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

An episode of [Bloomberg Tech](https://invisiblenetwork.ai/podcast/1611767434), 20 Jul 2026, 44 min.

Audio: https://podtrac.com/pts/redirect.mp3/traffic.omny.fm/d/clips/e73c998e-6e60-432f-8610-ae210140c5b1/41764a4f-fc64-4e11-89ba-ae7c0030ab5e/822d4e3d-4921-4792-87ab-b48d013b012d/audio.mp3?utm_source=Podcast&in_playlist=9caafc41-289c-4115-995d-ae7c0030ab75

Bloomberg’s Ed Ludlow breaks down what Moonshot's Kimi K3 means for the race with OpenAI, Anthropic and the chipmakers powering it. Plus, SpaceX gears up for a second attempt at a Starship test flight later this week, and Andreessen Horowitz signs defense investor and military vet Connor Love to the firm's American Dynamism team. See omnystudio.com/listener for privacy information.

## Transcript

[0:02] Speaker 1: Bloomberg Audio Studios, Podcasts, radio news. Bloomberg Tech is live
[0:15] Speaker 1: from the heart of Silicon Valley with Ed la Though
[0:18] Speaker 1: in San Francisco.
[0:22] Speaker 2: This is Bloomberg Tech coming up China's latest AI breakthrough
[0:26] Speaker 2: shaking up the industry. What moonshots Kimmy K three means
[0:29] Speaker 2: for the race with open AI Andthropic and the chip
[0:32] Speaker 2: makers powering it, plus SpaceX gears up for a second
[0:35] Speaker 2: attempt at a starship test flight later this week. The
[0:38] Speaker 2: Stog's underperforming most of the other big nazac IPOs and
[0:43] Speaker 2: andresen Horowitz signed defense investor and military vet Connor Love
[0:47] Speaker 2: to the firm's American dynamism team. Love joins the show.
[0:52] Speaker 2: Good morning from San Francisco. The surprise release of Kimmy
[0:55] Speaker 2: came K three from China's moonshot. It's still everything that
[0:59] Speaker 2: people are talking about reading about. We are rebounding in
[1:03] Speaker 2: technology stocks modestly.
[1:06] Speaker 3: We're off session highs where we opened earlier.
[1:08] Speaker 2: Remember, chip stocks entered a bear market Friday in part
[1:12] Speaker 2: because of all of the questions pose by what Kimmy
[1:15] Speaker 2: K three a two point eight trillion parameters model represents.
[1:19] Speaker 2: Some of those were like, well, if we have more
[1:21] Speaker 2: efficient AI, do we need to spend as much on
[1:24] Speaker 2: commute compute? Nvidia is an example of a name that's rebounding.
[1:28] Speaker 2: A lot of focus has also shifted to the intense
[1:31] Speaker 2: memory requirements for what is happening in open weighted models.
[1:36] Speaker 2: Moonshop's gone from AI upstart to global contender almost overnight,
[1:40] Speaker 2: Kimmy K three drawing praise form rivaling the best models
[1:44] Speaker 2: from open AI and Antropic. But demand has been so
[1:46] Speaker 2: strong that the company says it's temporarily limiting new subscriptions
[1:51] Speaker 2: as it adds to work works that add more capacity.
[1:54] Speaker 2: Bloombergs executive editor for Asia tech piece elstroms with us
[1:57] Speaker 2: from DC. There are so many reports on the Bloomberg
[2:02] Speaker 2: this morning, so many places we could start, but I
[2:05] Speaker 2: want to go to one of the pieces that the
[2:07] Speaker 2: team did about Okay, it's Monday morning. We probably went
[2:11] Speaker 2: into the weekend thinking, oh no, we don't need to
[2:13] Speaker 2: spend as much on compute hardware in that context. Now
[2:17] Speaker 2: we have a piece on the Bloomberg saying this, there's
[2:20] Speaker 2: a lot of emphasis on memory.
[2:21] Speaker 3: What's the team and age you're writing about.
[2:25] Speaker 4: Well, yeah, let's broaden up the lens A little bit here.
[2:29] Speaker 4: What's going on is China is holding this World AI
[2:32] Speaker 4: conference in Shanghai, and we're seeing their companies come forward
[2:36] Speaker 4: with new offerings in all sorts of different areas. Moonshot
[2:39] Speaker 4: is certainly the one that got the most attention over
[2:43] Speaker 4: the weekend. We saw this company that was probably not
[2:46] Speaker 4: that well known outside of China if at all, and
[2:48] Speaker 4: all of a sudden people are paying attention to it.
[2:50] Speaker 4: As you mentioned, its performance is very close to the
[2:53] Speaker 4: frontier models two point eight trillion parameters. Essentially the synapses
[2:59] Speaker 4: of this the brain synapses of this model are very significant.
[3:03] Speaker 4: But also you saw Ali Baba come out with a
[3:05] Speaker 4: new version of their Glen model that is very close
[3:08] Speaker 4: behind it's two point four trillion parameters. It's a very
[3:11] Speaker 4: powerful model too. So I think the big lesson, the
[3:13] Speaker 4: big takeaway here that China AI is much closer behind
[3:17] Speaker 4: the US in terms of AI capabilities than we really realized.
[3:21] Speaker 4: If you thought China was six to twelve months behind before,
[3:25] Speaker 4: now it looks like maybe they're only a couple of
[3:27] Speaker 4: months behind, and they're catching.
[3:28] Speaker 3: Up quite quickly.
[3:29] Speaker 4: And it's a broader selection of competitors that we really anticipated.
[3:33] Speaker 4: People have been focusing mostly on Deep Seek in particular,
[3:36] Speaker 4: maybe a little bit on Ali Baba and then a
[3:38] Speaker 4: couple of the smaller models. But now we're seeing companies
[3:41] Speaker 4: like Moonshot come forward and they're very competitive and a
[3:43] Speaker 4: lot of people didn't even realize that they were going
[3:45] Speaker 4: to be in this race.
[3:47] Speaker 2: There's some really important Bloomberg reporting on what Moonshot's looking
[3:50] Speaker 2: at financially. We're going to get to that with Bloomberg's
[3:52] Speaker 2: baty dip shots in just a moment. But away from
[3:54] Speaker 2: sort of the numbers, what do we know about the
[3:57] Speaker 2: people behind Moonshot, Like, what do we actually know about
[4:01] Speaker 2: this startup? And it's team that have been just thrust
[4:05] Speaker 2: into the limelight in such a quick and short space
[4:08] Speaker 2: of time.
[4:10] Speaker 4: Well, the founder of Moonshot is a thirty three year
[4:13] Speaker 4: old Yang Jie Lynn. He graduated from Carnegie Mellen. He
[4:17] Speaker 4: got his degree in the United States. He decided to
[4:19] Speaker 4: go back to China because of the possibilities that they
[4:22] Speaker 4: had there and really before Deep Seek, he was kind
[4:24] Speaker 4: of seen as the rising star within the AI universe.
[4:28] Speaker 4: He's a big music fan. He named the company in
[4:31] Speaker 4: Chinese after Dark Side of the Moon, The Pink Floyd album,
[4:34] Speaker 4: which apparently is his favorite. They have conference rooms. But
[4:37] Speaker 4: beyond being this geek and this music fan, at the
[4:40] Speaker 4: same time, he's been at kind of the cutting edge
[4:42] Speaker 4: of AI developments there and he made a very conscious choice.
[4:46] Speaker 4: Rather than trying to make a lean, efficient model, he
[4:49] Speaker 4: decided to go very big with this version of Kimmy
[4:52] Speaker 4: Kimmy three and try to make something that was very
[4:55] Speaker 4: close to the frontier edge that we're seeing in the US.
[4:57] Speaker 4: So I think Yang is somebody that we're going to
[5:00] Speaker 4: hear more about the same way that we talk about
[5:02] Speaker 4: deep seeks found early on. I think we're going to
[5:05] Speaker 4: see more breakthroughs, more innovations, and people are going to
[5:07] Speaker 4: start watching him much more closely than we've seen in
[5:10] Speaker 4: the past. We had a nice scoop over the weekend
[5:12] Speaker 4: also that Moonshad is planning an ongoing public so we're
[5:15] Speaker 4: going to be able to look at the financials of
[5:17] Speaker 4: their company in the way that we've seen some of
[5:19] Speaker 4: the other Chinese AA models, and by the way, we
[5:21] Speaker 4: haven't seen that from the US players so far.
[5:24] Speaker 2: A lot of key figures critical to the Bloomberg Tech
[5:26] Speaker 2: audience just kind of remarked over the weekend that this
[5:29] Speaker 2: could have been an American startup. You know, there's a
[5:32] Speaker 2: lot of focus on the US's own policy with regards
[5:36] Speaker 2: to talent immigration. We'll get to all of that later
[5:38] Speaker 2: in the program. Bloomboa, speede Elstrom, thank you very much.
[5:40] Speaker 2: We'll stay with moonshop you They just mentioned it. The
[5:43] Speaker 2: company isn't just making waves with its latest AI model,
[5:45] Speaker 2: it's now setting its sites on the public markets. The
[5:48] Speaker 2: startup is said to be preparing for a Hong Kong
[5:51] Speaker 2: IPO that could come in the next six months following
[5:54] Speaker 2: an absolute surge in demand. Its valuation above thirty billion
[5:58] Speaker 2: dollars right now, according to source, has been expated at
[6:01] Speaker 2: Schultz All Things IPO. Was so interesting to read through
[6:04] Speaker 2: this one. Let's start with the basics. What are we reporting,
[6:08] Speaker 2: what do we know about how quickly they're moving to
[6:11] Speaker 2: do a Hong Kong listing, and any of the sort
[6:13] Speaker 2: of valuation considerations around that.
[6:15] Speaker 3: Yeah, and you kind of nailed it on the head
[6:17] Speaker 3: with the intro.
[6:18] Speaker 5: So closing up around right now about thirty billion dollars, again,
[6:21] Speaker 5: a small drop in the bucket relative to what we've
[6:23] Speaker 5: been reporting around the Anthropics and open eyes of the
[6:26] Speaker 5: world here in the US and the expectation that in
[6:28] Speaker 5: the next six months they could be pricing that IPO
[6:31] Speaker 5: in Hong Kong. The main thing to keep in mind,
[6:33] Speaker 5: Bloomberg is reported a deep Seek is looking at a
[6:36] Speaker 5: mainland listing sometime in twenty twenty seven. So these two
[6:39] Speaker 5: now premiere companies in Asia in China could actually be
[6:44] Speaker 5: racing to go public, kind of similar to what we're
[6:47] Speaker 5: seeing with Anthropic and Open Eye. But again, mainland China
[6:50] Speaker 5: versus Hong Kong versus US listings open up completely different
[6:54] Speaker 5: investing universes.
[6:56] Speaker 3: So go into that in a little bit more.
[6:58] Speaker 2: De So I find that really interesting, Right, A big
[7:01] Speaker 2: pulse of the story over the weekend is that this
[7:04] Speaker 2: is a Chinese AI startup. Right, we are saying that
[7:07] Speaker 2: looking at Hong Kong as a listing, what are the
[7:10] Speaker 2: kind of political considerations around that as opposed to the US.
[7:14] Speaker 2: Where do the pools of capital tend to come from?
[7:16] Speaker 2: If you do an HK listing.
[7:18] Speaker 5: HK listing opens up an entirely different investing universe, So
[7:22] Speaker 5: you have the rest of really all of Asia. So
[7:24] Speaker 5: you think about Hong Kong based investors, do you think
[7:26] Speaker 5: about people in Australia and Singapore and other really countries
[7:30] Speaker 5: that could tap that listing. Again, US investors sometimes do
[7:34] Speaker 5: invest in Hong Kong, but similar story to what we
[7:35] Speaker 5: track with sk Heiniz, it's a bit of a higher
[7:38] Speaker 5: bar to clear and there had been some consternation around that,
[7:41] Speaker 5: whereas on the flip side, a deep seek listing in
[7:44] Speaker 5: mainland China, completely different investing universe. It's a really kind
[7:48] Speaker 5: of local driven investing community. You do have retail numbers
[7:52] Speaker 5: that are kind of eyepopping. It looks a bit kind
[7:54] Speaker 5: of like gambling and trying to get in on some
[7:56] Speaker 5: of those options. So the big push in actually the
[7:58] Speaker 5: big news is the fact that Moonshot wants to tap
[8:01] Speaker 5: the HK market just again because that does open an
[8:04] Speaker 5: entirely different investing universe, and we kind of opening up
[8:07] Speaker 5: more similarly to an Ali Barber or some of those
[8:09] Speaker 5: other companies JD dot Com that are listed in Hong Kong.
[8:12] Speaker 2: And also let's sit here in the US Bloombex Beley Lipschutz,
[8:16] Speaker 2: thank you very much. Indeed, in competition among Chinese AI
[8:19] Speaker 2: developers also continues to intensify. Ali Barber just unveiled its
[8:24] Speaker 2: latest flagship Queen Model which it says is second only
[8:28] Speaker 2: to Anthropics latest system, which sent Hong Kong shares of
[8:31] Speaker 2: Vali Barber rising on that news. The Sunday release came
[8:35] Speaker 2: only days, of course, after Moonshot unveiled its latest K
[8:39] Speaker 2: three model. We have some breaking news crossing the terminal
[8:42] Speaker 2: in the last few minutes. UK's Rachel Reeves has resigned
[8:46] Speaker 2: as Chancellor of the Exchequer. In a post on X,
[8:49] Speaker 2: she says it's been quote the privilege of my life
[8:52] Speaker 2: to serve as the Chancellor of the Exchequer, the UK's
[8:55] Speaker 2: equivalent of a finance minister or Treasury secretary. You can
[8:59] Speaker 2: see that as an immediate response in the pound dropping
[9:03] Speaker 2: against the dollar one thirty four, where it currently stands. Meanwhile,
[9:08] Speaker 2: new Prime Minister Andy Burnham just told reporters the Labor
[9:12] Speaker 2: government will use quote any flexibility within its fiscal rules.
[9:16] Speaker 2: He said he would stick to the self imposed rules
[9:18] Speaker 2: which were adjusted by Rachel Reeves two years ago to
[9:22] Speaker 2: allow for greater capital spending.
[9:24] Speaker 3: It's a dynamic situation in the UK.
[9:26] Speaker 2: Government will keep you up to date as things change.
[9:29] Speaker 2: Coming up the countdowns back on SpaceX targets Thursday for
[9:34] Speaker 2: another attempt for Starship's thirteenth test flight.
[9:38] Speaker 3: We've got more on that next.
[9:39] Speaker 6: This is Bloomberg TECHX.
[10:00] Speaker 2: Is gearing up for another attempt at a thirteenth Starship
[10:03] Speaker 2: test flight.
[10:04] Speaker 3: The company is set.
[10:04] Speaker 2: Thursday as the new launch date after last week's attempt
[10:08] Speaker 2: was scrubbed and right at the last moment, shares we're
[10:12] Speaker 2: down about one and a half percent. This has been
[10:15] Speaker 2: a broader story post IPO on SpaceX kind of underperforming.
[10:19] Speaker 2: For more on SpaceX, Brett Lindsay ma Zoi, managing director
[10:23] Speaker 2: and head of Industrial sector Research joins US, recently initiated
[10:26] Speaker 2: on SpaceX has an outperformed rating with the stock price
[10:30] Speaker 2: target of two hundred dollars per share. You know, Brett,
[10:34] Speaker 2: I appreciate that Jas there's a clear thesis at the
[10:38] Speaker 2: core of your coverage of SpaceX, right, but I wanted
[10:40] Speaker 2: to get to the idea that I'm sat here at
[10:42] Speaker 2: this desk during the Starship attempt last week, and it
[10:47] Speaker 2: was amazing to see how quickly in after hours trading
[10:51] Speaker 2: the stock dropped right four percent in the moment because
[10:55] Speaker 2: the test was aborted. Right at the final moments, all
[10:58] Speaker 2: of these future business lines on the AI side, NEO
[11:02] Speaker 2: cloud side, starlink orbsort data center. They're all predicated on
[11:07] Speaker 2: Starship not just working, but being rapidly reusable. So can
[11:12] Speaker 2: we just start there and sort of your response to
[11:15] Speaker 2: tracking the cadence of Starships development.
[11:19] Speaker 3: Yeah, you bet, Thanks for having me on.
[11:20] Speaker 7: So, yeah, this was like thirteen recent delay was some
[11:25] Speaker 7: minor technical issues that were identified pre launch. And I
[11:29] Speaker 7: remind you this is version three, so it is a bigger,
[11:32] Speaker 7: more powerful vehicle they've debuted about two months ago. The
[11:36] Speaker 7: automatic abart ad ignition was really the function of for
[11:39] Speaker 7: the thirty three Rapptor engines on the booster didn't immediately
[11:43] Speaker 7: light up and the onboard system caught it. It stood
[11:46] Speaker 7: the vehicle down before the liftoff, and that's a safety
[11:49] Speaker 7: system really doing, you know what it's supposed to do.
[11:51] Speaker 7: Musk in the team, you know they've identified the root cause.
[11:55] Speaker 7: Within a few hours, two engines are being swapped out.
[11:58] Speaker 7: They're already back on the pad targeting as you mentioned
[12:01] Speaker 7: this Thursday for relaunch. That's about a week of turnaround
[12:05] Speaker 7: on a pretty complex rocket. I think these development programs
[12:09] Speaker 7: that you've seen historically at this scale, you know, from
[12:13] Speaker 7: a diagnose and fixed cycle that's that's relatively quick versus
[12:17] Speaker 7: anything that's structurally wrong with you know, with the ship.
[12:20] Speaker 2: So Brett, here's the thing I'm reading your research investment thesis.
[12:25] Speaker 2: SpaceX is not a rocket company. We're just talking about rockets.
[12:31] Speaker 2: Explain your thesis.
[12:33] Speaker 7: Yeah, No, absolutely, I would say that the you know,
[12:37] Speaker 7: the enabler of the starlink opportunity and the connectivity opportunity
[12:43] Speaker 7: longer term is going to be dependent on the success
[12:47] Speaker 7: in the reusability cadence of of of Starship. Falcon nine
[12:52] Speaker 7: is the current vehicle they use this year. They've got
[12:55] Speaker 7: really good visibility on some of the you know, the
[12:57] Speaker 7: payload to orbit densifying the constant relation on starlink for
[13:01] Speaker 7: starlink on the actual Facon nine. But we do think
[13:04] Speaker 7: long term it will hinge upon the success of Starship.
[13:09] Speaker 7: You know, right now they do have a very you know,
[13:11] Speaker 7: big running start relative to the other competitive set. About
[13:15] Speaker 7: eighty five percent of all mass to orbit was you know,
[13:18] Speaker 7: was enabled by SpaceX over the last three years. And
[13:23] Speaker 7: you know, operational cads will be important. These launches this
[13:26] Speaker 7: year will be important, and they're you know, they're making
[13:29] Speaker 7: inroads in there. You know, well, we'll see on Thursday.
[13:32] Speaker 7: But yeah, that will be a big enabler and gating
[13:34] Speaker 7: factor for the success over the next couple of years.
[13:37] Speaker 2: Within your two hundred dollars price target, how do you
[13:40] Speaker 2: sort of break out each business line within that price target.
[13:46] Speaker 7: Yeah, no, it's a great question. So when we look
[13:48] Speaker 7: at the you know, the next three year roadmap, we're
[13:52] Speaker 7: really taking a large discount relative to the AI platform.
[13:56] Speaker 7: So near term, a lot of our price target is
[13:59] Speaker 7: going to hinge upon this success of star Link and
[14:02] Speaker 7: then some of what they're doing on space with the
[14:04] Speaker 7: government contracts and some of the enterprise most recently some
[14:07] Speaker 7: of the contracts with the airlines. We do think there
[14:10] Speaker 7: is pretty good line of site here in the near term,
[14:13] Speaker 7: and we anchored our evaluation on twenty twenty nine and
[14:16] Speaker 7: actually took a pretty large aircut relative to what the
[14:20] Speaker 7: internal plant is. And we're you know, thinking that connectivity
[14:25] Speaker 7: is more optionality around the stock and underwriting it on
[14:27] Speaker 7: some of the near term fundamentals versus a long upshot
[14:30] Speaker 7: on orbital data center in twenty thirty and beyond.
[14:35] Speaker 2: Gret do you model at all for any key man
[14:39] Speaker 2: risk with Elon Musk, You know, how do you kind
[14:42] Speaker 2: of factor that into how you value the company as
[14:45] Speaker 2: much as anything.
[14:48] Speaker 7: Well, that's that's certainly a consideration, right, He's a visionary,
[14:51] Speaker 7: he's you know, very involved in the operations and really
[14:56] Speaker 7: across every single segment and division within the organization. And
[15:00] Speaker 7: you know, all that being said, you know, there is
[15:02] Speaker 7: a there is a blueprint that's that's in place the
[15:06] Speaker 7: you know, i'd say the mostest, they've been able to
[15:08] Speaker 7: develop some of the secular strengths across the business. And
[15:12] Speaker 7: really what they have is you know, ownership within you know,
[15:14] Speaker 7: the launch capabilities in the US and across the globe.
[15:18] Speaker 7: That is the key differentiator as we're thinking about some
[15:21] Speaker 7: of the optionality with the other segments down the road.
[15:25] Speaker 7: So yeah, something always we consider. But you know, he's
[15:28] Speaker 7: built something from what was really a you know, you know,
[15:31] Speaker 7: a vision you know years ago, you know, over decades ago. Uh,
[15:34] Speaker 7: and it's a you know, it's a real company with
[15:36] Speaker 7: real motes and defensibility in our view.
[15:39] Speaker 2: Brett lindsay a Massur on Bloomberg Tech, Thank you very much.
[15:42] Speaker 3: Indeed, now coming up, this.
[15:45] Speaker 2: Space infrastructure startup just found a way to address a
[15:48] Speaker 2: bottleneck in scaling quantum computing in lind CEO and co
[15:53] Speaker 2: founder Rob Mason joins us next this is Bloomberg Tech.
[16:08] Speaker 8: It's time now for talking tech on yahira on on
[16:10] Speaker 8: First Up. CUSP, AAI of Bezos back to British AI startup,
[16:14] Speaker 8: has raised four hundred and fifty million dollars to accelerate
[16:17] Speaker 8: the developments of next generation chip materials using AI. The
[16:21] Speaker 8: company is also launching the AI Materials Foundry, bringing together
[16:26] Speaker 8: more than forty eight organizations including Nvidia, Meta, Hyundai, the
[16:29] Speaker 8: Speed Up Chip Innovation plus Deep Seek founder Li Ang.
[16:33] Speaker 8: When Fund's investment firm fell sixteen percent after China's quant
[16:38] Speaker 8: firms were hit with an AI route, The volatility and
[16:41] Speaker 8: Chinese hedge funds highlight the wild swings LAI related stocks
[16:45] Speaker 8: as bubble concerns intensify, and Boeing says it will need
[16:50] Speaker 8: until the end of the next decade to bring its
[16:53] Speaker 8: next generation jet to market as it continues to repair
[16:56] Speaker 8: its finances for the huge investment and as customers keep
[17:00] Speaker 8: buying its current model. CEO Kelly Orberg spoke with our
[17:03] Speaker 8: own Guy Johnson earlier today at this year's firm both
[17:07] Speaker 8: air show, OH.
[17:08] Speaker 9: The market does not quite ready for the new airplane,
[17:11] Speaker 9: and the customers are telling us, let's focus on the
[17:14] Speaker 9: existing product. We'd like to see better durability of that
[17:17] Speaker 9: product in the marketplace. And as I mentioned, we've got
[17:21] Speaker 9: big backlogs and they're looking for us to deliver on
[17:23] Speaker 9: those backlogs. We are studying a new airplane and we'll
[17:27] Speaker 9: be ready when the market's ready.
[17:30] Speaker 2: This startup is tackling a bottleneck facing quantum computing. After
[17:35] Speaker 2: President Trump made the technology a strategic National priority. Interlun
[17:40] Speaker 2: says it's discovered a new way to harvest helium three,
[17:44] Speaker 2: essential to cool the computers and to scale the industry.
[17:48] Speaker 2: Joining us now is Interlun CEO and co founder Rob Myerson.
[17:52] Speaker 2: I have to say, when this idea hit my inbox,
[17:54] Speaker 2: I was intrigued. Interlune's goal is to harvest helium three
[18:00] Speaker 2: in industrial quantities from the Moon. But in the here
[18:04] Speaker 2: and now, there just isn't enough helium three for the
[18:07] Speaker 2: quantum computing industry on Earth to scale. So you seem
[18:11] Speaker 2: to have found a stopgap. What have you done?
[18:15] Speaker 10: That's right, and thanks for having me. We've we've devised
[18:19] Speaker 10: a technique. Our team has gone back to first principles,
[18:23] Speaker 10: the crack the textbooks and invented a way to separate
[18:27] Speaker 10: helium three using cryogenic distillation from grade A helium here
[18:31] Speaker 10: on Earth.
[18:32] Speaker 5: Uh.
[18:32] Speaker 10: This is an idea that's been around for a while
[18:34] Speaker 10: but never pursued commercially, and as far as we know what,
[18:37] Speaker 10: we're the first to demonstrate the ability to separate helium
[18:41] Speaker 10: three from commercial grade A helium.
[18:45] Speaker 2: The Bloombirtek audience is familiar with the use of helium
[18:48] Speaker 2: and cooling and stabilization in the chip industry, for example,
[18:52] Speaker 2: and so with the war in Iran and the straight
[18:54] Speaker 2: of for Mu's bottleneck, we came to understand that that
[18:57] Speaker 2: there was a shortage there.
[18:59] Speaker 3: How was this for you?
[19:00] Speaker 2: Now, Rob, what's your solution looking like on a timeline basis?
[19:05] Speaker 10: Well, so, the shortage we're talking about is helium three,
[19:08] Speaker 10: which is an isotope of helium that's extremely rare on Earth.
[19:11] Speaker 10: It comes from the decay of tritium and it's used
[19:14] Speaker 10: to chill down superconducting quantum computing chips down to milliclvin temperatures.
[19:21] Speaker 10: We've for decades secured our helium three from tretium decay
[19:26] Speaker 10: in the US and overseas, including Canada, but there's not
[19:30] Speaker 10: much treatium made anymore, and tritium has a twelve point
[19:34] Speaker 10: three year half life, so it decays to helium three,
[19:37] Speaker 10: and we're just not producing helium three anymore, and so
[19:41] Speaker 10: interlun Is identified the Moon as a target market in
[19:46] Speaker 10: the future. The Sun produces helium three and large quantities.
[19:50] Speaker 10: There's over a million metric tons of helium three on
[19:53] Speaker 10: the Moon and it's lunar regular. In developing that plan
[20:00] Speaker 10: to go to the Moon and extract helien three, we
[20:03] Speaker 10: devised this technology we call cold capture to healium to
[20:08] Speaker 10: separate helium three here on Earth.
[20:11] Speaker 2: So it's the same technique we are Yes, you heard right,
[20:14] Speaker 2: blinbotech audience, We're talking about harvesting helium three from the Moon.
[20:17] Speaker 3: That's the ultimate goal, right Rob.
[20:19] Speaker 2: But I guess from an economics perspective, you know that
[20:23] Speaker 2: technique it has a higher yield on the Moon. Get
[20:27] Speaker 2: us to the point where this is real.
[20:30] Speaker 10: Yeah, it absolutely will have a higher yield on the Moon.
[20:33] Speaker 10: We plan with our go to market in the early
[20:35] Speaker 10: twenty thirties to produce ten kilograms of helium three on
[20:39] Speaker 10: the Moon here on Earth by deploying cold capture to
[20:43] Speaker 10: greater a helium plants on Earth. We expect to produce
[20:46] Speaker 10: about two and a half kilograms of helium three annually,
[20:50] Speaker 10: and we have customers that are already signed up to
[20:53] Speaker 10: buy helium three from US over well just under five
[20:57] Speaker 10: hundred million dollars in purchase or binding contracts that extend
[21:03] Speaker 10: from twenty twenty eight out to seven to ten years
[21:05] Speaker 10: in the future. So we had to find a way
[21:08] Speaker 10: to service those customers earlier, and by applying this technology
[21:12] Speaker 10: that we've developed for the Moon to this Earth application,
[21:15] Speaker 10: we're able to do that.
[21:17] Speaker 3: Rob just very quickly.
[21:19] Speaker 2: Have you a pathway or plan in place to get
[21:22] Speaker 2: your technology to the Moon's surface. Who is the launch
[21:25] Speaker 2: provider who's going to help you do that operationally?
[21:28] Speaker 10: Well, well, we have a little bit of time to
[21:30] Speaker 10: decide on that, ed, but we are looking to the
[21:33] Speaker 10: Artemis program and the launch companies like SpaceX and Blue Origin,
[21:39] Speaker 10: but other companies that are building landers, like Firefly, Intuitive
[21:43] Speaker 10: Machines and others. We plan to invest alongside the Artemis
[21:48] Speaker 10: program the NASA program to go back to the Moon
[21:51] Speaker 10: and then on to Mars, and we will use that
[21:54] Speaker 10: infrastructure that they developed now of course, add the infrastructure
[21:58] Speaker 10: we need to harvest large volumes of healium three from
[22:01] Speaker 10: regular is infrastructure that also can help now, so build
[22:05] Speaker 10: a moon base, build roads and build do construction on
[22:09] Speaker 10: the moon that will support it right, large scale and
[22:11] Speaker 10: long term a lunar economy.
[22:15] Speaker 2: Into lun CEO and co founder Rob Us and thank
[22:17] Speaker 2: you very much for your time. Now coming up, It's
[22:20] Speaker 2: been a fascinating weekend in AI and there is a
[22:23] Speaker 2: lot to unpack yep, Kimmy K three, but also bigger
[22:26] Speaker 2: questions about the future of American AI. Mary Dino Frio,
[22:30] Speaker 2: partner at Crosslink Capital, joins us to discuss the next
[22:33] Speaker 2: frontier in AI investing.
[22:35] Speaker 11: This is Bloomberg Tech. Welcome back to Bloomberg Tech.
[22:48] Speaker 2: We're rebounding in the technology sector when it comes to
[22:50] Speaker 2: public markets. Remember Chip stocks fell into a bear market
[22:54] Speaker 2: Friday in part because of questions posed by the release
[22:58] Speaker 2: of Chemy K three by China's Moonshot. We're off session
[23:01] Speaker 2: highs right now, but then as that one hundred adding
[23:03] Speaker 2: almost a percentage point, the socks up two percent. And
[23:07] Speaker 2: it's a lot of the compute names and video but
[23:09] Speaker 2: also memory names that are driving some upside in those
[23:12] Speaker 2: public markets since Moonshot.
[23:15] Speaker 3: Released' Kimmi K three model last week.
[23:16] Speaker 2: Everyone's debating what it really means. Does more efficient AI
[23:20] Speaker 2: mean less hardware? Is China closing the gap faster than
[23:24] Speaker 2: we expected? Where does AI's real value ultimately accrue? Joining
[23:29] Speaker 2: us now is Crosslink Capital partner Mary donofrio focused on
[23:33] Speaker 2: venture growth investments, but across software, AI and infrastructure. And
[23:39] Speaker 2: for the last three days, I've said Kimmy K three
[23:41] Speaker 2: many times, but it was interesting to take a sort
[23:44] Speaker 2: of straw pole of VC's founders, people in industry of
[23:49] Speaker 2: how they react to it so differently. Some people are
[23:52] Speaker 2: very fearful, defensive. Others say this is very good for
[23:56] Speaker 2: AI at a whole, very difficult job, first thing on a Monday.
[23:59] Speaker 3: But your reaction to what happened.
[24:01] Speaker 12: Well, Kimmy K three, as you said, was phenomenal. And
[24:04] Speaker 12: it's similar to the deep seek moment we had last year.
[24:07] Speaker 12: You know, two point eight trillion parameter model, and what
[24:11] Speaker 12: it demonstrates, I think is that you know, China and
[24:14] Speaker 12: Frontier pardon me, China and open weight models are closing
[24:17] Speaker 12: the gap with frontier models. You know, obviously it's done,
[24:20] Speaker 12: it was done based on their own benchmarking. But according
[24:23] Speaker 12: to the benchmarks. You know they're fourth and while you
[24:27] Speaker 12: know you can debate the merits of those benchmarks themselves,
[24:30] Speaker 12: the productivity gains are tremendous, and I think that in
[24:34] Speaker 12: general it is part of the pathway towards more widespread
[24:39] Speaker 12: open weight model usage, and in the context of enterprise
[24:42] Speaker 12: budgets ballooning in AI, I actually think that it's probably
[24:45] Speaker 12: a boon for AI ubiquity over time.
[24:47] Speaker 2: When you think about your your existing portfolio of companies,
[24:52] Speaker 2: were you sort of like, Okay, how does this impact
[24:54] Speaker 2: the trajectory? Is anything actually fundamentally change for the different
[24:59] Speaker 2: piece is the AI five layer cake that you're invested in.
[25:03] Speaker 12: I don't know if it changes the cake itself, but
[25:07] Speaker 12: I do think it's another demonstration that models keep leap
[25:09] Speaker 12: frogging one another, and if companies don't out innovate, they're
[25:13] Speaker 12: going to be leap frog two. Whether it's you know,
[25:16] Speaker 12: every single model seems to outperform the last, and it's
[25:19] Speaker 12: a matter of out innovating by creating modes of data,
[25:23] Speaker 12: of performance, of customer resonance and otherwise. Unfortunately, you know,
[25:30] Speaker 12: the innovations in both frontier and open weight models is tremendous.
[25:35] Speaker 2: You're on the crossover investment team, but basically focused on
[25:38] Speaker 2: private and maybe later stage or venture growth investments. Seven
[25:43] Speaker 2: years at Bessemer as well prior to being at cross
[25:45] Speaker 2: Link sort of introduce us to your overall investment focus,
[25:49] Speaker 2: your investment thesis.
[25:51] Speaker 3: Yeah, of course.
[25:51] Speaker 12: So, as you mentioned, I'm a partner Crosslink Capital. I
[25:55] Speaker 12: helped to lead the mid stage venture practice. I spend
[25:58] Speaker 12: an asymmetric amount of time in vertical AI and AI infrastructure.
[26:02] Speaker 2: What are some examples of what you call vertical AI
[26:05] Speaker 2: and AI infrastructure?
[26:06] Speaker 12: Yeah, of course, So AI infrastructure is a place I
[26:08] Speaker 12: spend a lot of time. That's you know, the physical
[26:10] Speaker 12: and software systems required to build, train and run models.
[26:14] Speaker 12: And in the first wave it was about how to
[26:16] Speaker 12: build a scalable model, and now we have those as
[26:19] Speaker 12: we're talking about right now, and so in the next wave,
[26:22] Speaker 12: it's how to deploy them quickly, fastly, cheaply agents of
[26:26] Speaker 12: course being one of one of the ways to do
[26:28] Speaker 12: so in an efficient way. You know, taking models is
[26:31] Speaker 12: the reasoning layer, adding tools in infrastructure around it to
[26:35] Speaker 12: connect outside data, data and apps, and using memory in
[26:39] Speaker 12: order to do it in a way that produces outcomes,
[26:43] Speaker 12: and agent infrastructure itself has become a sector that I've
[26:47] Speaker 12: spent a lot of time in, whether it's agent harnesses
[26:50] Speaker 12: or the identity layer around it. One of my portfolio companies, Teleport,
[26:53] Speaker 12: is a leader in that space, and it's a place
[26:55] Speaker 12: that I'm spending an asymmetric amount of time.
[26:57] Speaker 2: I wanted to get to that kind of back ground
[27:00] Speaker 2: on your focus to try and answer some of the
[27:02] Speaker 2: questions that the Communicay three moment posed.
[27:04] Speaker 3: One idea is that.
[27:07] Speaker 2: Basically, by lowering the margins at the model layer, it's
[27:11] Speaker 2: an enabler, it opens up the market a little bit,
[27:14] Speaker 2: it's a driver of demand essentially.
[27:17] Speaker 3: Do you agree with that kind of thinking In response.
[27:20] Speaker 12: To this, I do think that it demonstrates what you
[27:23] Speaker 12: know a lot of even the inference infrastructure layer is demonstrating,
[27:27] Speaker 12: which is, you know, as cost has gone down by
[27:29] Speaker 12: ninety five percent, does that displace the total dollar value
[27:34] Speaker 12: expended by customers? And I would argue it probably won't,
[27:37] Speaker 12: because the migration will just be from workloads that are
[27:41] Speaker 12: at the frontier, that are most expensive and of course
[27:43] Speaker 12: that require the most reasoning capabilities and are the most advanced.
[27:47] Speaker 12: But they'll go to cheaper models that you know, still
[27:52] Speaker 12: provide the same amount the requisite amount of compute for
[27:56] Speaker 12: the question being posed, but do so at much cheaper
[28:00] Speaker 12: while not sacrificing on things like latenccene performance.
[28:03] Speaker 2: It's interesting because like kim K three kind of pricing
[28:08] Speaker 2: three dollar per input fifteen dollars per output on the
[28:11] Speaker 2: token side, Like it's it's quite in line with what
[28:16] Speaker 2: a lot of leading models are also priced. That I
[28:18] Speaker 2: found that interesting. Maybe we should talk a little bit
[28:20] Speaker 2: about valuations. It's not a sophisticated argument, but when this
[28:24] Speaker 2: will happened, a lot of people came out and said, well,
[28:27] Speaker 2: if a twenty billion dollar Chinese startup can do a
[28:31] Speaker 2: two point eight trillion parameter model with these economics, why
[28:35] Speaker 2: are we signing open AI and anthropic evaluation of a trillion?
[28:38] Speaker 2: You know, they looked at the sort of fundamentals of that.
[28:42] Speaker 2: Is that the right way to think about it?
[28:45] Speaker 12: Frankly, I don't know yet, and I don't think anybody
[28:48] Speaker 12: knows yet, because even in the context of those questions,
[28:52] Speaker 12: Anthropic and open AI are still producing billions and billions
[28:55] Speaker 12: of revenue. And it's not just from their you know,
[28:58] Speaker 12: they have first party businesses, they have third party businesses,
[29:01] Speaker 12: and I think you know, obviously the API opened for
[29:06] Speaker 12: a moonshot, but the open weights have not yet.
[29:08] Speaker 3: Yeah, twenty seventh of July.
[29:09] Speaker 12: And it's worth noting that because it's open weight doesn't
[29:12] Speaker 12: mean it's free to run. And so I think that's
[29:15] Speaker 12: a dynamic to these models that people potentially underappreciate as well.
[29:18] Speaker 2: Say the argument you just outlined, the same that graylocks
[29:22] Speaker 2: on mod of Many made on Friday on the show,
[29:24] Speaker 2: that you know, these are very robust businesses that have
[29:27] Speaker 2: multiple revenue streams. You are focused on the private markets,
[29:32] Speaker 2: but we've had some really interesting IPOs of late, you know,
[29:36] Speaker 2: not just SpaceX, but also s k Heinez doing its ADRs.
[29:39] Speaker 3: I was in New York City for that. What does
[29:41] Speaker 3: that signal to you? I think it's two things.
[29:45] Speaker 12: Obviously, those are fantastic businesses, and I think when it
[29:48] Speaker 12: comes to what it says for the markets is that
[29:51] Speaker 12: investors are really looking for anything having to do with
[29:53] Speaker 12: the AI build out AI infrastructure. Obviously you made the
[29:56] Speaker 12: reference to the performance of public market stocks today. Also
[30:00] Speaker 12: think it's one other thing for the capital markets, which
[30:02] Speaker 12: is that you know, we have this narrative, but potentially
[30:05] Speaker 12: the IPO window is reopening. But it's doing so with
[30:07] Speaker 12: these long standing businesses with billions of dollars of revenue
[30:11] Speaker 12: that are or frankly older. You know, they're at the
[30:14] Speaker 12: cutting edge of innovation, but they've been around for decades,
[30:17] Speaker 12: and so I think it's saying, you know, there is
[30:20] Speaker 12: a desire for things having to do with AI, but nonetheless,
[30:24] Speaker 12: investors are still they're hardly speculative. Investors are looking for
[30:27] Speaker 12: things that are relatively secure.
[30:29] Speaker 2: Cross Link partner married and for making the debut on
[30:32] Speaker 2: Bloomberg Techs.
[30:33] Speaker 3: Really great to have you here, and Alisa, thank you
[30:34] Speaker 3: very much.
[30:35] Speaker 2: TSMC is spending at an unprecedented pace, committing a record
[30:40] Speaker 2: two hundred and sixty five billion dollars to its US expansion.
[30:43] Speaker 2: So with AI fueling enormous capital needs across the chip industry,
[30:47] Speaker 2: the question becomes, how will it pay for it? Bloomberg
[30:50] Speaker 2: Stephen Engel our CFO Wendel Wang whether another trip to
[30:54] Speaker 2: the bond market could be on the table.
[30:57] Speaker 13: Our fundamental belief is that WEISH should be able to
[31:00] Speaker 13: self sustain using our operating cash row to support the
[31:05] Speaker 13: KPEX and leave enough free cash flow to return cash
[31:10] Speaker 13: to shareholder through a sustainable and steadily increasing cash dividends,
[31:16] Speaker 13: which we have been doing that in the past several years.
[31:19] Speaker 13: We have very good track record of it. Now from
[31:23] Speaker 13: time to time we can go out to the market,
[31:26] Speaker 13: to the bone market. If the timing is right. We've
[31:30] Speaker 13: been doing it. We've done thirty billion dollars fundraising in
[31:36] Speaker 13: bond market a few years back, about half of it
[31:39] Speaker 13: in US dollar, half of it in NT. NT is very.
[31:47] Speaker 3: Cheap.
[31:47] Speaker 13: Funding for US US dollar is becoming more and more expensive,
[31:52] Speaker 13: so we'll see if the timing is correct. If the
[31:56] Speaker 13: timing is right, we do not rule out that we
[31:59] Speaker 13: can go to the bomb market again.
[32:01] Speaker 3: Well, given what you just laid out, would maybe a
[32:04] Speaker 3: mix of more empty dollar denominated bonds be more likely
[32:09] Speaker 3: or yields to a high.
[32:10] Speaker 13: ANTY dollar bomb market is much shallower than the US
[32:15] Speaker 13: dollar bomb market, so it has its limitations.
[32:20] Speaker 2: That was tsmc CFO Wendl Juan with Bloomberg Stephen Engel
[32:24] Speaker 2: in Taipei. Now coming up, Silicon Valley is still betting
[32:26] Speaker 2: big on defense. We're gonna speak with Andres and Horowitz
[32:30] Speaker 2: new signing Connor Love about the startup shaping the future
[32:34] Speaker 2: of national security kind of just joined the American Dynamism.
[32:37] Speaker 3: Team that's next. This is Bloomberg Tech.
[32:45] Speaker 2: Venture firm and recent Horowitz is doubling down on its
[32:48] Speaker 2: American Dynamism thesis and practice, adding defense and space investor
[32:52] Speaker 2: and former Army captain Connor Love to the team and
[32:55] Speaker 2: Connor joins us now joining from light Speed. It's great
[32:58] Speaker 2: to have you on the show day one. You're very
[33:01] Speaker 2: welcome Day one on American Dynamism. You, you know, reading
[33:06] Speaker 2: about your investments at the last firm and.
[33:11] Speaker 3: Like how you've approached your career.
[33:13] Speaker 2: You would call it American exceptionalism, right, but there's a
[33:17] Speaker 2: close alignment between those two ideas. Just reflect on what
[33:20] Speaker 2: it was that prompted you to join this firm.
[33:23] Speaker 14: Yeah, well, first of all, thanks for having me. We're
[33:27] Speaker 14: in a pretty amazing opportunity right now to invest really
[33:30] Speaker 14: in the fabric of America. And when Mark and Ben
[33:33] Speaker 14: and team kind of approached me, look, this is a
[33:35] Speaker 14: category that four years ago, they didn't jump on the bandwagon.
[33:39] Speaker 14: They said, hey, this is important to the fabric of America.
[33:41] Speaker 14: This is important for the entrepreneurs and the technology future
[33:44] Speaker 14: for our nation. And when given that opportunity. You know,
[33:48] Speaker 14: you kind of get on you have to get on
[33:50] Speaker 14: board and join the team. So it's been amazing, you know,
[33:53] Speaker 14: you know so far just to you know, dive in
[33:54] Speaker 14: and get to know everyone. I think in the end,
[33:56] Speaker 14: when you take a step back and just look at
[33:57] Speaker 14: the opportunity here. You know, as I said before, what's
[34:01] Speaker 14: happening is a once in a generation opportunity where we're
[34:03] Speaker 14: redefining what it really means to build in America. It's
[34:06] Speaker 14: not just software primitives. I mean, it's it's things that
[34:09] Speaker 14: are core to the fabric of America. It's defense technology,
[34:12] Speaker 14: it's energy, it's many things. And so to join a
[34:15] Speaker 14: team like Indres and Horowitz and kind of continue to
[34:18] Speaker 14: invest in that back entrepreneurs in this ecosystem, I'm just
[34:21] Speaker 14: really thrilled and excited to be.
[34:22] Speaker 3: Able to do it conin.
[34:24] Speaker 2: Do you mind if I ask a bit more about
[34:25] Speaker 2: the backstory. You know, you mentioned Mark and Drees and
[34:28] Speaker 2: Ben Horowitz that they approached you.
[34:31] Speaker 3: You know, how did this all come together.
[34:33] Speaker 14: It's actually funny. One of the partners in, you know,
[34:35] Speaker 14: founders of American Dynamism, Catherine Boyle. I actually spoke to
[34:39] Speaker 14: her about seven years ago, I had the privilege of
[34:42] Speaker 14: serving in the military and was actually for deployed in Iraq,
[34:44] Speaker 14: you know, looking for what the next opportunity was in
[34:47] Speaker 14: my professional career. And I connected with her really and
[34:50] Speaker 14: just kind of a you know, haphazard out of the
[34:52] Speaker 14: blue way, and she said, Hey, there's this company called
[34:54] Speaker 14: Androl you should look at andrel Obviously everyone knows what
[34:57] Speaker 14: Andreil is these days, but I think importantly I just
[35:00] Speaker 14: got exposed to Silicon Valley and I saw then what
[35:04] Speaker 14: I know now that really technology will be a force
[35:07] Speaker 14: for good, not just for the defense fabric of our
[35:10] Speaker 14: nation as I mentioned, but also just you can't have
[35:13] Speaker 14: an industrial base in America if you don't have a
[35:15] Speaker 14: vision in defense base. So it's to me a once
[35:18] Speaker 14: in a lifetime opportunity that is very much rooted in
[35:20] Speaker 14: my background and just happens to connect at you know,
[35:24] Speaker 14: three in the morning with Catherine Boyle at you know,
[35:27] Speaker 14: when she was then at her previous firm.
[35:30] Speaker 2: You've explained it right that it's not exclusively defense technology.
[35:34] Speaker 3: American dynamism is broad in that bit right now.
[35:37] Speaker 2: Anderil is the best known case study, yeah, right, seronic.
[35:41] Speaker 2: More recently in the news cycle, what is the next
[35:44] Speaker 2: phase of American dynamism, do you think Conna.
[35:47] Speaker 14: Yeah, well I put it in that lens, which is
[35:49] Speaker 14: you really cannot have a defense industrial base without an
[35:53] Speaker 14: industrial base. And when you look into what goes into
[35:56] Speaker 14: these autonomous systems, these new platforms, it's really the subcomet,
[36:00] Speaker 14: it's really the parts, it's really the you know, metal
[36:02] Speaker 14: and the end components of what's going on. So when
[36:05] Speaker 14: I think, you take a step back and you look
[36:06] Speaker 14: at hey, obviously defense, we're creating a new opportunity for
[36:10] Speaker 14: neoprimes to really capture and serve not just the United
[36:13] Speaker 14: States but also our allies. You see things like energy,
[36:17] Speaker 14: you see things like core manufacturing. Importantly, you see things
[36:21] Speaker 14: like space. Yes, there's a huge defense component to space,
[36:23] Speaker 14: but as we put more kilograms of mass up to
[36:26] Speaker 14: space in a lot cheaper way, you know, really it is,
[36:29] Speaker 14: you know, kind of the next frontier of where we're
[36:30] Speaker 14: going to be building not just for defense and kind
[36:32] Speaker 14: of government use cases, but all other parts of the economy.
[36:36] Speaker 14: So I think when I take a step back and
[36:38] Speaker 14: I look at American dynamism as we call it now,
[36:41] Speaker 14: I actually just viewed as it's a proxy for how
[36:43] Speaker 14: some of the most valuable companies in the world are created.
[36:46] Speaker 14: I mean, historically you look at Boeing, and historically you
[36:49] Speaker 14: look at you know, all these these companies that have
[36:50] Speaker 14: been important to America. They've been very important to you know,
[36:54] Speaker 14: the American individual, they necessarily have not always been the
[36:57] Speaker 14: most valuable companies. Technology is changing that. Technology is saying
[37:01] Speaker 14: you can be both important to America and our allies
[37:04] Speaker 14: and be some of the most valuable companies in the world.
[37:06] Speaker 14: And that's what we and myself at Andrews and Horwitz
[37:09] Speaker 14: are really excited to invest into.
[37:11] Speaker 2: And Reil is a big focus for the Bloomberg Tech audience.
[37:14] Speaker 2: You would have seen the news this morning with Archer
[37:17] Speaker 2: finally showing the autonomous Vitol platform. I think what'd be
[37:21] Speaker 2: interesting is to you know, you participated in light speeds
[37:25] Speaker 2: and aero investment and a real clearly important part of
[37:28] Speaker 2: the American dynamism in.
[37:30] Speaker 3: A sixteen team.
[37:31] Speaker 2: More broadly, portfolio, how are you going to approach that?
[37:35] Speaker 14: Yeah, well, I think I'll say a couple of things
[37:38] Speaker 14: about Andreil. The first thing is when you think they've
[37:41] Speaker 14: got to their last product, you're wrong. There's another product.
[37:44] Speaker 14: And I think the beauty is when you have such
[37:47] Speaker 14: a complex use case that is the kind of you know,
[37:50] Speaker 14: Department of War and the American warfighter. There are a
[37:53] Speaker 14: lot of problems that need to be solved. So you know,
[37:55] Speaker 14: to see this kind of partnership with Archer, I think
[37:58] Speaker 14: is only not good for the American warfighter and the
[38:00] Speaker 14: American public. I think when you take a step back
[38:02] Speaker 14: and say, you know, really this market is massive. I
[38:05] Speaker 14: mean there are a lot of problems to be solved
[38:07] Speaker 14: in the defense space. I mean there's everything's from missiles
[38:09] Speaker 14: and what Castilian's doing to ground autonomy systems. I mean
[38:13] Speaker 14: it is really a massively large market where there's actually
[38:16] Speaker 14: a ton of technology problems to be solved below it.
[38:19] Speaker 14: I think the underlying substrate that I think is important
[38:23] Speaker 14: and what we look at, Andresa and I've looked at
[38:24] Speaker 14: in the past is really what's happening in defense is
[38:27] Speaker 14: this low cost autonomy thesis that is just you know,
[38:31] Speaker 14: it's playing out in the battlefield. Go look at what
[38:33] Speaker 14: Saronic's done in kind of the conflict in Iran, Go
[38:36] Speaker 14: look at the front battlefield of Ukraine. When you can
[38:40] Speaker 14: introduce something that is literally ten times cheaper and as
[38:44] Speaker 14: performance and in mass it really changes the game in
[38:47] Speaker 14: terms of deterrence and what you can do for our nation.
[38:49] Speaker 14: So I think, again, you take a step back. The
[38:52] Speaker 14: market's big, and I think there's going to be a
[38:53] Speaker 14: ton of opportunity.
[38:55] Speaker 3: Kind of really quick.
[38:56] Speaker 2: You know, you mentioned Cafrine and airin regular contributors to
[38:59] Speaker 2: the show, but with you, somebody joined the team that
[39:02] Speaker 2: is a veteran just in the thirty seconds we have.
[39:04] Speaker 2: Would you reflect on that a little bit, the skill
[39:07] Speaker 2: set and experience you bring.
[39:08] Speaker 14: Yeah, I mean there are more impressive people that have
[39:11] Speaker 14: done some pretty amazing things for our country that I
[39:13] Speaker 14: want to put in front of me. I do think, though,
[39:16] Speaker 14: that time of service really just grounds me and why
[39:19] Speaker 14: we are doing this. Like I said at the start, importantly,
[39:23] Speaker 14: this is a category where you're not just doing good economically,
[39:26] Speaker 14: but I do believe you're doing good for the world
[39:28] Speaker 14: and for the American public. So having worn that uniform
[39:30] Speaker 14: puts me in a position just to take a step
[39:32] Speaker 14: back and say, hey, this means a lot, not just
[39:35] Speaker 14: for our shareholders and for everyone you know backing in dreason,
[39:38] Speaker 14: but also for America. So I'm deeply thankful and I'm
[39:41] Speaker 14: going to keep doing it for decades.
[39:43] Speaker 2: Connor Love now on the American Dynamison team at A
[39:46] Speaker 2: sixteen S great to have you on the show.
[39:48] Speaker 3: Thank thank you very much.
[39:49] Speaker 2: Indeed, coming up Alphabet, Tesla, Intel and more. Posting earnings.
[39:54] Speaker 2: This week, we're going to look ahead. We're going to
[39:56] Speaker 2: discuss what to expect. This is Bloomberg Tech Okay, tech
[40:07] Speaker 2: Earning's kicking off and we're expecting some big names this week,
[40:10] Speaker 2: including Alphabet, Tesla and Intel. Investors will be looking for
[40:15] Speaker 2: proof that billions in AI spending are starting to pay off,
[40:19] Speaker 2: even as cheap AI models from China have raised some questions.
[40:22] Speaker 2: Bloomberg Intelligence Senior analyst man Deep Singh joins us on
[40:26] Speaker 2: what to expect, and you know, it's been so interesting
[40:29] Speaker 2: to read the team's research.
[40:31] Speaker 3: There will be short.
[40:32] Speaker 2: Term a focus on CAPEX, but I think looking past that,
[40:36] Speaker 2: what is it that you want to see? What's the
[40:38] Speaker 2: data point that shows this AI things real?
[40:41] Speaker 15: I mean, cloud revenue is still the key metric when
[40:44] Speaker 15: it comes to hyperscalers who are at the forefront of kapex,
[40:48] Speaker 15: and everyone is expecting those CAPEX numbers to go up.
[40:52] Speaker 15: I mean, the question is is next year going to
[40:54] Speaker 15: be thirty to forty percent, or we are going to
[40:58] Speaker 15: have another year of fifty percent plus increase in capex.
[41:01] Speaker 15: I mean, to my mind, the numbers are getting very
[41:04] Speaker 15: big in terms of trillion dollars in analyzed capex from
[41:08] Speaker 15: the hyperscaler. So there is a limit to how far
[41:12] Speaker 15: they can go. But clearly you know twenty twenty seven
[41:15] Speaker 15: numbers will be higher and we are expecting positive upward
[41:19] Speaker 15: revisions justified by accelerating cloud growth from these hyperscalers.
[41:26] Speaker 2: Let's stick with the Google parent is the case study.
[41:28] Speaker 2: You know, overall revenue growth is expected to be about
[41:31] Speaker 2: twenty percent, but Google Cloud I think sixty five percent.
[41:34] Speaker 3: You're only a growth you know, that's.
[41:36] Speaker 2: The story, right that Google Cloud is where the traction
[41:40] Speaker 2: with AI shows up.
[41:43] Speaker 15: Yeah, no, you're absolutely right, And that's where you know,
[41:46] Speaker 15: the comparisons would be drawn to Azure and the fact
[41:51] Speaker 15: that Google Cloud is growing at least fifteen percentage faster
[41:55] Speaker 15: points faster than Azure and AWLUSS. I mean, seems to
[42:00] Speaker 15: be things seem to be picking up. And now OpenAI
[42:03] Speaker 15: is deploying on Amazon Ablus as well as Azure. So
[42:08] Speaker 15: the impact of that on Microsoft that's going to be
[42:11] Speaker 15: a front end center, but there is no doubt that
[42:14] Speaker 15: cloud infrastructure is where the monetization will be evident. And look,
[42:18] Speaker 15: all the frontier labs are monetizing tokens, but you're not
[42:22] Speaker 15: going to be able to glean that from the earnings
[42:25] Speaker 15: of companies. It's too early to see that token monetization
[42:29] Speaker 15: being visible.
[42:31] Speaker 3: This is what the week looks like. All kicks off Wednesday.
[42:34] Speaker 2: Alphabet and tesla IBM of course pre released. But maybe
[42:38] Speaker 2: we'll learn a bit more than Intel on Thursday. Bloomberg
[42:40] Speaker 2: intelligencing around this, Mandip Singh, thank you very much. Indeed,
[42:44] Speaker 2: a busy week form mendy two. As we head out,
[42:47] Speaker 2: let's take a look at today's big number. Two hundred
[42:50] Speaker 2: and sixty four point one million dollars. That's how much
[42:53] Speaker 2: The Odyssey, the new movie from Blockbuster director Christopher Nolan,
[42:57] Speaker 2: took in sales globally opening weekend. It was the number
[43:01] Speaker 2: one film in the US and Canada over the weekend,
[43:05] Speaker 2: with one hundred and twenty four point five million dollars
[43:07] Speaker 2: in ticket sales in North America, a large portion of
[43:10] Speaker 2: which was generated from screenings in IMAX's giant auditoriums, many
[43:16] Speaker 2: of them were sold out many months in advance. I'm
[43:21] Speaker 2: still yet to see it. That does it for this
[43:25] Speaker 2: edition of Bloomberg Tech. A really big focus still on
[43:29] Speaker 2: Moonshot's Kimmy K three model two point eight trillion parameters,
[43:33] Speaker 2: and a lot of questions posed for the world for AI,
[43:36] Speaker 2: for markets for investors. We got through all of that
[43:39] Speaker 2: in the program, so recap it on the podcast. You
[43:40] Speaker 2: know where to find it on the Bloomberg terminal. As
[43:43] Speaker 2: one audience member reminded me, we call it the Blue Meet,
[43:46] Speaker 2: but it's online too.
[43:47] Speaker 3: Have a great week. This is Bloomberg Tech.

Shows are listed from Apple Podcasts' charts and directory, and Podcast Index where it is used. Every show and episode belongs to its publisher and plays from the publisher's own feed; this site hosts none of them.
