# AI Boom Faces New Tests From Chips to Immigration

An episode of [Bloomberg Tech](https://invisiblenetwork.ai/podcast/1611767434), 8 Oct 2026, 40 min.

Audio: https://podtrac.com/pts/redirect.mp3/traffic.omny.fm/d/clips/e73c998e-6e60-432f-8610-ae210140c5b1/41764a4f-fc64-4e11-89ba-ae7c0030ab5e/babb1df5-1b0f-484f-ba3d-b4dd0145018b/audio.mp3?utm_source=Podcast&in_playlist=9caafc41-289c-4115-995d-ae7c0030ab75

Bloomberg’s Ed Ludlow breaks down the Trump administration's decision to suspend an immigration program for multiple tech firms, including Microsoft, alleging widespread abuse of that worker visa system. Plus: Samsung's quarterly profit jumps and TSMC sales surge, but investors are still worried about the staying power of the AI investment boom. And, NASA's SpaceX Crew-12 makes it way back down to Earth after 237 days in space. See omnystudio.com/listener for privacy information.

## Transcript

[0:05] Speaker 1: Bloomberg Tech is live from the heart of Silicon Valley
[0:09] Speaker 1: with Ed Ludlow in San Francisco.
[0:14] Speaker 2: This is Bloomberg Tech coming up.
[0:16] Speaker 3: The Trump administration says it's suspending an immigration program for
[0:20] Speaker 3: multiple tech firms, including Microsoft, alleging widespread abuse of a
[0:24] Speaker 3: worker visa system. Plus, Samsung's quarterly profit jumps, TSMC sales surge,
[0:30] Speaker 3: but investors are still worried about the staying power of
[0:32] Speaker 3: the AI investment boom. And NASA's SpaceX Crew 12 makes
[0:37] Speaker 3: its way back down to Earth after 237 days in space.
[0:41] Speaker 3: We'll bring you live coverage of the splashdown. Breaking news
[0:45] Speaker 3: just before we came to air.
[0:46] Speaker 2: The U.S.
[0:46] Speaker 3: Is suspending an immigration program for multiple tech firms. This
[0:50] Speaker 3: after Vice President J.D. Vance called out Microsoft, saying it
[0:55] Speaker 3: had abused the system. Bloomberg's Kayleigh Lyons joins us with
[0:57] Speaker 3: the latest.
[0:58] Speaker 2: Kayleigh.
[1:00] Speaker 4: Yeah, Ed, this news came from a fraud and protecting
[1:03] Speaker 4: American workers event that the vice president, the secretary of labor,
[1:06] Speaker 4: and the attorney general are all taking part in this morning.
[1:09] Speaker 4: And it pertains to the permanent labor certification program. Basically,
[1:13] Speaker 4: this is a program managed by the Department of Labor
[1:16] Speaker 4: that does allow firms in the U.S. to give permanent
[1:18] Speaker 4: residency to employees who if they can prove that they
[1:21] Speaker 4: qualify for the program because there are not enough available
[1:25] Speaker 4: and qualified workers in the United States for the jobs
[1:28] Speaker 4: in question. The Labor Secretary, Keith Sunderling, who was just
[1:31] Speaker 4: approved to be the Labor Secretary by the Senate at
[1:34] Speaker 4: the end of last month, said this applies to a
[1:36] Speaker 4: number of firms. After investigations, Adobe and Microsoft will see
[1:41] Speaker 4: their access to this program suspended, as well as Cognizant, Infosys, TAT, Wipro, HCL.
[1:46] Speaker 4: And Capgemini. And as you mentioned, the vice president, J.D. Bance,
[1:49] Speaker 4: he did say that Microsoft has been the biggest abuser
[1:52] Speaker 4: of this. The administration pointing to figures, including Microsoft, laying
[1:55] Speaker 4: off 6,000 workers in the U.S.
[1:58] Speaker 5: Last year.
[1:58] Speaker 4: At the same time, they applied for 6,000 H-1B visas
[2:00] Speaker 4: and about 3,700 applications to this permanence program. Vance says
[2:06] Speaker 4: that the Microsoft suspension to that program will last as
[2:09] Speaker 4: long as needed. And he said his message to Adobe
[2:12] Speaker 4: and the other firms is to stop defrauding U.S.
[2:15] Speaker 2: Workers.
[2:15] Speaker 4: We heard from the Attorney General Todd Blanch on this
[2:18] Speaker 4: as well. I'll quote him directly. He says the announcement
[2:20] Speaker 4: today should send and does send a strong signal. to
[2:23] Speaker 4: every single employer in the U.S. that the Trump administration
[2:26] Speaker 4: will not tolerate discrimination against American workers. I would point
[2:30] Speaker 4: out as well that the action on visas here is
[2:32] Speaker 4: extending to universities. The Trump administration is probing alleged J-1
[2:37] Speaker 4: visa issues at nine American universities, including Harvard, Yale, MIT, Stanford, Brown, Caltech,
[2:44] Speaker 4: and Arizona State. Those J-1 visas apply to things like researchers.
[2:48] Speaker 4: And professors who, through certain programs, can come to the
[2:51] Speaker 4: United States. But it looks like they're probing that and, again,
[2:53] Speaker 4: cracking down on the permanent labor certification program for companies,
[2:56] Speaker 4: including Microsoft Edge.
[2:58] Speaker 3: Okay, we're keeping an eye on ongoing headlines and a
[3:00] Speaker 3: lot of the names that you mentioned in the tech
[3:02] Speaker 3: sector have either paired their gains or gone into negative territory.
[3:05] Speaker 3: Bloomberg's Kayleigh Lyons, thank you very much indeed. This all
[3:08] Speaker 3: comes as Nvidia says it will commit a billion dollars
[3:11] Speaker 3: to advance US science over the next five years. And
[3:15] Speaker 3: as President Trump is putting AI at the center of
[3:17] Speaker 3: his push. for what he calls a new golden age
[3:20] Speaker 3: of American innovation. Later today, the White House will be
[3:23] Speaker 3: holding a summit where the president will award the National
[3:26] Speaker 3: Medal of Science to SpaceX CEO Elon Musk, Google co-founder
[3:30] Speaker 3: Sergey Brin, NVIDIA CEO Jensen Wang, and AMD CEO Lisa Su.
[3:35] Speaker 3: He will also award the National Medal of Technology and
[3:38] Speaker 3: Innovation to Dell CEO Michael Dell and Microsoft's And Microsoft
[3:43] Speaker 3: CEO Satya Nadella, Bloomberg's Mike Shepard, senior tech editors with us,
[3:48] Speaker 3: go back to the breaking story and the idea that
[3:52] Speaker 3: later today the president will be with Satya Nadella, the
[3:55] Speaker 3: Microsoft CEO, having just removed them from the PERM program.
[4:01] Speaker 6: It does create a bit of an awkward look there, Ed,
[4:04] Speaker 6: but it is not entirely inconsistent. As part of this
[4:07] Speaker 6: initiative today, this golden age of science, there are going
[4:11] Speaker 6: to be a number of announcements, and you mentioned the
[4:13] Speaker 6: one from NVIDIA, but there's another that our colleague Liam
[4:16] Speaker 6: Knox was first to report yesterday, and that is that
[4:19] Speaker 6: the administration is setting aside $ 100 million in an effort
[4:24] Speaker 6: to help train homegrown American scientists.
[4:27] Speaker 7: There's this concern that there.
[4:28] Speaker 6: Is and excess reliance on foreign talent in the science
[4:33] Speaker 6: talent pool here in the United States at the academic level,
[4:36] Speaker 6: but also at the corporate level. So we're seeing that initiative.
[4:39] Speaker 6: And yet at the same time, this crackdown on visa
[4:42] Speaker 6: abuses that they have are citing on the administration side
[4:45] Speaker 6: by some of these companies, including Microsoft and others. Now,
[4:49] Speaker 6: we have not heard comments yet from the companies. It's
[4:52] Speaker 6: important to note. So we'll wait to hear back to
[4:54] Speaker 6: see what they have to say. But it does make
[4:57] Speaker 6: a an awkward contrast, certainly at the event.
[5:01] Speaker 2: All right, Shep, bear with me.
[5:02] Speaker 3: I appreciate this just broke before we came to air.
[5:04] Speaker 3: We now do have those comments from Vice President J.D.
[5:07] Speaker 7: Vance.
[5:07] Speaker 2: This is what he said.
[5:09] Speaker 5: There has been no company in the United States, unfortunately,
[5:12] Speaker 5: that has abused this system more than Microsoft.
[5:14] Speaker 7: Thanks to an ongoing investigation, we.
[5:16] Speaker 5: Have decided to suspend the PERM program for Microsoft, which
[5:21] Speaker 5: means they will no longer be able to take those
[5:26] Speaker 5: H-1Bs and apply for permanent resident status within the United
[5:30] Speaker 5: States of America.
[5:34] Speaker 3: Shep, that was pretty clear in his explanation, the vice president,
[5:37] Speaker 3: of the action that the government has taken.
[5:39] Speaker 2: Microsoft, one of several companies.
[5:41] Speaker 3: Again, recap it and remind us the timing of this
[5:44] Speaker 3: with the ceremony taking place later today in which Microsoft
[5:48] Speaker 3: CEO Satya Nadella is a recipient of an award from
[5:51] Speaker 3: the president of the United States.
[5:54] Speaker 6: This is all set to take place at 1.30. The
[5:56] Speaker 6: president will be on his way over then to hang
[5:59] Speaker 6: the medals on the necks of these tech leaders, including
[6:04] Speaker 6: Satya Nadella. And, you know, Nadella is no stranger to
[6:07] Speaker 6: the White House. He has been among the many executives
[6:10] Speaker 6: who have visited and met with the president, including just
[6:14] Speaker 6: last week when he held that luncheon to talk about
[6:17] Speaker 6: AI technology. safety risks, safety policy, and the president's desire
[6:21] Speaker 6: to keep developing this technology. So you do see a
[6:25] Speaker 6: little bit of dissonance here, certainly in terms of how
[6:29] Speaker 6: welcome the company is in terms of the administration's view
[6:32] Speaker 6: on its policy. We'll have to see whether the company
[6:35] Speaker 6: will be able to sort of bifurcate its view, separating
[6:39] Speaker 6: the good that would be celebrated later today by the
[6:43] Speaker 6: president with that medal and from the sanction that the
[6:46] Speaker 6: administration is opposing just now, as Vice President J.D. Vance
[6:51] Speaker 6: laid out in pretty stark terms.
[6:54] Speaker 3: Bloomberg's Mike Shepard, Senior Tech Editor in Washington, D.C., with
[6:58] Speaker 3: the breaking news.
[6:58] Speaker 2: You can follow that D.C.
[6:59] Speaker 3: Summit that we discussed on the Bloomberg Terminal, a live
[7:02] Speaker 3: go later today. When all of this was happening just
[7:05] Speaker 3: before Bloomberg Tech kicked off, the Nasdaq 100 sort of
[7:09] Speaker 3: accelerated its decline. You can see sort of leading up
[7:12] Speaker 3: to 11 a.m.
[7:12] Speaker 2: Eastern, 8 a.m.
[7:13] Speaker 3: Pacific, we're now down half a percentage point on the
[7:16] Speaker 3: Nasdaq 100. But the sort of macro point is is
[7:19] Speaker 3: that stocks, technology stocks in particular, have been at record
[7:21] Speaker 3: highs and taken a breather in the last 24 hours
[7:24] Speaker 3: or so. Joining us now is Matt Weir, Goldman Sachs
[7:27] Speaker 3: Wealth Management Co-Head of the Investment Strategy Group. And actually,
[7:31] Speaker 3: coming into this morning, my question for you is whether
[7:33] Speaker 3: we've hit a market top or not. And there were
[7:35] Speaker 3: lots of, in the technology sector in particular, indicators that
[7:38] Speaker 3: you could argue either way where you guys sit right now.
[7:41] Speaker 7: Yeah, thank you very much for having me, Ed.
[7:42] Speaker 8: I think a key focus over the last few weeks
[7:45] Speaker 8: has been a very small number of large tech stocks
[7:49] Speaker 8: have been driving this market higher.
[7:51] Speaker 7: To a new all-time high.
[7:53] Speaker 8: That is not necessarily the ideal setup, but we are
[7:57] Speaker 8: of the view that that alone by no means suggests
[8:00] Speaker 8: that this is a market peak or that we've seen
[8:03] Speaker 8: the all-time high this cycle.
[8:05] Speaker 7: I think a couple of things we should be pointing out.
[8:06] Speaker 8: The first is, It's been a very short period when
[8:09] Speaker 8: you've seen this divergence between, for example, the Magnificent Seven
[8:13] Speaker 8: and the rest of the 493 stocks in the S &
[8:15] Speaker 8: P 500. We've also seen divergences like this in 2023, 2024.
[8:21] Speaker 8: That didn't derail this bull market we've been on. And
[8:24] Speaker 8: then I think the most interesting thing is if you
[8:27] Speaker 8: zoom out and take a longer term view, so year
[8:29] Speaker 8: to date, the S & P 500 is up. The
[8:34] Speaker 8: Magnificent 7 is up by about the same magnitude, 15%,
[8:38] Speaker 8: as the S & P 493 stocks outside. So the
[8:42] Speaker 8: two are actually performing quite comparably.
[8:44] Speaker 3: Yeah, and you would say, well, go back three years.
[8:47] Speaker 3: Go back to 2023.
[8:48] Speaker 2: I mean, you say year-to-date. I mean, it's October already.
[8:51] Speaker 3: But that's an interesting distinction. A lot of people post
[8:56] Speaker 3: questions to me like, Are we in a cycle? Like,
[8:58] Speaker 3: for example, in memory, is there actually a cycle? Where
[9:01] Speaker 3: are we in it? But more broadly for the technology sector,
[9:05] Speaker 3: a cycle where the conversation is distorted by sort of
[9:08] Speaker 3: line goes up in AI investment terms.
[9:11] Speaker 7: Yeah.
[9:11] Speaker 8: So we think in the third quarter, we did see
[9:14] Speaker 8: the peak growth rate.
[9:15] Speaker 7: In hyperscaler CapEx, which is driving.
[9:18] Speaker 2: That's the key number.
[9:19] Speaker 7: Exactly. Exactly. Now, that sounds like bad news.
[9:23] Speaker 8: The rate of growth at 116% in year-over-year terms is
[9:26] Speaker 8: unsustainably high. We think that that rate of growth is
[9:30] Speaker 8: going to come down such that at the end of
[9:32] Speaker 8: next year, fourth quarter 2027, the growth rate we think
[9:36] Speaker 8: will be 22%. If you view that in the context
[9:39] Speaker 8: of nominal GDP growth in the U.S. of 4% to 5%,
[9:42] Speaker 8: even at that growth rate in the fourth quarter next year,
[9:45] Speaker 8: We're still seeing CapEx growth four to five times higher
[9:49] Speaker 8: than nominal GDP growth. So nothing to sneeze at. It's
[9:52] Speaker 8: just a deceleration of a very fast growth rate.
[9:55] Speaker 3: You know, higher rates discount the present value of future
[9:58] Speaker 3: cash flows, like go by the book when it comes
[10:00] Speaker 3: to the tech sector. But like, you know, valuations have
[10:02] Speaker 3: been pretty steady. You know, now we confront earnings season
[10:06] Speaker 3: at the same time as a midterms election, you know,
[10:09] Speaker 3: which sort of headwind, if we call it headwind, wins
[10:12] Speaker 3: out the investor community.
[10:14] Speaker 8: It's so interesting when you see macroeconomic events like a
[10:18] Speaker 8: midterm election, At the same time as an earnings season,
[10:21] Speaker 8: in some quarterly earnings season, you see the market distracted
[10:25] Speaker 8: by what's happening at the macro level.
[10:27] Speaker 7: Yes.
[10:27] Speaker 8: And not necessarily focused on what's happening at the earnings level,
[10:30] Speaker 8: which we think is a mistake. Earnings ultimately determine the
[10:33] Speaker 8: path of the market. This could be a quarter like that.
[10:36] Speaker 8: But what we do, though, see, Ed, is. Even if
[10:38] Speaker 8: there's an economic event or macro event that distracts investors
[10:42] Speaker 8: from looking at earnings, ultimately, when the dust settles, investors
[10:46] Speaker 8: will come back to earnings. We think, well, earnings season
[10:49] Speaker 8: is starting soon here for the third quarter, given, as
[10:52] Speaker 8: you mentioned, we're already in October. Investors are increasingly focused
[10:56] Speaker 8: on 2027, and we do think that the earnings growth
[10:58] Speaker 8: outlook still looks good for next year.
[11:00] Speaker 3: Matt, let's put the markets in a way to one side,
[11:04] Speaker 3: even AI. Let's talk about human beings, your clients. You
[11:08] Speaker 3: guys represent a lot of net wealth, right? What are
[11:13] Speaker 3: their behaviors? How are people having conviction or a lack
[11:18] Speaker 3: of conviction about their portfolios right now?
[11:21] Speaker 8: Yeah, so especially here in the Bay Area, A lot
[11:24] Speaker 8: of our clients are making wealth in the tech sector,
[11:28] Speaker 8: in AI.
[11:29] Speaker 3: Soon to be newly minted millionaires and more from IPOs
[11:31] Speaker 3: that may or may not happen.
[11:33] Speaker 7: Precisely.
[11:33] Speaker 8: And so there is a lot of planning that's taking place.
[11:37] Speaker 8: We're getting a lot of incoming phone calls from folks
[11:39] Speaker 8: who stand to see a very large wealth creation event
[11:42] Speaker 8: if we do see these IPOs.
[11:43] Speaker 2: Are you having to become a real estate guy, by
[11:45] Speaker 2: the way?
[11:46] Speaker 7: You know what I mean?
[11:47] Speaker 6: Yeah.
[11:47] Speaker 8: And by the way, that is a very frequent question
[11:49] Speaker 8: that's coming up. We actually had this conversation yesterday, which is,
[11:53] Speaker 8: The first things people ask about when an event like
[11:56] Speaker 8: this happens is how much home can I afford? Right now,
[12:00] Speaker 8: that is something we help our clients with. But our
[12:01] Speaker 8: main focus is on the investment portfolio. And what do
[12:03] Speaker 8: they do with these concentrated stakes in either a private
[12:06] Speaker 8: company or a public company?
[12:08] Speaker 7: That's where a lot of these conversations are. And the
[12:10] Speaker 7: right answer is.
[12:11] Speaker 8: For each family or client does differ. Some clients have
[12:14] Speaker 8: a very high risk appetite. They very much believe in
[12:17] Speaker 8: the company where they have a concentrated stock. They're going
[12:19] Speaker 8: to hold on to that position. Others though say, oh
[12:22] Speaker 8: my gosh, I've now ended up with much more money
[12:24] Speaker 8: than I ever thought I would.
[12:25] Speaker 7: And they start to diversify.
[12:27] Speaker 3: You will appreciate like many of these people watch this show, right?
[12:29] Speaker 3: What we're talking about here is Anthropix IPO and how
[12:32] Speaker 3: much of a liquidity event over time that is for
[12:34] Speaker 3: certain individuals. What was the lesson from the SpaceX IPO?
[12:37] Speaker 7: Yeah. You know, I think there are a few lessons.
[12:40] Speaker 8: One is the SpaceX IPO was well telegraphed. I think
[12:44] Speaker 8: a lot of people knew that it was coming. And
[12:46] Speaker 8: so there was a lot of preparation ahead of time.
[12:50] Speaker 8: SpaceX as a company has been around a lot longer
[12:53] Speaker 8: than some of these anthropic open AI and other AI companies.
[12:57] Speaker 7: So this is happening over a very compressed time.
[13:00] Speaker 8: Timeline for employees and engineers and founders and executives at
[13:06] Speaker 8: these companies. So unfortunately there's a lot of really quick
[13:09] Speaker 8: planning that has to take place such that when these
[13:12] Speaker 8: IPOs occur, They know what they're going to do. They
[13:14] Speaker 8: have a plan, and that's something that we're frequently helping
[13:16] Speaker 8: clients with.
[13:17] Speaker 3: Matt Weir of Goldman Sachs, thank you for bearing with me.
[13:20] Speaker 3: Extraordinary morning so far. Thank you. Go back really quickly
[13:23] Speaker 3: to the NASDAQ 100. So the breaking news that Microsoft
[13:26] Speaker 3: and other technology companies have been suspended from the PERM program.
[13:29] Speaker 3: That's the program by which They can get green cards
[13:32] Speaker 3: for foreign employees. It kind of coincides with the NASDAQ
[13:37] Speaker 3: 100 accelerating declines sort of from 10.30 Eastern to 11 a.m.
[13:41] Speaker 2: Eastern, down just half a percentage point.
[13:43] Speaker 3: But remember, technology stocks in NASDAQ 100 specifically coming off
[13:47] Speaker 3: fresh record highs, multiple consecutive record highs over the last
[13:51] Speaker 3: seven trading days or so.
[13:52] Speaker 2: Coming up.
[13:53] Speaker 3: Alphabet's big bet on AI-powered drug discovery. It's Isomorphic Labs
[13:58] Speaker 3: looking to raise fresh funding at a massive valuation.
[14:01] Speaker 2: This is Bloomberg Tech.
[14:10] Speaker 3: Record profits surging sales, yet investors still aren't impressed. Samsung's
[14:14] Speaker 3: quarterly profit jumped nearly ninefold, while TSMC saw its sales
[14:19] Speaker 3: surge 51%. Investors still wanted more. Bloomberg's global tech editor,
[14:23] Speaker 3: Peter Elstrom.
[14:24] Speaker 2: Joins us. Let's take it one by one and start
[14:26] Speaker 2: with Samsung.
[14:27] Speaker 3: You know, this is a story that I have a
[14:29] Speaker 3: little deja vu over, but continued growth and the market
[14:32] Speaker 3: still wants more.
[14:34] Speaker 9: It's a pattern that you've been seeing from Samsung and
[14:36] Speaker 9: some of the other memory chip names out there. But yeah,
[14:39] Speaker 9: Samsung reported preliminary results. Revenues were up a lot. And
[14:44] Speaker 9: as you say, operating profit was up ninefold to $ 80 billion,
[14:48] Speaker 9: which is among the most– money that any company is
[14:52] Speaker 9: going to make this quarter. But it's still, as you say,
[14:54] Speaker 9: it still wasn't enough for investors. Their shares slid after
[14:59] Speaker 9: those results. Investors are nervous about how long the AI
[15:02] Speaker 9: boom is going to continue at this point, and especially
[15:06] Speaker 9: the debt financings that's underlying. the investments that all these
[15:10] Speaker 9: companies are making at this point. It's a little bit curious, though,
[15:13] Speaker 9: because investors are anticipating that perhaps the memory chip spikes
[15:17] Speaker 9: that we've seen over the past few months are going
[15:19] Speaker 9: to slow down at this point. But in fact, the
[15:22] Speaker 9: research firms are saying that's not the case so far.
[15:25] Speaker 9: They're anticipating that memory chip prices are probably going to
[15:28] Speaker 9: go up 10 to 20 percent in the third quarter.
[15:30] Speaker 9: DRAM price, HBM prices in particular, are going up quite
[15:34] Speaker 9: a bit. Citi Analysts are forecasting HBM4 prices could go
[15:39] Speaker 9: up as much as 100% in the next year. So
[15:41] Speaker 9: it seems like there's still a lot of momentum there,
[15:43] Speaker 9: but investors are a little bit skeptical about things. Very
[15:46] Speaker 9: similar on the TSMC side, as you say, they reported
[15:49] Speaker 9: their revenue results. We don't know their profits yet, but
[15:52] Speaker 9: revenue are up 51%. At this point, again, investors are
[15:56] Speaker 9: a little bit skeptical about those results.
[15:58] Speaker 7: And it's surprising.
[15:59] Speaker 9: You saw some of the suppliers also take a hit today.
[16:02] Speaker 3: Bloomberg's Peter Elstrom summarizing what came out of Asia overnight.
[16:05] Speaker 2: Thank you very much.
[16:06] Speaker 3: If record chip profits aren't enough for investors, where do
[16:09] Speaker 3: they want to be in the AI trade? Well, Wall
[16:11] Speaker 3: Street is still betting on the company's pairing at all
[16:14] Speaker 3: chips and infrastructure. Bloomberg executives reporter Ryan Blastellica joins us
[16:18] Speaker 3: with something a lot of people reading about on the
[16:20] Speaker 3: terminal this morning.
[16:21] Speaker 2: Give me the report.
[16:23] Speaker 10: Hey, well, obviously, as you know, there's been a lot
[16:25] Speaker 10: of focus on so-called AI agents, especially since Meta released
[16:30] Speaker 10: Muse earlier last month. That was obviously very positive for
[16:33] Speaker 10: Meta shares, but a lot of investors say that the
[16:36] Speaker 10: best way to play the rise of agents is to
[16:38] Speaker 10: stick with the part of the AI trade that has
[16:41] Speaker 10: really been the focus since Chachi PT first took off
[16:44] Speaker 10: back in 2023, which is chip makers, it's infrastructure companies,
[16:48] Speaker 10: The idea is that these agents require a lot more
[16:52] Speaker 10: computing power than the previous generation of AI tools, which
[16:56] Speaker 10: is largely chatbots. So given that increased computing need, people
[17:00] Speaker 10: are expecting a lot more demand for chips, for memory,
[17:04] Speaker 10: for networking, for optical components, for all of the different
[17:08] Speaker 10: parts of the ecosystem that make up AI infrastructure.
[17:12] Speaker 2: There's a great.
[17:12] Speaker 3: Quote in your report about the leadership almost like a
[17:16] Speaker 3: game of hot potato. I mean, my takeaway from reading
[17:18] Speaker 3: the report is investors basically see the infrastructure play as
[17:21] Speaker 3: more reliable because the agent play is still a bit volatile.
[17:27] Speaker 2: It's a bit unknown.
[17:28] Speaker 7: Yeah, absolutely.
[17:29] Speaker 10: The leadership among the actual products is pretty severe. We
[17:33] Speaker 10: are seeing, you know, Meta comes out with something and
[17:35] Speaker 10: then OpenAI launches Dots. I saw earlier today that Alphabet
[17:39] Speaker 10: launched something with Google Cloud. So we're seeing just so
[17:42] Speaker 10: many different agents entering the market. Right now it's very
[17:45] Speaker 10: unclear which of these is the best, which of these
[17:48] Speaker 10: is going to see the most adoption, which of these
[17:50] Speaker 10: is going to be the most effective. So rather than
[17:52] Speaker 10: playing sort of the companies that are actually launching these products,
[17:56] Speaker 10: people are saying to play sort of the foundation for
[17:59] Speaker 10: the technology overall, which again is the sort of infrastructure
[18:02] Speaker 10: layer that allows all these products to be working in
[18:05] Speaker 10: the first place.
[18:06] Speaker 3: Bloomberg's Ryan for Celica with a must read on the
[18:08] Speaker 3: Bloomberg today. Thank you very much indeed. Coming up, drug
[18:11] Speaker 3: discovery startup Isomorphic Labs is in early talks to raise
[18:15] Speaker 3: new funds at a valuation of at least $ 40 billion,
[18:19] Speaker 3: which is a jump. We'll have the details next. This
[18:22] Speaker 3: is Bloomberg Tech. Welcome back to Bloomberg Tech. So you're
[18:36] Speaker 3: looking at live pictures of SpaceX's Crew 12 Dragon Capsule,
[18:39] Speaker 3: which has entered Earth's atmosphere, making its way back after
[18:43] Speaker 3: a 27-hour journey that began at the International Space Station.
[18:46] Speaker 3: The crew returning to Earth after 237 days at space,
[18:53] Speaker 3: 17,500 miles per hour at the point of reentry. Mach 23,
[18:58] Speaker 3: four Gs of peak force on that crew.
[19:00] Speaker 2: And very shortly.
[19:02] Speaker 3: The expectation is that the drogue chutes deploy, the four
[19:05] Speaker 3: main chutes deploy. And we'll show you those pictures as
[19:08] Speaker 3: and when that happens. Somewhat routine for NASA and for SpaceX.
[19:12] Speaker 3: But let's bring in Bloomberg's Lauren Grush, who leads our
[19:15] Speaker 3: space coverage. And again, Lauren, we got these live pics, right?
[19:18] Speaker 3: Dragon on its way back down. Hope to splash down
[19:23] Speaker 3: off the coast of California. This is routine, right? At
[19:27] Speaker 3: this at this juncture.
[19:28] Speaker 1: I mean, that is ultimately what SpaceX has achieved with
[19:33] Speaker 1: its Dragon program. I mean, this is Crew-12. It's their
[19:36] Speaker 1: 12th operational mission for the ISS program. We've got Crew-13. Oh, look,
[19:42] Speaker 1: we can see the drogues showing up there. So that's
[19:47] Speaker 1: a great sign.
[19:48] Speaker 2: But yes, yes.
[19:50] Speaker 1: Crew 13 already in orbit. They launched last week, did
[19:54] Speaker 1: their typical week-long handover, and now Crew 12 is coming home.
[19:58] Speaker 1: And this is a cool crew. They've been up in
[20:01] Speaker 1: space for about seven months, more than seven months. They
[20:04] Speaker 1: got to speak with the Artemis II crew, if you remember,
[20:07] Speaker 1: once they had swung around the moon. So it's a
[20:11] Speaker 1: nice welcome home for them after a really great time
[20:14] Speaker 1: in space.
[20:15] Speaker 3: Okay, two drogue chutes making their way off. The deployment
[20:20] Speaker 3: of four main chutes. There, you got a second look
[20:23] Speaker 3: at it. 237 days in.
[20:24] Speaker 7: Space, February 13th.
[20:29] Speaker 3: This crew lifted off towards the ISS. You know, you
[20:34] Speaker 3: and I have been talking a lot about that there
[20:36] Speaker 3: is a life cycle for ISS. It's gonna be decommissioned,
[20:41] Speaker 3: but also like SpaceX's in transition on both the Falcon
[20:44] Speaker 3: program and the Dragon program. Like how's all that being managed?
[20:48] Speaker 1: Yeah, I think I hate to say it, but watching
[20:51] Speaker 1: these splashdowns, you know, they might become a precious resource
[20:55] Speaker 1: in the near future because questions are swirling about, you know,
[20:59] Speaker 1: how long the Dragon program is actually going to stay operational.
[21:05] Speaker 1: You know, signs seem to indicate SpaceX officials have have
[21:09] Speaker 1: made hints that the Dragon program is coming to an end,
[21:13] Speaker 1: possibly by the time the ISS program is retired. And
[21:19] Speaker 1: so we have to really enjoy these splashdowns while we
[21:22] Speaker 1: can because there might not be much more of them.
[21:25] Speaker 1: There's at least four operational Crew Dragon missions for NASA
[21:29] Speaker 1: that SpaceX is going to do through the end of
[21:31] Speaker 1: the ISS program. But beyond that, it's, you know, a
[21:35] Speaker 1: question mark. There might be some other Dragon missions here
[21:37] Speaker 1: and there. but ultimately SpaceX wants to put all of
[21:40] Speaker 1: its resources into its Starship rocket. That's the massive rocket
[21:44] Speaker 1: that it's building to send people to the moon and
[21:48] Speaker 1: to Mars. And so ultimately the idea is to put.
[21:51] Speaker 2: All of the resources into that program.
[21:54] Speaker 1: And so that doesn't mean much for Dragon and for
[21:56] Speaker 1: Falcon moving forward.
[21:58] Speaker 3: At the moment those main chutes deploy, Dragon slows down
[22:02] Speaker 3: to 119 miles per hour, and at the point of splashdown,
[22:04] Speaker 3: it's kind of like 15 miles per hour. But I
[22:07] Speaker 3: always think the moment of reentry into Earth's atmosphere at
[22:11] Speaker 3: 17,500 miles per hour, pretty astonishing. The crew's been up
[22:16] Speaker 3: there a long time, and an interesting crew, some with
[22:20] Speaker 3: vast experience. What's the next big thing for NASA and
[22:24] Speaker 3: for SpaceX in terms of human spaceflight, Lauren?
[22:27] Speaker 2: What's to come?
[22:28] Speaker 7: Well, absolutely.
[22:29] Speaker 1: NASA is all in on its Artemis program right now,
[22:32] Speaker 1: which is set on sending humans back to the moon
[22:36] Speaker 1: for the first time in over half a century. Obviously,
[22:39] Speaker 1: for those who have been following You might remember the
[22:42] Speaker 1: Artemis II mission that captured everyone's attention and spread moon
[22:46] Speaker 1: joy earlier this year. That sent a crew for slingshotting
[22:50] Speaker 1: around the moon, but ultimately the goal is to send
[22:53] Speaker 1: people to land on the lunar surface, and SpaceX is
[22:56] Speaker 1: a big part of that. That Starship rocket that I
[22:59] Speaker 1: mentioned earlier, that is supposed to be a human lander
[23:02] Speaker 1: that will take future astronauts, future Artemis astronauts, to the
[23:07] Speaker 1: surface of the moon. Still quite a long thing, list
[23:10] Speaker 1: of things to do in order for that to be
[23:12] Speaker 1: ready in order to carry humans. But that is a
[23:15] Speaker 1: crucial part of the Artemis program that SpaceX is building.
[23:19] Speaker 3: OK, so momentarily, Crew 12 will splash down in the
[23:25] Speaker 3: Pacific Ocean, not too far off from Los Angeles. It's
[23:32] Speaker 3: still magical, isn't it, Lauren? It's still something special when
[23:35] Speaker 3: we do these.
[23:36] Speaker 1: You know, as much as we say it's routine, it's
[23:38] Speaker 1: certainly not. It's still, you know, just awe-inspiring and shows
[23:42] Speaker 1: what can be achieved when we, you know, put all
[23:44] Speaker 1: of our resources together into something like this. So, yeah,
[23:47] Speaker 1: it never gets old for me, even if it has
[23:49] Speaker 1: become somewhat routine and standard.
[23:53] Speaker 3: Bloomberg's Lauren Grush, who leads our space coverage. Great to
[23:55] Speaker 3: have you back on the show to walk us through that.
[23:57] Speaker 3: Thank you very much. Let's go to private markets. A
[23:59] Speaker 3: new company developing AI agents to help run businesses is
[24:04] Speaker 3: joining this kind of crowded field of services using AI
[24:08] Speaker 3: to streamline work. Hone has announced today a $ 60 million
[24:11] Speaker 3: seed round, $ 285 million valuation. It's founded by former employees
[24:15] Speaker 3: from some of the fastest growing AI names, OpenAI and Cognition.
[24:19] Speaker 3: Hone co-founder and CEO, Moritz Stefan, is with us here
[24:22] Speaker 3: in SF. So this is interesting because what you're saying
[24:25] Speaker 3: is you want to build something that runs a business goal, right?
[24:29] Speaker 3: Actually take on the responsibility, not a task, the entire
[24:32] Speaker 3: responsibility of a part of an org.
[24:34] Speaker 2: How does that work? Yes, correct.
[24:36] Speaker 11: I mean, if you look at what's going on in
[24:38] Speaker 11: AI today, right, the frontier labs are able to throw
[24:40] Speaker 11: millions of dollars in compute at solving math problems. But
[24:44] Speaker 11: when it comes to real work, I think a lot
[24:46] Speaker 11: of the impact has been quite underwhelming. And we think
[24:49] Speaker 11: a big reason for that is the way folks deploy
[24:51] Speaker 11: agents today. You know, you have to build them one off,
[24:54] Speaker 11: you have to rely on people using them correctly, and
[24:56] Speaker 11: you have to integrate them into existing workflows. With Hone,
[25:00] Speaker 11: we're building systems that can actually own outcomes end-to-end, which
[25:03] Speaker 11: means you have to run for weeks or months and
[25:05] Speaker 11: solve all the problems that go into that.
[25:07] Speaker 2: So let's go through a case study.
[25:08] Speaker 3: But what I found interesting in researching about Hone is first,
[25:12] Speaker 3: there are steps.
[25:12] Speaker 2: First, build the engine. The engine.
[25:16] Speaker 3: Built a let's call them a fleet or a team
[25:19] Speaker 3: of agents but it's constant supervision and responsibility for let's
[25:23] Speaker 3: say go to market or revenues or whatever it might
[25:26] Speaker 3: be yeah.
[25:28] Speaker 11: Because i mean in order to run any of these
[25:29] Speaker 11: functions right you need so much context that's not even
[25:32] Speaker 11: written down you can only get it actually doing the
[25:35] Speaker 11: work and interacting with the people that know how to
[25:38] Speaker 11: And so that's a big focus of ours. How can
[25:39] Speaker 11: we make these systems seamless to adopt and integrate into
[25:42] Speaker 11: an organization?
[25:43] Speaker 3: So you have an interesting backstory. You were most recently
[25:47] Speaker 3: chief of staff at Cognition. Correct. Cognition covered on this
[25:50] Speaker 3: program a lot. Explain your history of Cognition and their
[25:54] Speaker 3: involvement with home.
[25:55] Speaker 6: Yeah.
[25:56] Speaker 11: I mean, I met the Cognition folks in early 24
[26:00] Speaker 11: shortly after the company got started and then ended up joining.
[26:03] Speaker 11: It was an incredible journey. I was there for a
[26:06] Speaker 11: little over two years, helped work on Devin, led different
[26:10] Speaker 11: product and edge teams, and then was also chief of
[26:12] Speaker 11: staff to Scott, the CEO.
[26:14] Speaker 6: Yeah.
[26:15] Speaker 2: A lot of respect for them.
[26:16] Speaker 11: And I think it's incredible what they've done to bring
[26:19] Speaker 11: AI into software engineering.
[26:20] Speaker 3: And in our reporting, our understanding is that Scott personally
[26:24] Speaker 3: will invest in Hone and Cognition will be one of
[26:29] Speaker 3: several customers of Hone.
[26:31] Speaker 11: So, I mean, Scott is like a mentor and supporter.
[26:33] Speaker 11: So we've always been working very closely and it's great
[26:36] Speaker 11: to be able to continue collaborating with him.
[26:40] Speaker 2: Why is what you're offering different?
[26:43] Speaker 3: What is the problem that you perceive you're solving for here?
[26:47] Speaker 2: Yeah, I think there's a few.
[26:49] Speaker 11: I mean, one is when you go look at how
[26:50] Speaker 11: agents are deployed, right? As I was saying, you have
[26:53] Speaker 11: to manually go build them for new use cases.
[26:56] Speaker 2: That takes a lot of time, effort, and context.
[26:59] Speaker 11: And then when you actually go and deploy them, people
[27:01] Speaker 11: need to change the way they work.
[27:03] Speaker 2: And we think in.
[27:04] Speaker 11: Order for AI to actually produce ROI and real economic outcomes,
[27:07] Speaker 11: you know, it has to diffuse really, really quickly. And
[27:11] Speaker 11: that means actually automating an outcome that can own end
[27:14] Speaker 11: to end, rather relying on the fact that, you know,
[27:16] Speaker 11: the sum of some tasks actually produces.
[27:18] Speaker 2: A business outcome.
[27:19] Speaker 3: You know, if you take the go to market example,
[27:21] Speaker 3: you know, at some point. There needs to be an
[27:24] Speaker 3: element of human supervision. How does that work in Hone's case?
[27:28] Speaker 12: Totally.
[27:29] Speaker 11: I mean, when we deploy, right, we work closely with
[27:32] Speaker 11: folks in order to make sure the systems behave the
[27:34] Speaker 11: way they want. They follow the guardrails that are important.
[27:37] Speaker 6: Right.
[27:37] Speaker 2: And the agent might have questions.
[27:39] Speaker 7: Exactly.
[27:39] Speaker 11: And I mean, that's like a mutual learning for humans
[27:42] Speaker 11: and the AI systems as well.
[27:44] Speaker 7: Yeah.
[27:45] Speaker 11: But over time you can then grant them more autonomy
[27:48] Speaker 11: as they learn.
[27:49] Speaker 3: Benchmark and Index led this round, find that interesting. Also,
[27:53] Speaker 3: you kind of see model agnostic.
[27:57] Speaker 2: Is that fair?
[27:58] Speaker 3: So how have you built home?
[28:00] Speaker 11: Yeah, I mean, the model agnosticism is an important part
[28:03] Speaker 11: for us because, you know, which lab has the best
[28:05] Speaker 11: model changes every other week?
[28:07] Speaker 2: To use which is ever best at the time. Yeah,
[28:10] Speaker 2: and there's also a big cost factor.
[28:11] Speaker 11: For a lot of the tasks that we work on,
[28:14] Speaker 11: it's really important to deeply understand the context of an organization,
[28:18] Speaker 11: which requires you to use a lot of compute in
[28:21] Speaker 11: order to traverse that and understand it. And for that,
[28:24] Speaker 11: you might want to use cheaper models in many cases.
[28:26] Speaker 3: On the model side right now, is there a clear
[28:29] Speaker 3: number one from an economic proposition point of view?
[28:33] Speaker 2: It really depends on the task.
[28:34] Speaker 11: I mean, for some of the workloads that require heavier reasoning,
[28:38] Speaker 11: we use models that are different from the ones that
[28:40] Speaker 11: are actually used for executing. But that's something that we
[28:43] Speaker 11: calibrate internally continuously against our benchmarks.
[28:46] Speaker 3: $ 60 million seed's a big seed.
[28:49] Speaker 11: What are you going to do with that capital in
[28:51] Speaker 11: the first instance? Yeah, I mean, one is just hiring
[28:55] Speaker 11: an amazing team. The market is very competitive for talent,
[28:59] Speaker 11: but in order to build an amazing company, you need
[29:01] Speaker 11: to bring together amazing people. So that's a big component
[29:03] Speaker 11: of that. And then the other portion is just moving
[29:06] Speaker 11: really fast. Anything we can do to compress timelines and
[29:10] Speaker 11: move fast in building the business.
[29:11] Speaker 7: Is a big focus.
[29:13] Speaker 3: Home co-founder, Maurice Stephans. Great to have you here with
[29:16] Speaker 3: us in SF. Thank you very much indeed. Coming up,
[29:19] Speaker 3: we're going to speak to Nick Grossman from Union Square
[29:21] Speaker 3: Ventures about the firm's new $ 900 million in funds.
[29:25] Speaker 2: It's an effort to streamline the firm.
[29:27] Speaker 3: Take another look at these live pictures of SpaceX Crew 12.
[29:31] Speaker 3: It splashed down successfully in the Pacific off the coast
[29:35] Speaker 3: of Los Angeles. That crew, 237 days in space. Jessica Meir,
[29:40] Speaker 3: NASA commander. Jack Hathaway, NASA pilot. Sophie Adenaux, ESA of France.
[29:45] Speaker 3: And Andrei Fedayev, a Roscosmos astronaut, awaiting rescue and salvage
[29:51] Speaker 3: from that capsule.
[29:53] Speaker 2: This is Bloomberg Tech.
[30:02] Speaker 3: Okay, live pictures of SpaceX Crew 12 Dragon Capsule in
[30:06] Speaker 3: the Pacific off the coast of California. The crew returning
[30:10] Speaker 3: to Earth after 237 days in space at the International
[30:14] Speaker 3: Space Station. Wow, it was, again, a flawless execution. This
[30:20] Speaker 3: has become routine for SpaceX. The ISS has a limited
[30:24] Speaker 3: life decommissioning in 2030, 2031, but.
[30:27] Speaker 3: Good job, SpaceX. Okay, let's get to VC. Union Square
[30:31] Speaker 3: Ventures is nearly doubling its fund sizes, adding $ 900 million,
[30:35] Speaker 3: including half a billion for its latest early stage fund.
[30:38] Speaker 3: That's not the only change the firm's making. It's cut
[30:41] Speaker 3: ranks to just four general partners. One of them, Nick Grossman,
[30:44] Speaker 3: joins me now. I found this so interesting to read,
[30:48] Speaker 3: not just the specifics, but the rationale. And in some ways,
[30:51] Speaker 3: it's about right-sizing some of the funds, right? But right-sizing
[30:55] Speaker 3: from where?
[30:56] Speaker 2: What was wrong with them?
[30:58] Speaker 12: Well, there's nothing been wrong with our funds. We've had
[31:00] Speaker 12: an amazing history. And our focus has always been on
[31:05] Speaker 12: being a small firm, investing small funds, being targeted and focused.
[31:10] Speaker 12: That's worked for us for over 20 years now. And
[31:14] Speaker 12: we're in a different market right now. As you've seen
[31:17] Speaker 12: from everything you've been covering, it's not just that sort
[31:21] Speaker 12: of market prices are higher, but it takes a different
[31:24] Speaker 12: amount of capital to start certain kinds of companies in
[31:26] Speaker 12: this space. And so we took our previous early stage fund,
[31:30] Speaker 12: which was 275, and now we have our newest early
[31:34] Speaker 12: stage fund, which is 500 million, paired with our opportunity
[31:36] Speaker 12: fund of 400, for a total of 900 for this vintage.
[31:40] Speaker 2: Nick, how do you anticipate that manifesting?
[31:42] Speaker 3: Like you write larger checks, but do fewer deals in
[31:46] Speaker 3: a calendar year. I know that's a bit simplistic, but
[31:49] Speaker 3: that's something we've been talking about on the show a
[31:51] Speaker 3: little bit recently.
[31:52] Speaker 12: Yeah, I think we'll end up with almost a double
[31:54] Speaker 12: barbell where within our early stage fund, we're going to
[31:57] Speaker 12: go continue to go early and we may go way early.
[32:01] Speaker 12: We've been doing some investments at inception stage. And we
[32:05] Speaker 12: also want to make room at the other end for
[32:09] Speaker 12: the Series A market of today where there are, you know, 20,
[32:12] Speaker 12: $ 30 million checks needed to get into Series A. And
[32:15] Speaker 12: so within our early stage fund, there's a spread where
[32:18] Speaker 12: we're going to go early, like we've always done, and
[32:21] Speaker 12: maybe even earlier, and be ready to play in the
[32:24] Speaker 12: heart of what is today's Series A market. And then
[32:27] Speaker 12: with our opportunity fund, you know, we have always used
[32:30] Speaker 12: that to both invest in the breakout winners from our
[32:33] Speaker 12: early stage fund and also get into emerging category leaders
[32:37] Speaker 12: in the other thesis areas that we cover.
[32:40] Speaker 2: You want to lead more rounds.
[32:42] Speaker 3: So why does moving to just four general partners facilitate that?
[32:48] Speaker 12: Yeah, so we've been in, we've got an incredible team,
[32:51] Speaker 12: an incredible tight team. team that we've been working together
[32:54] Speaker 12: as a group for a long long time and we've
[32:57] Speaker 12: been in the midst of a extremely graceful generational transition
[33:02] Speaker 12: over the course of a decade so you know my
[33:06] Speaker 12: partners Brad and John sort of have scaled back you
[33:09] Speaker 12: know maybe ten years ago but they're still leading deals
[33:12] Speaker 12: here and there we've got this incredible brain trust around
[33:14] Speaker 12: the firm and And we've developed this pattern where as
[33:18] Speaker 12: the senior partners sort of move on in their careers,
[33:20] Speaker 12: they can dial back, sort of move into the backseat,
[33:23] Speaker 12: so to speak. And the younger partners can move more
[33:27] Speaker 12: into the front seat. And we've got an incredible GP
[33:30] Speaker 12: core for this fund. Rebecca Caden, my partner, who's been
[33:34] Speaker 12: my partner for over a decade, Mike Mignano coming over
[33:37] Speaker 12: from Lightspeed, who is just an incredible product.
[33:40] Speaker 2: Leader and investor.
[33:43] Speaker 12: And of course, Fred, who co-founded USV and is a
[33:46] Speaker 12: living legend. And so we've got a really spectacular core
[33:50] Speaker 12: group for this next fund cycle while also having the
[33:53] Speaker 12: broader USV brain trust in the mix.
[33:55] Speaker 3: We were just showing some of your exits, right? Twitter, Duolingo.
[34:00] Speaker 3: But go back to 2003, the firm's founding, like very
[34:03] Speaker 3: New York, early stage, those names in particular. How is
[34:07] Speaker 3: the firm different today in other ways beyond the structural
[34:10] Speaker 3: changes you just made?
[34:11] Speaker 12: You know, in a lot of ways, the firm is
[34:14] Speaker 12: the same as it's ever been. We like to say
[34:16] Speaker 12: USV is a conversation. We're just deeply collaborative as a group,
[34:21] Speaker 12: tight knit. We really work together on everything across the portfolio,
[34:26] Speaker 12: new investments, new areas that we're trying to learn on.
[34:28] Speaker 12: So that idea of of sort of a tight firm
[34:32] Speaker 12: is consistent. The other thing we've prided ourselves on over
[34:37] Speaker 12: the decades is being able to adapt to wherever the
[34:40] Speaker 12: market is, because every five or 10 years on a cycle,
[34:43] Speaker 12: the market is different and you need new moves. And
[34:46] Speaker 12: so we're always ready to do that. 10 years ago,
[34:49] Speaker 12: that was getting into crypto and thinking we might need
[34:52] Speaker 12: to buy tokens. Today, in the AI market, it's like, okay,
[34:55] Speaker 12: where's where is the market in terms of capital, what
[35:00] Speaker 12: changes do we need to make, or on the firm
[35:03] Speaker 12: and internal side, how can we implement AI and build
[35:06] Speaker 12: with agents and do new things. And so we always
[35:08] Speaker 12: try to strike the right balance between keeping the core
[35:12] Speaker 12: of the thing that's made us great for decades and
[35:15] Speaker 12: being really open to trying new moves along the way.
[35:18] Speaker 3: You know, a feature of today's market is like we
[35:21] Speaker 3: did Hone in the last segment, right? You have alumni
[35:24] Speaker 3: of now established, but short history.
[35:27] Speaker 2: Labs or neo labs.
[35:30] Speaker 3: And it's just a few people, but they're able to
[35:32] Speaker 3: raise such a big seed or a series A.
[35:35] Speaker 2: How are you going to respond to that?
[35:37] Speaker 12: Yeah, I mean, the reason that our current set of
[35:40] Speaker 12: funds is larger is that we need to be where
[35:45] Speaker 12: the market is to lead rounds at the seed stage
[35:48] Speaker 12: and at the A stage. We are still small and focused,
[35:52] Speaker 12: and so our job is not to do coverage and
[35:55] Speaker 12: be leading every round and.
[35:57] Speaker 2: Blasting capital everywhere.
[35:58] Speaker 12: We're still small and focused, but we are very much
[36:02] Speaker 12: positioned to lead A rounds, lead seed rounds, in sort
[36:06] Speaker 12: of the sizing profile of today's market.
[36:10] Speaker 3: Nick Grossman, Union Square Ventures partner. It's been great to
[36:12] Speaker 3: catch up with you. Thanks so much for your time
[36:14] Speaker 3: on the show.
[36:14] Speaker 2: Thank you.
[36:15] Speaker 3: Now coming up, anthropic IPO investors struggle to put a
[36:19] Speaker 3: price on the risk of rogue AI.
[36:22] Speaker 2: We've got that story next. This is Bloomberg Tech.
[36:36] Speaker 3: Public companies must disclose the risks they face to investors.
[36:40] Speaker 3: But what if those risks include possibly helping to wipe
[36:43] Speaker 3: out the human race? That's what investors considering Anthropix IPO
[36:47] Speaker 3: are trying to put a price on ahead of the
[36:49] Speaker 3: company's planned public market debut. Bloomberg's Anthony Hughes joins us.
[36:54] Speaker 3: Just fascinating read this morning. They file confidentially, right? That's
[37:00] Speaker 3: important context. But the question is how they disclose this risk,
[37:04] Speaker 3: what they say about it.
[37:06] Speaker 13: Yes, well, I think there's already been a lot of
[37:07] Speaker 13: discussion about the fact the risk factors are going to
[37:09] Speaker 13: include all of these, you know, is going to talk
[37:12] Speaker 13: about some of these existential risks. But I think, you know,
[37:16] Speaker 13: stopping short of these doomsday scenarios, we also have to
[37:19] Speaker 13: think about the risks that they face around, you know,
[37:22] Speaker 13: the liabilities they face if AI goes rogue and causes
[37:26] Speaker 13: things like steals data or makes unauthorised transactions, that sort
[37:30] Speaker 13: of thing. So, you know, people are very interested in
[37:31] Speaker 13: the fact that or interested in the extent to which
[37:34] Speaker 13: Anthropic's going to address this in the IPO roadshow. And
[37:38] Speaker 13: there's a whole lot of views that are percolating about this.
[37:41] Speaker 13: It's obviously something that's going to be very difficult to quantify.
[37:45] Speaker 13: But some people think that Anthropic and OpenAI should reserve money,
[37:50] Speaker 13: reserve funds to cover future liabilities and are looking to some...
[37:55] Speaker 13: precedence in the past around big natural disasters or big
[38:02] Speaker 13: corporate disasters for guidance as to how this might be
[38:07] Speaker 13: handled and also what risks there might be further out. Obviously,
[38:11] Speaker 13: it's such an emerging area of AI that it's very
[38:14] Speaker 13: difficult for people to pinpoint the risks. Then, obviously, you
[38:18] Speaker 13: have to think about things like insurance as well.
[38:21] Speaker 2: Very quick, Anthony, there's still demand for this IPO based
[38:24] Speaker 2: on our reporting, right?
[38:25] Speaker 13: Yeah, so from talking to the big end of town,
[38:27] Speaker 13: I think they're still very interested in the fact that
[38:29] Speaker 13: Anthropic is a very fast-growing company and is really at
[38:33] Speaker 13: the forefront of the AI revolution we're seeing. So, I mean,
[38:36] Speaker 13: all the indications are that people will put those market
[38:39] Speaker 13: risks to one side and chase this IPO.
[38:42] Speaker 2: Although, you know, I think people will have to think.
[38:44] Speaker 13: Carefully about how real these risks are in the future
[38:48] Speaker 13: and And try to sort of handicap them. And, you know,
[38:51] Speaker 13: I think it's just, you know, obviously there's some people
[38:54] Speaker 13: who think that these risks could be worth trillions and
[38:57] Speaker 13: trillions of dollars, but, you know, very hard to quantify.
[39:01] Speaker 2: That right now. Bloomberg's Anthony Hughes. Great to have you
[39:05] Speaker 2: on the show. Thank you.
[39:06] Speaker 3: A final look at SpaceX Crew 12 successfully splashing down
[39:09] Speaker 3: in the Pacific off the coast of California. 237 days
[39:14] Speaker 3: in space for that crew. Now the recovery effort. SpaceX
[39:17] Speaker 3: and NASA staff retrieving the crew from the Dragon capsule.
[39:21] Speaker 3: That does it for this edition of Bloomberg Tech. So
[39:23] Speaker 3: much to recap. A lot of breaking news. A lot
[39:26] Speaker 3: of market moves. Do that with the podcast. You know
[39:28] Speaker 3: where to find it. On the Bloomberg terminal. as well
[39:30] Speaker 3: as online on Apple, Spotify, and on iHeart. From San Francisco,
[39:35] Speaker 3: thanks for being with us. 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.