That's bombas.com and use code audio. On June 10th, Bloomberg Invest is back in Hong Kong. We look at the role Hong Kong plays between China and the world as major powers compete and markets realign. As global investors rethink risk, we'll explore the forces driving Asian demand and the future of private capital. Catch exclusive interviews with top newsmakers, plus a live recording of Bloomberg's Odd Lots podcast. Visit BloombergLive.com forward slash InvestHongKong to learn more.

23:38Supporting sponsor Deutsche Bank. In the race to build out AI infrastructure, Oracle is moving to solve that power pinch by striking another big deal with Bloom Energy for its fuel cell power. Bloomberg's Brody Ford has more. The share prices have reacted, Brody. And what is it that Bloom Energy brings that others don't? They bring power, right? I mean, that's the big bottleneck. Right now, you can really get the chips, you can get the customers, certainly, but in a lot of cases, you just can't get the power to get these big data centers running. And so what Bloom Energy offers is fuel cells that's going to be a lot quicker than if you were going to go out and try to buy some gas turbines yourself.

24:20And, you know, Oracle has clearly said that they are willing to pay what is likely a pretty hefty premium here to get power online quickly. and again they have the giant contract with open ai they're on the hook for a number of gigawatts and within a year or two and so oracle is ready to give out the cash to bloom to make sure that

24:41Caroline Hyde:this plan goes accordingly that scale is interesting because we talk about the hyperscalers and then oracle from a capacity standpoint we cover in great detail the hyperscalers securing long-term energy supply. Talk a bit more about Oracle's broader strategy on making sure it can power all of that capacity. Effectively, Oracle promised something pretty large to OpenAI. They promised four and a half gigawatts in a pretty unprecedented timeline. And so, yes, they are not a traditional hyperscaler, but right now they are building at hyperscale. They have a similar deal with a company called VoltaGrid, which offers kind of gas turbines as a service.

25:24And they're paying them a lot of money, right? I mean, out of their Shackelford campus in Texas, that's one data center they're paying over a billion dollars per year to power it. And so we have to imagine this is a similar scale deal. And that's why the Bloom shares are so happy this morning.

25:41Caroline Hyde:Right. Bloomberg's Brody Ford, thank you very much. All right. So the state of play is that the Nasdaq 100, which is a very tech heavy index, is up for a 10th straight day. Best run of gains since 2021. Despite everything that's happening around the world, technology is plowing on. Coming up, we're going to speak to Denny Fish, Portfolio Manager on the Global Technology and Innovation Team at Janice Henderson Investors. Somebody that has funds in the billions and billions of dollars is zeroed in on the biggest names of technology and I hope can tell us why Tech powers forward amid all these geopolitical crises.

26:20Caroline Hyde:It is halftime, and this is Bloomberg Tech.

26:31Welcome back to Bloomberg Tech, a check on these markets that are continuing to ramp higher. We've had 10 straight days of gains in the Nasdaq 100. We'll dig into that in a moment. But today, what fuels us higher? optimism that we'll have a second round of peace talks between the United States and Iran, and we're up more than a percentage point on the Nasdaq 100. Amazon, M &A spending$11.6 billion to buy a satellite company, and Global Star jumps almost 10%, and that's on the back of a nice little premium being paid by Amazon. Move on and have a little look at what has been an extraordinary stretch.

27:01For technology, amid all the volatility, geopolitical angst, we saw the S &P 500 shake off those losses of the conflict yesterday. And now we have a tenth day of gains,$3 trillion added to the Nasdaq 100. The question is, can we continue?

27:16Caroline Hyde:Let's get more on the perspective and the markets with Denny Fish, portfolio manager on the global technology and innovation team at Janus Henderson Investors. Something we looked at yesterday is one case study was Intel, right, and gaining more than 50 % over nine sessions. But the Nasdaq 100 is on track for 10 straight days of gains while we have a war in Iran, while energy prices are spiking. Given the technology sector has momentum, what is it that investors are basing that on, what are the factors they're assessing in their optimism? Yeah, I think the primary factors are that regardless of what the geopolitical environment looks like, and obviously I have to couch out a little bit because if it got extreme, this won't hold.

28:02But the trends that have persisted for the last three years because of the advances and expected advances in artificial intelligence just won't slow. And so unless you have a belief that we are starting to plateau in some form as it relates to the models, or that there's some other major constraint that's going to constrain our ability to actually advance these technologies, you kind of have to participate in the investment cycle because this could go on for several years. And, you know, having covered technology or worked in the technology industry for over 30 years, you know, these trends persist much longer than investors could ever imagine.

28:46And this is the most profound technology shift in our lifetimes. And I think that's why the market has quickly come back to the stocks that were working before the conflict actually started. Denny, before the conflict started, software wasn't working. And actually today it is. But that hasn't been consistent in the last few days, particularly when you think about Mythos, the latest models coming out of Anthropic and what that means for software writ large. What's your take on how beaten up the sector's got? Yeah, it's really hard because the fundamental thing that's going on right now in software is that the market, broadly speaking, is questioning the terminal values of the companies.

29:24And once you start questioning the terminal values of the companies, it's really, really hard to disprove a bear case. And the reason that this just continues to get questioned, and we've seen it, really, it started in mid-December with the launches of Claude 4.5 and Codex from OpenAI. Those were step function changes. And I think in an environment where investors were expecting maybe more moderate increases in capabilities, we saw a step function change. And then with 4.6 from Claude, we just saw another step function change. And what's important is, you know, these context windows are getting much longer.

30:04The agenting capabilities are becoming amazing. You combine those two things, these products now, these models can do things that humans, the best humans can barely do when it comes to software development. So really what it comes down to then is what is going to be the barrier to entry for software and how sustainable are the competitive advantages of the existing incumbents. And I think one of the challenges is that not only do you have threats from the models at the moment, but all these capital market cycles generally kind of look and feel the same. And as we get into 2027, 2028, we're likely to see an environment where you see a whole slew of agentic first software companies that have been funded in the venture community over the last three or four years that are then going to be coming to scaled enterprises growing really fast.

30:59And typically that's a more attractive environment. And so don't buy into software at these levels, Denny? Like how do you make that decision or is it safe to be with the picks of the shovels, with the energy, with the infrastructure, the chips? I think it's still safer to be with the picks and shovels and the infrastructure that's building out for these models. And, I mean, if we just look, I think a great data point that we just recently got, if you read Andy Jassy's shareholder letter from Amazon, and he just clearly lays out what's going on within their infrastructure. And once you really understand kind of the constraints that we're still facing and the returns that these companies expect to get over time, you can start to get more comfortable with that.

31:43But, you know, as far as the eye can see, I was just in Asia. I mean, we are supply constrained in AI infrastructure, likely, you know, well through 2027, maybe into 2028, depending on, you know, what components that you're looking at. But nonetheless, this building, this build out is going to continue. And I think we're going to continue to see the same thing that we've seen the last three years where we get closer to the end of the year. And then the expectations around CapEx, whether it's the hyperscalers, enterprises or governments, just continues to exceed expectations.

32:17Caroline Hyde:So to illustrate that point, the top two holdings in the two biggest funds that you're across, Denny, are NVIDIA and then TSMC. And the state of play so far in 2026 is that the NASDAQ 100 is up modestly and the SOX, excuse me, is massively outperformed. I think going back to Caroline's last question, when does that change? When do you, which point in time do you adjust your holdings to reflect more dominance of the software names benefiting from the investment? Yeah, I think a couple of things would need to happen. I mean, the reason that AI infrastructure continues to work is, number one, the fundamentals remain quite strong and continue to exceed expectations.

32:56That's not just TSMC and NVIDIA. That's the memory complex. That's optics. That's energy. I mean, we can, you know, data center infrastructure. You can go, you know, down the whole list and they all continue to participate. but you would either need those fundamentals really to start to slow, you know, lower than what investor expectations are, or the valuations to blow out the levels where then you just get really uncomfortable at what your return could be despite really strong fundamentals. And so far, fundamentals continue to be strong and valuations are reasonable across a lot of the ecosystem.

33:32And you contrast that to software, for example, where maybe fundamentals are stable, Well, maybe they're not declining yet, but as I mentioned, you have terminal value risk. And then you actually, if you're looking at these things and comparing them on, say, like a gap basis or adjusted for stock-based compensation or the things that a lot of generalist investors would do, the valuations actually might not look that attractive in many cases. And that's why I think you have to be careful. We're still investing in software, to be clear, but we're taking what I call a rifle shot approach versus a shotgun approach when we think about the names that we'll invest in.

34:09And more likely than not, the better place to be invested is in infrastructure software that's based on consumption. It's really going to benefit from this build out as well versus the application space. Danny Fisher, Janice Henderson, we so appreciate it as you've been on the road a lot. Thanks for checking in. And look, let's check out another area that NVIDIA is injecting some life into, quantum. Now, we understand NVIDIA is launching its world's first family of open-source quantum AI models called NVIDIA Icing. That's named after a mathematical model. And really, this is designed to help researchers, enterprises, build quantum processes actually capable of running useful applications.

34:46Look at the stock market reaction, D-Wave, IonQ, Inflection, which is part of this partnership in particular, all up and to the right end.

34:53Caroline Hyde:Okay, coming up, the Treasury Department's tech team wants access to Anthropik's mythos. We'll discuss the AI model cybersecurity impact next. This is Bloomberg Tech.

35:23lines, in-depth analysis, and big interviews. All the stories that hit home on your days off. And I'm Lisa Mateo. Watch and listen to Bloomberg this weekend for thoughtful, enlightening conversations about business, lifestyle, people, and culture. On Saturday mornings, we put the past week's events into context, examining what happened in the markets and the world. Then on Sundays, we speak with journalists, columnists, and key political figures to prepare you for the week ahead. Join us as soon as you wake up and bring us with you wherever your weekend plans take you. Watch us on Bloomberg Television, listen on Bloomberg Radio, stream the show live on the Bloomberg Business app, or listen to the podcast.

36:00That's Bloomberg this weekend, Saturdays and Sundays, starting at 7 a.m. Eastern. Make us part of your weekend routine on Bloomberg Television, radio, and wherever you get your podcasts.

36:21The U.S. Treasury Department is seeking access to Anthropic's Mythos model so it can begin hunting for vulnerabilities. Now, that's according to a source. Anthropic has warned the powerful model may be capable of powering cyber attacks and has released it only to a limited group of institutions thus far. Let's get the latest Bloomberg cyber editor, Jeff Stone. Is it? As powerful as all this is a good reason why part of the government, we might say, while the Pentagon is currently in a legal fight with Anthropic, should be getting access to the model. Anthropic hasn't released this model publicly, which is really important for everyone, especially in the general public.

36:56Cybersecurity researchers don't have access to this tool, so we can't independently verify that. Initial feedback has been that this tool is quite powerful. This is the kind of thing that it's essentially what Anthropic describes as the ultimate kind of red team tool, that if you unleash it within your systems, you're going to use it to find as many bugs and vulnerabilities and potential points of exploit before the bad guys can.

37:20Caroline Hyde:So according to our source, right, the Treasury Department want access to it to identify vulnerability. Is there any more flesh to that that you can take us through, Jeff, what the Treasury Department would do with it? There's also the background that Treasury Secretary Besant was part of the official group that warned Wall Street about taking this seriously, right? That's right. The Besant meeting that he had with Fed Chair Powell is a really important piece of context to this. When they summoned all these Wall Street leaders to a meeting with them in Washington to urge them to say, hey, you should unleash this tool.

37:56You should try to find your own kind of flaws and vulnerabilities before outside attackers do. You know, at that time, we know that Treasury and some government agencies were in discussions with Anthropic about using this tool. They want it for the same reasons that they're encouraging so many of their partners to use it.

38:15Caroline Hyde:Bloomberg's Jeff Stone with the reporting. Thank you very much. Let's get more on the capabilities and impact of Anthropic's mythos. Let's speak with Jaya Baloo. She's the chief operating officer and chief information security officer at the AI cybersecurity firm Aisle. A week ago, I spoke to Mike Krieger, who leads Anthropics Labs team. And in the context of Mythos, he was saying, when you have a model this powerful, you can't actually isolate one capability. But what's so confusing about that is we are only talking about Mythos' ability to identify vulnerabilities quicker at great expense. But that's the very narrow use case.

38:54Caroline Hyde:What have you made of that? Well, so actually it was never in doubt that we can use AI to discover vulnerabilities. And at Aisle, we've been doing that since August 2025. We found all of the vulnerabilities in something that's been a very hard code base, like OpenSSL. We've been finding those vulnerabilities. So using AI to find vulnerabilities is not a unique capability. What I think is interesting about Mythos is it's supposed to be the most powerful model yet. And I have to say, like, there is a contention here. I disagree with that because I think one of the things you see is that, you know, the vulnerabilities that they've actually surfaced and disclosed are not impossible to find, not with a closed source model like Mythos, but also with small, tiny open source models.

39:40And that's incredibly significant because Mythos has now been restricted use to only like 40 companies. They are incredibly deterministic about who gets access to it. And I think that what we should expect is that if we really want to maintain a state of security, specifically with open source security vendors, then we need to make sure that they also have the capability to do this. So one of the things that we're really intent on is releasing something open source that is comparable to finding those same kind of vulnerabilities. And we've proven that we can actually do it. How can you, Jaya, do that safely?

40:17And do you think, therefore, that this is being done safely from a mythos general purpose model perspective? Or is it more the hype that's happening around it? Well, I think there's always, you know, I think with all of this stuff, there is always a degree of hype. And the real effect, you can only actually validate it when you have access to the model and you compare it to other models. But if you just take a look at the vulnerabilities that they disclosed, we can confirm that we can find those vulnerabilities. not with a closed source model from Anthropic, but an open source model like GPT 5.4.

40:51So and that is the tiniest model that we can test with. And we can confirm that we can find those same vulnerabilities. And the reason to make this more widely available, this open source, tiny LLM model, is this is something that I believe that we should always do, because defenders now globally are worried about who else can build this. If we've already proven previously you can do this with open source, it's just a matter of time before someone else figures out how to do it. So without making it open source, we're actually giving defenders a disadvantage. And I think the real issue is how can you get defenders up to scale globally, and not just 40 companies in Silicon Valley, but everybody on a level playing field in order to defend their stuff.

41:37So open source maintainers need to be able to find their own villages in order to fix them. Are attackers already looking at these smaller open source models? What have you seen, particularly as we are in the midst of a global conflict where Iran has very sophisticated cyber attackers, at least? I think the best example that we saw is the recent example that was hit by the towards the Mexican government, where you saw a bunch of compromise happening because of AI, because of an attacker using AI to hack the government of Mexico. So I think we are definitely going to see more of this. I worry about, again, the ability for people who are trying to build foundational Internet capabilities to have access to things, to find issues in their own code base and then defend them proactively versus having a sort of small community only have access.

42:35I think we need to democratize this because the building blocks of the Internet are democratized. They are in the hands of everyone.

42:42Caroline Hyde:Jaya, this is the view of Bloomberg Intelligence, our in-house analysts, right? And that is that mythos on paper, because you quite rightly point out the limited scope of access, lacks the capability to replace leading cybersecurity providers like yourself. Beyond vulnerability assessment, right? It's good at identifying memory corruption, logic vulnerabilities. Do you accept that, that that's where its limits end? because the conversations I've had every day for seven days is that this is a model more capable than any we've seen before. And you know what? I think Mythos is actually doing, let me just maybe caveat this by saying I think they're doing a really good job.

43:22I think it's hard, though, for us to actually validate some of that it's the greatest model ever unless you can actually do it because we can reproduce this with three, and I want to emphasize the word free here, free open source models, the smallest models are capable of finding the same thing. So it's not about the model. Although Anthropic makes it about the model, it's really about the system and the scaffold and everything you built around it. Because what we see is that this capability is quite jagged. So it's not good at everything equally as well. It's good at some things, like you pointed out, but it's not good at everything.

43:58And that's just something that we also see in the open source model. But given enough open source model, you can highly parallelize all of this stuff. So instead of relying on one exceptionally intelligent model, you can take this kind of raw intelligence capability, farm it across many models, and then actually find more. This is something that we've proven. We can actually find bugs that Mythos has missed. And we're doing this together with the open source community. For example, Hugging Face models and GPT 5.4. And again, doing this for free means that there's a certain degree of asymmetry right now that's being held by Anthropik.

44:38That asymmetry will decline. And to your point, we will have other actors also capable of doing this. So it's about giving defenders the leverage here to figure out their own issues and fix them. Jaya, it's great catching up with you. Jaya Baloo, Avile, we really appreciate your time now coming up. U.S. residential electricity prices, they have spiked in the last five years. And how much blame is AI getting for that jump? What's it going to mean for the midterms? More on that next. This is Bloomberg Tech.

45:12The rapid build-out of AI infrastructure has seen a spike in energy prices. Wholesale electricity in areas with significant data center activity has risen as much as 267 % in five years. Now, that's making the cost of electricity for the 2026 midterms well, what the price of eggs was in the 2024 election. That's the focus of the latest Bloomberg Big Take and one of our reporters on it, energy reporter Josh Saul, is here with us now. And I want to clear something up first and foremost because is the rap, the bad rap coming to data centers, the right target right now? In amidst oil price issues, amid just supply demand constraints, the upgrade of the utilities more broadly?

45:51People are really angry about their utility bills. Data centers are part of the reason that those bills are getting higher, but they're not the whole reason. There's other things like investments needed for an aging grid, overall inflation. But when you ask people, when I was in the Lehigh Valley in Pennsylvania, when I would ask them about their utility bills,

46:09Caroline Hyde:they would say, I'm so mad about the data centers. They would go right to that. Right. Josh, you know, there is, when it comes to data center impact on any given zip code, there's not in my backyard, right, NIMBY. That's one movement. But specifically with energy prices and how it relates to the midterms, how do we measure the movement of the electorate in how they feel about this issue, in how it might sway votes in key races? The NIMBY issue is more about green space. It's more about what's right next to you. People worry about power prices, but the way that the power prices travel across the grid, whether you can see a data center or whether it's 10 or even 100 miles away, it's still going to be affecting your power bills.

46:53And look, it's affecting Japanese power bills. It's affecting UK power bills. And some of them are data center related and some of them are also energy and import related. But take us to Pennsylvania and take us to whether or not this is being deemed more one side, more Democrat issue, more Republican issue. Because, of course, it is the administration, the current administration that is all in on data center expansion. It's shaping up as an anti-incumbent issue. So there's a Republican incumbent in the Lehigh Valley and people are angry about their power bills. and often that anger is taken out on an incumbent, whether it's a local representative in the Lehigh Valley

47:30Caroline Hyde:or whether it's President Trump. I want to thank Bloomberg's Josh Scholl, who has the big take, and it's a must read. Let's get back to Disney, the story that we broke at the top of the hour. Disney is cutting around 1 ,000 jobs. That's according to Bloomberg sources. Bloomberg reporting that the cuts will be company-wide but impact marketing most. This would be the first layoffs under new CEO Josh DeMauro, who took over from Bob Iger in March. Separately, Bloomberg is reporting and has seen a memo from Josh DeMauro to Disney staff that says that they will start eliminating roles in some parts of the company.

48:09Caroline Hyde:And they've begun notifying those impacted employees. The stock didn't really move when the story broke. Carrie, you know, there's been reports on this in prior weeks, but a big piece of news. Yeah, more agile, technologically enabled workforce is what they want. Well, we talk about that technology and the enablement, and it does it for this edition of Bloomberg Tech. Yeah, recap on the pod. You know where to find it. A lot of important markets, news, technology, conversations. That's where you find it. From New York and San Francisco, this is Bloomberg Tech.

48:46You can get the news whenever you want it with Bloomberg News Now. I'm Amy Morris. And I'm Karen Moscow, here to tell you about our new on-demand news report delivered right to your podcast feed. Bloomberg News Now is a short five-minute audio report on the day's top stories. Episodes are published throughout the day with the latest information and data to keep you informed. Yes, there are other products like this from a variety of news organizations. But they usually rerun their radio newscasts throughout the day. That's not what we do. We create customized episodes that can only be heard on Bloomberg News Now.

49:22And we don't wait an hour to publish breaking news. When news breaks, we'll have an episode up in your podcast feed within minutes. So you're always getting the latest stories and developments. Get the reporting and the context from Bloomberg's 3 ,000 journalists and analysts. We're all over the world. Listen to the latest from Bloomberg News Now on Apple, Spotify, or anywhere you listen.