The AI Race, Starship and Silicon Valley's Defense Bet

23:08but there's not much tritium made anymore and tritium has a 12.3 year half-life so it decays to helium-3 and we're just not producing helium-3 anymore and so interluna's identified the moon as a target market in the future the the sun produces helium-3 in large quantities there's over a million metric tons of of helium-3 on the moon and its lunar regolith and in developing that plan to go to the moon and extract helium-3, we devised this technology we call cold capture to separate helium-3 here on Earth.

23:51Ed Ludlow:So it's the same technique. Yes, you heard right, Bloomberg Tech audience, we're talking about harvesting helium-3 from the moon. That's the ultimate goal, right, Rob? But I guess from an economics perspective, You know, that technique has a higher yield on the moon. Like, get us to the point where this is real. Yeah, it absolutely will have a higher yield on the moon. We plan with our go-to-market in the early 2030s to produce 10 kilograms of helium-3 on the moon. Here on Earth, by deploying cold capture to grade A helium plants on Earth, we expect to produce about two and a half kilograms of helium-3 annually.

24:29Italy. And we have customers that have already signed up to buy helium three from us over, well, just under$500 million in purchase orders, binding contracts that extend from 2028 out to seven to 10 years in the future. So we had to find a way to service those customers earlier. And by applying this technology that we've developed for the moon to this earth application, we're able to do that.

24:57Ed Ludlow:Rob, just very quickly, have you a pathway or a plan in place to get your technology to the moon's surface? Who is the launch provider? Who's going to help you do that operationally? Well, we have a little bit of time to decide on that, Ed, but we are looking to the Artemis program and launch companies like SpaceX and Blue Origin, but other companies that are building landers like Firefly, Intuitive Machines, and others. We plan to invest alongside the Artemis program, the NASA program to go back to the moon and then on to Mars, and we will use that infrastructure that they developed. Now, of course, Ed, the infrastructure we need to harvest large volumes of helium-3 from regolith is infrastructure that also can help massive, build a moon base, build roads and build, do construction on the moon that will support large scale and long term lunar economy.

25:55Ed Ludlow:Interlune CEO and co-founder, Rob Myerson, thank you very much for your time. Now, coming up, it's been a fascinating weekend in AI and there is a lot to unpack. Yep, Kimi K3, but also bigger questions about the future of American AI. Mary D 'Onofrio, partner at Crosslink Capital, joins us to discuss the next frontier in AI investing. This is Bloomberg Tech.

26:26Ed Ludlow:Welcome back to Bloomberg Tech. We're rebounding in the technology sector when it comes to public markets. Remember, chip stocks fell into a bear market Friday, in part because of questions posed by the release of Kemi K3 by China's Moonshot. We're off session highs right now, but the NASDAQ 100 adding almost a percentage point, the stock's up 2%. And it's a lot of the compute names, NVIDIA, but also memory names that are driving some upside in those public markets. Since Moonshot released its Kemi K3 model last week, everyone's debating what it really means. Does more efficient AI mean less hardware?

27:01Ed Ludlow:Is China closing the gap faster than we expected? Where does AI's real value ultimately accrue? Joining us now is Crosslink Capital partner, Mary D 'Onofrio, focused on venture growth investments, but across software, AI, and infrastructure. And for the last three days, I've said Kimi K3 many times. But it was interesting to take a sort of straw poll of VCs, founders, people in industry of how they react to it so differently. Some people are very fearful, defensive. Others say this is very good for AI as a whole. Very difficult job first thing on a Monday. But your reaction to what happened? Well, Kimmy K3, as you said, was phenomenal.

27:44And it's similar to the DeepSeek moment we had last year. You know, 2.8 trillion parameter model. And what it demonstrates, I think, is that China and open-weight models are closing the gap with frontier models. Obviously, it was done based on their own benchmarking. But according to the benchmarks, they're fourth. And while you can debate the merits of those benchmarks themselves, the productivity gains are tremendous. And I think that in general, it is part of the pathway towards more widespread open-weight model usage. And in the context of enterprise budgets ballooning in AI, I actually think that it's probably a boon for AI ubiquity over time.

28:27Ed Ludlow:When you think about your existing portfolio of companies, were you sort of like, OK, how does this impact their trajectory? Is anything actually fundamentally changed for the different pieces of the AI five layer cake that you're invested in? I don't know if it changes the cake itself, but I do think it's another demonstration that models keep leapfrogging one another. And if companies don't out-innovate, they're going to be leapfrogged too. Whether it's, you know, every single model seems to outperform the last. And it's a matter of out-innovating by creating modes of data, of performance, of customer resonance.

29:06And otherwise, unfortunately, you know, the innovations in both Frontier and Openweight models is tremendous.

29:14Ed Ludlow:You're on the crossover investment team, but basically focused on private and maybe later stage or venture growth investments. Seven years at Bessemer as well, prior to being at Crosslink. Sort of introduce us to your overall investment focus, your investment thesis. Yeah, of course. So as you mentioned, I'm a partner at Crosslink Capital. I help to lead the mid-stage venture practice. I spend an asymmetric amount of time in vertical AI and AI infrastructure. What are some examples of what you call vertical AI and AI infrastructure? Yeah, of course. So AI infrastructure is a place I spend a lot of time.

29:49That's the physical and software systems required to build, train, and run models. And in the first wave, it was about how to build a scalable model. And now we have those, as we're talking about right now. And so in the next wave, it's how to deploy them. quickly, fastly, cheaply. Agents, of course, being one of the ways to do so in an efficient way. Taking models as the reasoning layer, adding tools and infrastructure around it to connect outside data in apps, and using memory in order to do it in a way that produces outcomes. And agent infrastructure itself has become a sector that I've spent a lot of time in, whether it's agent harnesses or the identity layer around it.

30:31one of my portfolio companies, Teleport, is a leader in that space. And it's a place that I'm spending an asymmetric amount of time.

30:36Ed Ludlow:I wanted to get to that kind of background on your focus to try and answer some of the questions that the Kimmy K3 moment posed. One idea is that basically by lowering the margins at the model layer, it's an enabler. It opens up the market a little bit. It's a driver of demand, essentially. Do you agree with that kind of thinking in response to this? I do think that it demonstrates what a lot of even the inference infrastructure layer is demonstrating, which is as cost has gone down by 95%, does that displace the total dollar value expended by customers? And I would argue it probably won't because the migration will just be from workloads that are at the frontier, that are most expensive, and of course that require the most reasoning capabilities and are the most advanced, but they'll go to cheaper models that still provide the requisite amount of compute for the question being posed, but do so at much cheaper cost while not sacrificing on things like latency and performance.

31:43Ed Ludlow:It's interesting because Kim K3 pricing$3 per input,$15 per output on the token side, like it's it's quite in line with like what a lot of leading models are also priced at i found that interesting maybe we should talk a little bit about valuations it's not a sophisticated argument but when this all happened a lot of people came out and said well if a 20 billion dollar uh chinese startup can do a 2.8 trillion parameter model with these economics why are we assigning open ai and Anthropic evaluation of a trillion. They looked at the sort of fundamentals of that. Is that the right way to think about it?

32:24Frankly, I don't know yet. And I don't think anybody knows yet. Because even in the context of those questions, Anthropic and OpenAI are still producing billions and billions of revenue. And it's not just from their, you know, they have first-party businesses, they have third-party businesses. And I think, you know, obviously the the API opened for Moonshot, but the open waits have not yet.

32:48Ed Ludlow:Yeah, 27th of July. And it's worth noting that because it's open wait doesn't mean it's free to run. And so I think that's a dynamic to these models that people potentially underappreciate as well. The argument you just outlined is the same that Greylock saw Modern Medi made on Friday on the show that these are very robust businesses that have multiple revenue streams. You are focused on the private markets, but we've had some really interesting IPOs of late. You know, not just SpaceX, but also SK Hynix doing its ADRs. I was in New York City for that. What does that signal to you? I think it's two things.

33:25Obviously, those are fantastic businesses. And I think when it comes to what it says for the markets is that investors are really looking for anything having to do with the AI buildout, AI infrastructure. Obviously, you made the reference to the performance of public market stocks today. But I also think it's one other thing for the capital markets, which is that we have this narrative that potentially the IPO window is reopening, but is doing so with these longstanding businesses with billions of dollars of revenue that are, frankly, older. They're at the cutting edge of innovation, but they've been around for decades.

33:56And so I think it's saying, you know, there is a desire for things having to do with AI. But nonetheless, investors are still, they're hardly speculative. Investors are looking for things that are relatively secure.

34:09Ed Ludlow:Crosslink partner, Mary Denefrio, making the debut on Bloomberg Tech. It's really great to have you here. Thank you for having me. Thank you very much. TSMC is spending at an unprecedented pace, committing a record$265 billion to its U.S. expansion. So with AI fueling enormous capital needs across the chip industry, the question becomes, how will it pay for it? Bloomberg's Steven Engel asked CFO Wendell Huang whether another trip to the bond market could be on the table. Our fundamental belief is that we should be able to self-sustain using our operating cash flow to support the CAPEX and leave enough free cash flow to return cash to shareholder through a sustainable and steadily increasing cash dividends, which we have been doing that in the past several years.

34:59We have very good track record of it. Now from time to time, we can go out to the market, the bond market, if the timing is right. We've been doing it. We've done a$30 billion fundraising in bond market a few years back. About half of it in U.S. dollar, half of it in NT. NT is very cheap funding for us. U.S. dollar is becoming more and more expensive. So we'll see if the timing is correct. If the timing is right, we do not rule out that we can go to the bond market again. Well, given what you just laid out, would maybe a mix of more NT dollar denominated bonds be more likely or yields to a high?

35:50NT dollar bond market is much shallower than the U.S. dollar bond market. So it has its limitations.

35:59Ed Ludlow:That was TSMC CFO Wendell Huang with Bloomberg's Stephen Engel in Taipei. Now, coming up, Silicon Valley is still betting big on defense. We're going to speak with Andreessen Horowitz, new signing, Connor Love, about the startup shaping the future of national security. Connor just joined the American Dynamism team. That's next. This is Bloomberg Tech.

36:23Support for the show comes from Public.com. If you're actively involved in your portfolio, you probably catch yourself repeating the same actions. Buying the dip, manually sweeping idle cash, putting on a hedge. On public, you can now create AI agents that handle all these tasks on your behalf. Just describe what you want to do in plain English. Like, if the VIX hits 25, buy a put option on the S &P 500. Or, if my cash balance goes above$20 ,000, move the excess into my direct index. You approve of the workflow, and your agent handles the rest. Monitoring the market, watching for your conditions, and executing your strategies exactly as defined.

37:01An investing platform driven by your intent, not just your clicks. You can also get full read and write access to your account via the public API. Go to public.com slash market and fund your account in five minutes or less. That's public.com slash market. Paid for by Public Investing. Brokered services by Open to the Public Investing, Inc., member FINRA and SIPC. Advisory services by Public Advisors, LLC, SEC Registered Advisor. Complete disclosures available at public.com slash disclosures. Ask yourself, what are your best people spending their time on right now? Expense reports, receipt chasing, month-end close that takes weeks.

37:41You become what you spend on, and that's not what you're building toward. Brex is the intelligent finance platform that eliminates that work before it starts. AI agents that handle the manual stuff automatically, so your team can spend their time on what actually compounds. It's time to get Brex AF. Learn more at brex.com slash AF.

38:26Visit TheFootheans.com to learn more and where to buy. These statements have not been evaluated by the Food and Drug Administration. This product is not intended to diagnose, treat, cure, or prevent any disease.

38:35Ed Ludlow:Venture firm Andreessen Horowitz is doubling down on its American dynamism thesis and practice, adding defense and space investor and former Army captain Connor Love to the team. And Connor joins us now, joining from Lightspeed. It's great to have you on the show. Thanks for having me. Day one. You're very welcome. Day one on American dynamism. you you know reading about your investments at the the last firm and like how you've approached your career you you would you would call it american exceptionalism right but there's a close alignment between those two ideas just reflect on on what it was that prompted you to join this firm yeah well um first of all thanks for having me um we're in a pretty amazing opportunity right now to invest really in the fabric of america uh and when mark and ben and team kind of approached me.

39:25Look, this is a category that four years ago, they didn't jump on the bandwagon. They said, hey, this is important to the fabric of America. This is important for the entrepreneurs and the technology future for our nation. And when given that opportunity, you know, you kind of get on, you have to get on board and join the team. So it's been amazing, you know, so far just to, you know, dive in and get to know everyone. I think in the end, when you take a step back and just look at the opportunity here, you know, as I said before, what's happening is a once-in-a-generation opportunity where we're redefining what it really means to build in America.

39:56It's not just software primitives. I mean, it's things that are core to the fabric of America. It's defense technology. It's energy. It's many things. And so to join a team like Andreessen Horowitz and kind of continue to invest in back entrepreneurs in this ecosystem, I'm just really thrilled and excited to be able to do it.

40:14Ed Ludlow:Connor, do you mind if I ask a bit more about the backstory? You know, you mentioned Mark Andreessen Ben Horowitz, they approached you. How did this all come together? It's actually funny. One of the partners and founders of American Dynamism, Catherine Boyle, I actually spoke to her about seven years ago. I had the privilege of serving in the military and was actually forward deployed in Iraq, looking for what the next opportunity was in my professional career. And I connected with her really in just kind of a haphazard out of the blue way. And she said, hey, there's this company called Andrel.

40:45You should look at Andrel. Obviously, Everyone knows what Andrel is these days. But I think importantly, I just got exposed to Silicon Valley. And I saw then what I know now that really technology will be a force for good, not just for the defense fabric of our nation, as I mentioned, but also just you can't have an industrial base in America if you don't have a burgeoning defense base. So it's to me a once in a lifetime opportunity that is very much rooted in my background and just happens to connect at three in the morning with Catherine Boyle when she was then at her previous firm.

41:20Ed Ludlow:You've explained it, right, that it's not exclusively defense technology. American dynamism is broader than that. But right now, Anderil is the best known case study. Yeah. Right. Saronic more recently in the news cycle. What is the next phase of American dynamism, do you think, Connor? Yeah, well, I put it in that lens, which is you really cannot have a defense industrial base without an industrial base. And when you look into what goes into these autonomous systems, these new platforms, it's really the subcomponents. It's really the parts. It's really the metal and the end components of what's going on.

41:54So when I think you take a step back and you look at, hey, obviously, defense, we're creating a new opportunity for neoprimes to really capture and serve not just the United States but also our allies. You see things like energy. You see things like core manufacturing. Importantly, you see things like space. Yes, there's a huge defense component to space, but as we put more kilograms of mass up to space in a lot cheaper way, really it is kind of the next frontier of where we're going to be building, not just for defense and kind of government use cases, but all other parts of the economy. So I think when I take a step back and I look at American dynamism, as we call it now, I actually just view it as it's a proxy for how some of the most valuable companies in the world are created.

42:36I mean, historically, you look at Boeing, historically, you look at, you know, all these companies that have been important to America. They've been very important to, you know, the American individual. They necessarily have not always been the most valuable companies. Technology is changing that. Technology is saying you can be both important to America and our allies and be some of the most valuable companies in the world. And that's what we and myself at Andreessen Horowitz are really excited to invest into.

43:01Ed Ludlow:Anderil is a big focus for the Bloomberg Tech audience. You will have seen the news this morning with Archer finally showing the autonomous VTOL platform. I think what would be interesting is you participated in Lightspeed's Anderil investment, Anderil clearly important part of the American dynamism and A16Z more broadly portfolio. How are you going to approach that? Yeah. Well, I think I'll say a couple things about Andrew. The first thing is when you think they've got to their last product, you're wrong. There's another product. And I think the beauty is when you have such a complex use case that is the kind of, you know, Department of War and the American warfighter, there are a lot of problems that need to be solved.

43:44So, you know, to see this kind of partnership with Archer, I think is only net good for the American warfighter and the American public. I think when you take a step back and say, you know, really this market is massive. I mean, there are a lot of problems to be solved in the defense space. I mean, there's everything from missiles and what Castilian's doing to ground autonomy systems. I mean, it is really a massively large market where there's actually a ton of technology problems to be solved below it. I think the underlying substrate that I think is important and what we look at, and Andresa and I've looked at in the past, is really what's happening in defense is this low-cost autonomy thesis.

44:20That is just, you know, it's playing out in the battlefield. go look at what Saronix done in kind of the conflict in Iran, go look at, you know, the front battlefield of Ukraine. When you can introduce something that is literally 10 times cheaper and as performance and in mass, it really changes the game in terms of deterrence and what you can do for our nation. So I think, again, you take a step back, the market's big, and I think there's going to be a ton of opportunity.

44:45Ed Ludlow:Kind of really quick, you know, you mentioned Catherine and Aaron, regular contributors to the show. but with you, somebody joined the team that is a veteran, just in the 30 seconds we have, would you reflect on that a little bit, the skill set and experience you bring? Yeah, I mean, there are more impressive people that have done some pretty amazing things for our country that I want to put in front of me. I do think, though, that time of service really just grounds me in why we are doing this. Like I said at the start, importantly, this is a category where you're not just doing good economically, but I do believe you're doing good for the world and for the American public.

45:19So having worn that uniform puts me in a position just to take a step back and say, hey, this means a lot, not just for our shareholders and for everyone back in Andreessen, but also for America. So I'm deeply thankful and I'm going to keep doing it for decades.

45:33Ed Ludlow:Connor Love, now on the American Dynamism team at A16Z. Great to have you on the show. Thank you very much indeed. Now coming up, Alphabet, Tesla, Intel, and more. Posting earnings this week. We're going to look ahead. We're going to discuss what to expect. This is Bloomberg Tech.

45:56Ed Ludlow:okay tech earnings kicking off and we're expecting some big names this week including alphabet tesla and intel investors will be looking for proof that billions in ai spending are starting to pay off even as cheaper ai models from china have raised some questions bloomberg intelligence senior analyst mandeep singh joins us on what to expect and you know it's been so interesting to read the team's research. There will be short-term focus on CapEx, but I think looking past that, what is it that you want to see? What's the data point that shows this AI thing's real? I mean, cloud revenue is still the key metric when it comes to hyperscalers who are at the forefront of CapEx and everyone is expecting those CapEx numbers to go up.

46:42I mean, the question is, is next year going to be 30 % to 40 % or we are going to have another year of 50 % plus increase in CapEx. I mean, to my mind, the numbers are getting very big in terms of trillion dollars in annualized CapEx from the hyperscalers. So there is a limit to how far they can go, but clearly, you know, 2027 numbers will be higher and we are expecting positive upward revisions justified by accelerating cloud growth from these hyperscalers.

47:15Ed Ludlow:Let's stick with the Google parent as the case study. You know, overall revenue growth is expected to be about 20%, but Google Cloud, I think 65 % year-on-year growth. You know, that's the story, right? That Google Cloud is where the traction with AI shows up. Yeah, no, you're absolutely right. And that's where, you know, the comparisons would be drawn to Azure and the fact that Google Cloud is growing at least 15 % faster faster than Azure. And AWS, I mean, things seem to be picking up, and now OpenAI is deploying on Amazon AWS as well as Azure. So the impact of that on Microsoft, that's going to be front and center, but there's no doubt that cloud infrastructure is where the monetization will be evident.

48:07And look, all the frontier labs are monetizing tokens, but you're not going to be able to glean that from the earnings of companies. It's too early to see that token monetization being visible.

48:21Ed Ludlow:This is what the week looks like. It all kicks off Wednesday. Alphabet and Tesla. IBM, of course, pre-released. But maybe we'll learn a bit more than Intel on Thursday. Bloomberg Intelligence Senior Analyst Mandeep Singh. Thank you very much indeed. A busy week for Mandeep too. As we head out, let's take a look at today's big number. $264.1 million. That's how much The Odyssey, the new movie from blockbuster director Christopher Nolan, took in sales globally in its opening weekend. It was the number one film in the US and Canada over the weekend, with$124.5 million in ticket sales in North America, a large portion of which was generated from screenings in IMAX's giant auditoriums.

49:06Ed Ludlow:Many of them were sold out many months in advance. I'm still yet to see it. That does it for this edition of Bloomberg Tech. A really big focus still on Moonshot's Kimi K3 model, 2.8 trillion parameters, and a lot of questions posed for the world, for AI, for markets, for investors. We got through all of that in the program. So recap it on the podcast. You know where to find it? On the Bloomberg Terminal. As one audience member reminded me, we call it the Bloomy, but it's online too. Have a great week. This is Bloomberg Tech.

49:42Healthcare doesn't always work great. If you've ever waited on a refill or couldn't schedule an appointment, you get it. That's the kind of stuff Optum is changing. They're using data and technology to integrate patient care, pharmacy, and everything else. So healthcare is connected, not complicated. What's that look like? Cheaper prescriptions that are easier to get and care that looks at the whole person how you need it. Optum is helping make healthcare work as one for everyone. Learn more at business.optum.com.

50:39Learn more at brex.com slash AF. Aging is real, and so are the benefits of adding Vital Proteins Collagen Peptides to your daily routine. New Vital Proteins Collagen Sparkling Water. Your daily glow-up, now in three fresh flavors. Strawberry Blossom, Lemon Lime, and Blood Orange. Improved skin health in as little as 30 days thanks to collagen peptides. Cheers to that, or go with our classic collagen peptides, so you can stay vital, stay you. Visit vitalproteins.com to learn more and where to buy. These statements have not been evaluated by the Food and Drug Administration. This product is not intended to diagnose, treat, cure, or prevent any disease.