# US Considers Permits for Global Nvidia, AMD AI Chip Sales

An episode of [Bloomberg Tech](https://invisiblenetwork.ai/podcast/1611767434), 6 Mar 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/ef6dff9c-59d8-4f1a-bb02-b405012ca046/audio.mp3?utm_source=Podcast&in_playlist=9caafc41-289c-4115-995d-ae7c0030ab75

Bloomberg’s Caroline Hyde and Ed Ludlow discuss reports that the US Commerce Department has drafted regulations restricting AI chip shipments to anywhere in the world without American approval. Plus, Oracle plans to cut thousands of jobs as it handles a cash crunch from a massive AI data center expansion effort. And, the Pentagon has officially notified lawmakers that it has determined Anthropic and its products pose a risk to the US supply chain. See omnystudio.com/listener for privacy information.

## Transcript

[0:02] Speaker 1: Bloomberg Audio Studios, podcasts, radio news. Bloomberg Tech is a
[0:13] Speaker 1: live from Coast to coast with Carolline Hide in New
[0:17] Speaker 1: York and v Lovelow in Sentrancsco.
[0:22] Speaker 2: This is Bloomberg Tech coming up.
[0:24] Speaker 3: Bloomberg reports the US Commerce Department has drafted regulations restricting
[0:29] Speaker 3: AI chip shipments anywhere in the world without American approval.
[0:33] Speaker 4: Plus Oracle plans to cut thousands of jobs as it
[0:36] Speaker 4: handles a cash crunch from a massive AI data center
[0:39] Speaker 4: expansion effort.
[0:40] Speaker 3: And the Pentagon has officially notified lawmakers that it has
[0:44] Speaker 3: determined Anthropic and its products pose a risk to the
[0:47] Speaker 3: US supply chain.
[0:49] Speaker 4: Some extraordinary news on the private sector and indeed on
[0:51] Speaker 4: the public markets right now. And we're seeing ourselves, what
[0:53] Speaker 4: by about three quarters percent on the NASAK one hundred,
[0:56] Speaker 4: coming off of our lows.
[0:57] Speaker 5: But we're down on the week, the S and P
[0:58] Speaker 5: five hundred having.
[0:59] Speaker 4: Its worst week. It's October of last year, and in March.
[1:02] Speaker 4: Part it's because there was lots happening to oil prices,
[1:04] Speaker 4: and we're in a sixth day of conflict with the
[1:06] Speaker 4: Middle East with Iran, and we're seeing ail up six
[1:09] Speaker 4: percent at one point. We're now, of course, eclipsing ninety
[1:11] Speaker 4: one dollars on the BNT on the global contract. All
[1:14] Speaker 4: of this adding to inflation pressure when we're seeing weakness
[1:17] Speaker 4: in the labor market, non farm payrolls are surprise loss
[1:20] Speaker 4: in jobs, no additions. We're also seeing, therefore, money coming
[1:23] Speaker 4: out of some of those risk assets. Bitcoin off by
[1:26] Speaker 4: three point six percent today, ED, what are.
[1:28] Speaker 5: You looking at?
[1:29] Speaker 2: Yeah, chip stocks are under pressure.
[1:30] Speaker 3: The Philadelphia Semiconductor indexes down in this session, down on
[1:34] Speaker 3: the week, and headed to two straight weeks of declines
[1:36] Speaker 3: for the first time since November. Bloomberg News is reporting
[1:40] Speaker 3: that the Commerce Department has drafted new rules that would
[1:43] Speaker 3: restrict US shipments of AI chips worldwide, giving Washington broad
[1:48] Speaker 3: authority over whether other countries can build facilities to train
[1:52] Speaker 3: and run our official intelligence models.
[1:53] Speaker 2: That's all according to sources.
[1:55] Speaker 3: Bloomberg Tech and Industrial Policy Report of Maggie Eastland joins
[1:59] Speaker 3: us with the let's get into the content of those
[2:02] Speaker 3: draft rules and what Bloomberg's reported.
[2:04] Speaker 2: What do we need to know?
[2:07] Speaker 6: Of course, so as we've reported, there is a license
[2:10] Speaker 6: requirement for AI chips sent almost anywhere in the world. Now,
[2:15] Speaker 6: whether or not these licenses will be handed out. Depends
[2:18] Speaker 6: a lot on how many chips, as compared to a
[2:22] Speaker 6: Blackwell and video black Well, how many of those chips
[2:25] Speaker 6: each does each end user want to purchase, And now
[2:28] Speaker 6: on the higher end, for end users wanting to purchase
[2:31] Speaker 6: up to two hundred thousand, that may actually require and
[2:34] Speaker 6: the US may ask nations to invest in AI within
[2:38] Speaker 6: US borders. So this opens up an opportunity for chip
[2:41] Speaker 6: licenses to become a factor in global trade talks.
[2:46] Speaker 4: Something that I'm sure many a country of Bristol at, Maggie.
[2:48] Speaker 4: What's interesting is that the Commas Department has responded and
[2:53] Speaker 4: they say the draft rule isn't meant to function as
[2:55] Speaker 4: an Nvidia export band. Can you just give us the
[2:58] Speaker 4: nuance of how the government has responded.
[3:00] Speaker 5: You're reporting right.
[3:03] Speaker 6: So this response shows and tries to emphasize that this
[3:07] Speaker 6: is not the same as the Biden administration and.
[3:10] Speaker 5: Their diffusion rule.
[3:12] Speaker 6: It's one thing is it is not based on countries. However,
[3:15] Speaker 6: there is a big role here that the government is
[3:18] Speaker 6: playing as a gatekeeper, and whether or not this is
[3:21] Speaker 6: more restrictive is ultimately going to depend on how the
[3:24] Speaker 6: US decides to dole out these licenses, and it does
[3:28] Speaker 6: have the leeway to be more or less.
[3:30] Speaker 5: Strict with that.
[3:32] Speaker 7: Maggie.
[3:32] Speaker 3: An important point here is it doesn't necessarily change US
[3:37] Speaker 3: policy with regards to China right now. You outline that,
[3:40] Speaker 3: But there seems to be like this tiered response if
[3:43] Speaker 3: it's a small batch of chips for a small data
[3:45] Speaker 3: center project lighter review. But at the other end of
[3:48] Speaker 3: the extreme in our reporting, it seems to indicate that
[3:52] Speaker 3: you have a nation state country to country negotiation. If
[3:57] Speaker 3: the US is going to allow the export of cutting
[4:00] Speaker 3: edge AI accelerators.
[4:04] Speaker 6: Exactly as you say, for any shipments over two hundred
[4:08] Speaker 6: thousand Blackwell equivalents, there will have to be a nation
[4:11] Speaker 6: to nation negotiation, and it is sort of a tiered response.
[4:15] Speaker 6: If it's under a thousand chips, that's going to be
[4:18] Speaker 6: a lot easier to get approval. But again, as you
[4:20] Speaker 6: work up the chain and want more and more ships,
[4:22] Speaker 6: the US is going to be asking for more in return.
[4:26] Speaker 4: Maggie Eastlan on what the US we understand is currently considering.
[4:30] Speaker 4: We so appreciate it. Now, let's stick with Global Tech News.
[4:33] Speaker 4: Ted Mordenson is with US. He's bed managing director. He
[4:36] Speaker 4: oversees technology research. Boy, is there are a lot to
[4:40] Speaker 4: be trying to digest as someone who analyzes tech. Right now,
[4:44] Speaker 4: just focus in on what potentially the US could be
[4:46] Speaker 4: thinking about here and overseeing of basically how in video
[4:49] Speaker 4: and AMD export to the rest of the world.
[4:52] Speaker 8: Your reaction, I think it's a broader focus. It also,
[4:57] Speaker 8: let's face facts.
[4:58] Speaker 9: I mean, the Trump administration is providing a broad blanket
[5:02] Speaker 9: not only on energy policy, Look what we've done in Venezuela,
[5:05] Speaker 9: look what we're doing in our own right.
[5:07] Speaker 8: Now, but also on rare earth.
[5:09] Speaker 9: This all feeds together in one big ecosystem, and I
[5:14] Speaker 9: think they just want control of what's happening on our
[5:17] Speaker 9: next generation silicon at that three and two nanimeter node.
[5:23] Speaker 9: This is one of the reasons why we've built eight
[5:26] Speaker 9: fabs at TSM and Arizona. We have to control that technology,
[5:31] Speaker 9: specifically at that two nanimeter node, because that's where all
[5:34] Speaker 9: the AI chip development is going. So I think they're
[5:38] Speaker 9: just trying to get their arms around the whole situation
[5:41] Speaker 9: that the US is in more of a power solution
[5:46] Speaker 9: verse being ripped off globally a.
[5:49] Speaker 5: Power solution first.
[5:50] Speaker 4: I mean, at the moment, we're seeing a negative reaction
[5:53] Speaker 4: to the stocks in the back of that story. In fact,
[5:56] Speaker 4: very few stocks are in the green today and maybe
[5:58] Speaker 4: Marvel and broad Com out perform because of numbers we
[6:02] Speaker 4: see defense tech outperformed.
[6:04] Speaker 2: TED.
[6:04] Speaker 4: What do you do in this current geopolitical moment of tension?
[6:09] Speaker 9: So I'm on with some very big clients, you know, constantly,
[6:14] Speaker 9: whether it be during the day or even at night
[6:17] Speaker 9: lately during earnings, and I think what people are doing
[6:20] Speaker 9: is we're kind of post last Saturday, and there's not
[6:24] Speaker 9: a lot of cycle experience. I mean, I'm old, so
[6:27] Speaker 9: I've seen pretty much all this. It doesn't really surprise
[6:29] Speaker 9: me that much. But we're now in this I would
[6:32] Speaker 9: call it a capital preservation phase, where when you don't
[6:37] Speaker 9: have a you know, a clear picture of what could happen,
[6:42] Speaker 9: whether it be on the commodity side the raid environment,
[6:46] Speaker 9: you tend to go to a capital preservation mode. And
[6:49] Speaker 9: what I mean by that is there's a lot of
[6:52] Speaker 9: d risking going on in tech portfolios, and there's been
[6:56] Speaker 9: some very violent moves between software and also semiconductors, and
[7:02] Speaker 9: I think there's also some pms that are quite frankly
[7:05] Speaker 9: just moving to other sectors that are a little bit
[7:09] Speaker 9: more defensive.
[7:10] Speaker 3: So you know, Caroline outlined right at the top of
[7:13] Speaker 3: the show, the sort of the trajectory of the Nasdaq
[7:15] Speaker 3: one hundred, right, I think, on track for its worst
[7:18] Speaker 3: week since October. Then I said, chips are following a
[7:21] Speaker 3: similar pattern, back to back weekly declines for the first
[7:25] Speaker 3: time since November. It seems a strange question, But what
[7:29] Speaker 3: is the main catalyst in this market right now? What
[7:31] Speaker 3: is it that your clients are saying is driving their
[7:34] Speaker 3: psychology when you're on the phone with them.
[7:37] Speaker 8: They need clarity. You know, everybody's looking at the tenure.
[7:40] Speaker 9: I mean it revers down to the four to twelve
[7:42] Speaker 9: range earlier in the day, now it's back last time
[7:45] Speaker 9: I checked out of the four seventeen range. They're worried
[7:48] Speaker 9: about a potential spike in rakes. There's inflation everywhere in tech.
[7:53] Speaker 9: Every single conference call I'm on and I listened to
[7:56] Speaker 9: over seventy during earning season, everybody is passing on price.
[8:01] Speaker 9: So I think the institutional investor is looking at locking
[8:05] Speaker 9: in gains on some really leadership stocks, building some more
[8:12] Speaker 9: cash levels just to be more nimble. Buying from the
[8:15] Speaker 9: fearful selling to the needy is a discussion I have
[8:19] Speaker 9: almost every single conversation. And then the other things that
[8:22] Speaker 9: portfolio managers are doing is they're upgrading their portfolio. And
[8:27] Speaker 9: this is a great opportunity when the markets are down
[8:30] Speaker 9: to really look at these A companies with unbelievable management teams,
[8:35] Speaker 9: product cycles that move into AGENTIIC and free cash LOWI
[8:40] Speaker 9: is king and if you do a screen on those names,
[8:43] Speaker 9: there's a bunch of tech companies that are way off
[8:46] Speaker 9: their highs and that's where that's where we're doing most
[8:49] Speaker 9: of the work.
[8:50] Speaker 3: It's actually the S and P five hundred I think
[8:52] Speaker 3: that's having the rough week relatives in AWE. That one
[8:54] Speaker 3: hundred one name that is having a very good week
[8:58] Speaker 3: is Palenteer.
[8:59] Speaker 2: You up almost thirteen.
[9:01] Speaker 3: Percent, and that's like logical, right, Like Caroline was making
[9:04] Speaker 3: the point, I think it was Monday, maybe Tuesday that
[9:08] Speaker 3: with everything that's happening with the war in Iran and
[9:10] Speaker 3: indeed the back and forth between the Department of Defense,
[9:15] Speaker 3: the Pentagon and AI Providers, that you'd expect that name
[9:19] Speaker 3: to be higher, you know, you'd expect it to be
[9:21] Speaker 3: in the spotlight right now.
[9:24] Speaker 9: It truly is one of these next generation companies that
[9:27] Speaker 9: has been put in pretty much every single agency. They
[9:30] Speaker 9: started with the army, but they've gone to the Space Force,
[9:35] Speaker 9: the Navy, pretty much every single industry. And if you
[9:38] Speaker 9: just connect the dots on our success in Iran is
[9:43] Speaker 9: based upon Pallanteer being embedded in the IDEF, the Israeli
[9:48] Speaker 9: Defense Force as well as.
[9:53] Speaker 8: Pretty much I.
[9:54] Speaker 9: Would assume all of our agencies in relationship to our
[9:58] Speaker 9: Department of War.
[9:59] Speaker 8: This gives the warfighter.
[10:01] Speaker 9: And you know, I flew for you know a lot
[10:03] Speaker 9: of years and I feel so antiquated on what my
[10:08] Speaker 9: new naval aviators have at their disposal. They can see
[10:11] Speaker 9: around corners, they can see the battlefield in real time,
[10:17] Speaker 9: and that's that's really because of a lot of the
[10:20] Speaker 9: Paneteer software and their ability to really take on an
[10:25] Speaker 9: immense amount of data and make sense of it.
[10:28] Speaker 4: I love hearing about that history and that understanding you
[10:31] Speaker 4: have ted. Can you therefore opine a little bit on
[10:34] Speaker 4: how difficult it might be for them to unentwine themselves.
[10:38] Speaker 5: That's even a word with anthropic, because Panta has had
[10:42] Speaker 5: a partnership with them. If we are seeing them forced out,
[10:46] Speaker 5: is that going to be an issue for the business.
[10:48] Speaker 9: I think it's I think it's a really political I
[10:50] Speaker 9: think at some point of the Defense Department, and you know,
[10:54] Speaker 9: the government will need and tropic and they'll still find
[10:58] Speaker 9: a way of somewhere in the middle. Same thing with
[11:01] Speaker 9: open eye, open AI. If you go back in the
[11:04] Speaker 9: way back machine. I mean, this has happened at Amazon,
[11:08] Speaker 9: it's happened at Google, it's happened at Microsoft.
[11:11] Speaker 8: We'll get there somehow.
[11:13] Speaker 9: What I think your viewers need to understand is the
[11:17] Speaker 9: US government, specifically the Defense Department, needs the best technology,
[11:23] Speaker 9: and I think what we're using is kind of a
[11:25] Speaker 9: strong arm tactic to get them back at the table
[11:28] Speaker 9: and to get to some middle.
[11:29] Speaker 3: Grown Ted Wordenson of Bed with a very close view
[11:36] Speaker 3: of what's happening between technology and the world, the state
[11:40] Speaker 3: of the world.
[11:41] Speaker 2: Right now, I really appreciate it.
[11:42] Speaker 3: Now coming up in the program, Oracle plans to act
[11:45] Speaker 3: thousands of roles as soon as this month. We're going
[11:48] Speaker 3: to discuss what is behind that move and the Bloomberg reporting.
[11:51] Speaker 2: This is Bloomberg Tech.
[12:02] Speaker 3: Oracle is planning to cut thousands of jobs as it
[12:05] Speaker 3: looks to manage a cash crunch driven by AI spending.
[12:08] Speaker 3: According to sources, Let's get the details of Bloomberg's Brady Ford,
[12:11] Speaker 3: who broke the story. Lots of detail in the report,
[12:15] Speaker 3: the size, the scope, the when, what do we need
[12:17] Speaker 3: to know?
[12:19] Speaker 10: It costs a lot of money to build data centers.
[12:21] Speaker 11: Right, Oracle is betting its company on the fact that
[12:24] Speaker 11: it will be an AI infrastructure provider, and over the
[12:28] Speaker 11: next couple of years that means negative cash flow. It's
[12:32] Speaker 11: feeling a crunch, and so it's responding, right, it's planning
[12:35] Speaker 11: what appears to be one of the largest layoffs on
[12:38] Speaker 11: record for the company. Thousands of workers impacted as soon
[12:41] Speaker 11: as this month. And yeah, I mean it is certainly
[12:45] Speaker 11: a rough market out there for jobs in the tech industry,
[12:49] Speaker 11: and this is just one more data point.
[12:50] Speaker 5: For that, Brady.
[12:53] Speaker 4: It needs money for AI and the infrastructure of it.
[12:55] Speaker 4: It needs less people because of AI. And you say
[12:58] Speaker 4: that the job cuts are going to be coming to
[13:00] Speaker 4: those where I assume the latest greatest AI agent is
[13:04] Speaker 4: going to be able to perform for them.
[13:07] Speaker 10: Exactly.
[13:07] Speaker 7: Yeah.
[13:08] Speaker 10: I think every management team right now, everybody who has.
[13:10] Speaker 11: A P and L is being asked, Hey, when you're
[13:13] Speaker 11: doing your headcom planning, let's think about what roles we
[13:17] Speaker 11: need less of. We all know these AI and tools
[13:20] Speaker 11: aren't quite there in terms of replacing people, but in
[13:23] Speaker 11: terms of time saving, in terms of hey, maybe if
[13:27] Speaker 11: you had a team of ten people, can we do
[13:29] Speaker 11: it with eight people.
[13:30] Speaker 10: From what we are hearing.
[13:32] Speaker 11: From sources, that is certainly a factor in this layoff.
[13:35] Speaker 10: Though at the end of the day, it's cost cutting.
[13:38] Speaker 10: I think that's the big story rather than it.
[13:41] Speaker 11: Being a you know, AI is so amazing that we
[13:43] Speaker 11: can replace people.
[13:44] Speaker 4: Yeah, then you want some whether it's AI washing or not, continues.
[13:48] Speaker 4: It's proty forward. Thank you so much for joining us
[13:50] Speaker 4: on that story. Look, another tech leader has been acknowledging
[13:53] Speaker 4: the impact of AI on jobs Page a Duty CEO
[13:56] Speaker 4: Jennifer tahadah I spoke with her at Bloomberg's New Voices
[14:00] Speaker 4: event earlier this week, where she said AI's impact from
[14:02] Speaker 4: productivity is changing the way her company fills vacant positions.
[14:07] Speaker 5: Yeah, we've tried.
[14:08] Speaker 12: I moved the word backfill from our vocabulary because that
[14:12] Speaker 12: sort of infers that you need something exactly like what
[14:15] Speaker 12: you have before, and the market is moving so fast
[14:18] Speaker 12: that by the time one person exit the role, you
[14:20] Speaker 12: actually need to think completely differently about what the opportunity
[14:23] Speaker 12: is to add capacity to your team in a new way.
[14:26] Speaker 4: So are you seeing few of people working at Patrigg
[14:30] Speaker 4: and how does that perhaps.
[14:32] Speaker 5: Change the conversation you're having with employees.
[14:35] Speaker 12: Well, I think employees understand that we're looking for each
[14:39] Speaker 12: of them to demonstrate not only productivity and efficiency, but
[14:45] Speaker 12: higher velocity in the innovation. And I stand up at
[14:48] Speaker 12: town hall almost every month and say, I'm not asking
[14:51] Speaker 12: you to work more, I'm not asking you to work harder.
[14:53] Speaker 12: I'm asking to work differently. I'm asking you to embrace
[14:57] Speaker 12: and leverage AI and service of competitive advantage, but mostly
[15:00] Speaker 12: in service of our customers. Right, And I think that's
[15:04] Speaker 12: exciting for most employees.
[15:07] Speaker 3: That was page Judy CEO Jennifer Tahada. Now, US employers
[15:11] Speaker 3: did unexpectedly slash ninety two thousand jobs last month, raising
[15:15] Speaker 3: the unemployment rate to four point four percent. Those cuts
[15:19] Speaker 3: are being blamed on strikes at major healthcare providers and servers.
[15:23] Speaker 2: Whether not AI.
[15:25] Speaker 3: How much longer could it be until AI does impact
[15:29] Speaker 3: to the labor market market? Gimball, executive director of the
[15:32] Speaker 3: Yale Budget Lab, has been doing research on AI labor
[15:35] Speaker 3: market impact and joins us WE every week month and
[15:40] Speaker 3: quarter wait for the definitive data set that says AI
[15:45] Speaker 3: is doing something here.
[15:48] Speaker 2: Have you seen it yet? Will we see it soon?
[15:52] Speaker 13: I really haven't seen it yet. I think everyone is
[15:55] Speaker 13: so focused on what the technology can do, because the
[15:58] Speaker 13: technology is so amazing that they're jumping straight to and
[16:03] Speaker 13: therefore we're going to really quickly see these labor market impacts.
[16:08] Speaker 13: I think it's really important to keep in mind there
[16:10] Speaker 13: are a lot of things that affect deployment of technology.
[16:14] Speaker 13: It policies, economic pressures, demographic changes, liability concerns, and so
[16:22] Speaker 13: the question isn't just what can the technology do, but
[16:26] Speaker 13: how quickly is society going to and companies going to
[16:30] Speaker 13: rearrange itself around it.
[16:32] Speaker 3: The news is that, you know, the unexpected cut to
[16:36] Speaker 3: jobs in February. That's the news. That's what's driving markets today.
[16:40] Speaker 3: Within the data set, did you see anything? Did you
[16:43] Speaker 3: go deep and say, Okay, what can I put from
[16:46] Speaker 3: this into my own model?
[16:49] Speaker 9: Yeah?
[16:49] Speaker 13: And when you look at this data, I mean, strikes
[16:51] Speaker 13: did play a role, so we should see those jobs
[16:53] Speaker 13: coming back.
[16:54] Speaker 7: But even if you.
[16:55] Speaker 13: Take the strikes out, about fifty percent of the job
[16:58] Speaker 13: loss was in goods producing sectors approached to private service
[17:01] Speaker 13: providing sectors. So things like manufacturing, manufacturing, construction, mining, and logging,
[17:08] Speaker 13: those are not sectors that we expect to be affected
[17:10] Speaker 13: by AI in the same way. And for context, right,
[17:13] Speaker 13: they were about fifty percent of the cuts, but they're
[17:15] Speaker 13: only sixteen percent of employment, and so that doesn't suggest
[17:19] Speaker 13: that this is AI. It suggests that, you know, you
[17:22] Speaker 13: maybe have impacts of tariffs plus just some general cyclical weakening.
[17:27] Speaker 4: Martha, Yeah, we saw in information about eleven thousand jobs gone.
[17:30] Speaker 5: Out of those ninety two thousand.
[17:33] Speaker 4: Can you tell us what data points we should look
[17:35] Speaker 4: at then, if it is going to be about productivity gains,
[17:38] Speaker 4: if it is going to be more about different areas
[17:41] Speaker 4: that is starting to see the impact of an AI
[17:44] Speaker 4: agent for example.
[17:47] Speaker 13: Yeah, one of the things that I've been looking at
[17:48] Speaker 13: is how quickly the just composition of jobs in our
[17:52] Speaker 13: economy is changing. Because we gain and lose jobs all
[17:55] Speaker 13: the time, that's a pretty standard part of an economy.
[17:58] Speaker 13: And so just because we lose jobs in one month
[18:02] Speaker 13: doesn't mean that we should immediately assume that it's AI
[18:04] Speaker 13: related job loss. But if we're losing jobs predominantly in
[18:08] Speaker 13: one sector or a couple of sectors that are unusual
[18:11] Speaker 13: that look like AI. That's a signal that maybe this
[18:15] Speaker 13: is starting to happen. And that's why you know, at
[18:17] Speaker 13: Budget Lab we've been tracking this measure of occupational distribution
[18:21] Speaker 13: every month and it still isn't really flashing anything. I
[18:26] Speaker 13: think it's really important to separate out the vibes on
[18:29] Speaker 13: AI in the labor market from what's happening in the data.
[18:32] Speaker 13: And you know, I think you're just seeing broad agreement
[18:35] Speaker 13: that it's not in the data yet. You just saw
[18:36] Speaker 13: a report from Anthropic come out yesterday saying, yeah, we're
[18:40] Speaker 13: really not seeing macroeconomic impacts here yet.
[18:43] Speaker 4: Vibe is one thing, but Jack Dawcy saying he's cutting
[18:46] Speaker 4: half of his staff is a data point. And yes,
[18:50] Speaker 4: we can say whether there's some AI washing within it,
[18:52] Speaker 4: but still there are executives saying we are not backfilling,
[18:56] Speaker 4: we are using AI.
[18:58] Speaker 5: We are cutting half of our.
[18:59] Speaker 4: Staff because we think AI is at a cusp where
[19:01] Speaker 4: we can do more with it.
[19:02] Speaker 5: Martha, when does that start to leach in?
[19:06] Speaker 13: Do you think tracking CEO statements is not the ideal
[19:10] Speaker 13: way to look at this for a couple of reasons.
[19:13] Speaker 8: Right.
[19:13] Speaker 13: One is there's huge selection and who even makes a
[19:16] Speaker 13: statement about layoffs. Your hardware store down the street is
[19:19] Speaker 13: not issuing a statement every time they lay someone off.
[19:22] Speaker 13: There's selection effects in whose statements get covered, right. The
[19:26] Speaker 13: statement from Jack Dorsey was you know, very eye catching,
[19:29] Speaker 13: and so I got picked up. And also, frankly, you know,
[19:32] Speaker 13: CEOs have incentives to talk about things in a certain way.
[19:36] Speaker 13: Shareholders want to hear that they're making AI investments and
[19:40] Speaker 13: driving productivity, and they may not even have a perfect
[19:43] Speaker 13: sense of what hiring would look like in the absence
[19:46] Speaker 13: of AI. And so I do think it's a moment
[19:49] Speaker 13: where you know, yes, we should listen to people about
[19:51] Speaker 13: what's going on. We should listen to CEOs. But it's
[19:54] Speaker 13: really important to go back to the data and make
[19:56] Speaker 13: sure we're taking a really data driven approach here rather
[19:58] Speaker 13: than you know, following the latest.
[20:03] Speaker 10: Company announcement.
[20:04] Speaker 4: And that is why your data approach is so important.
[20:06] Speaker 4: Some of the research you do is just great, go
[20:08] Speaker 4: and read it. Labor market a exposure, what do we know?
[20:10] Speaker 4: For example, Martha Gimmell, executive director and co founder at
[20:13] Speaker 4: the Yale Budget Lab, we appreciate you now coming up
[20:16] Speaker 4: soft Bank in search of forty million dollars to fuel.
[20:20] Speaker 5: Its AI ambitions. More on that next, this is bloombag Tech.
[20:33] Speaker 5: It's time now for talking tech and first up.
[20:34] Speaker 4: Robin Hood raised six hundred and fifty eight million dollars for
[20:37] Speaker 4: its closed end fund designed to give US retail investors
[20:40] Speaker 4: access to private companies.
[20:42] Speaker 5: Robinhood, CEO of Latana, joined Open Interest. Just take a listen.
[20:46] Speaker 14: Robin Adventure is a closed end fund listed on the
[20:48] Speaker 14: Nicey that invests in private companies. The idea is we
[20:54] Speaker 14: raise the capital in an IPO and then we invest
[20:58] Speaker 14: that capital in private companies. And so far we've got
[21:03] Speaker 14: a portfolio frontier names in their industries. You mentioned Data
[21:08] Speaker 14: Bricks or a Mercore, Revolute Ramp. Those are just some
[21:13] Speaker 14: of the companies that are are part of this portfolio.
[21:16] Speaker 4: Plus uber Well it is posted job listings that suggest
[21:20] Speaker 4: the right share company is stepping our efforts to test
[21:22] Speaker 4: a subscription offering for drivers and couriers around the world.
[21:26] Speaker 4: Now the move underscores a sharper focus on subscriptions as
[21:29] Speaker 4: emerging competitors load drivers with flat fee models and soft
[21:33] Speaker 4: bank Well it's seeking up to forty billion dollars from
[21:36] Speaker 4: banks to help finances investment in Open AI. Now that's
[21:38] Speaker 4: according to sources, the roughly twelve month bridge loan will
[21:41] Speaker 4: be unwritten by fall enders including JP Morgan Chase. So
[21:44] Speaker 4: SoftBank has been unloading assets in fact, including its taken
[21:47] Speaker 4: VideA to bankroll.
[21:48] Speaker 5: It's growing better on open AI, but ED is an
[21:51] Speaker 5: interesting time to be.
[21:52] Speaker 4: Wanting more banks to be syndicating dead at the moment exposed.
[21:55] Speaker 3: To AI because credit pressure is building on soft Bank
[21:59] Speaker 3: in particularly a twelvemonth bridge loan. And remind yourself that
[22:03] Speaker 3: forty billion dollars is their largest ever US denominated borrowing caroc,
[22:08] Speaker 3: isn't it.
[22:08] Speaker 4: Also within that conversation, we're just honing from Vlad Tenorth
[22:11] Speaker 4: and Robin Hood on one side, people still wanting access
[22:14] Speaker 4: to private markets. But then we have today the news
[22:16] Speaker 4: around Blackrock and it's private market credit fund and having
[22:19] Speaker 4: to say, look, you can't keep taking.
[22:21] Speaker 5: Money from at the moment. It's really extraordinary moves.
[22:24] Speaker 3: Yeah, and we've been on top of it all week.
[22:25] Speaker 3: That credit story is not going away.
[22:27] Speaker 2: Okay.
[22:27] Speaker 3: Coming up, we're going to discuss the data center threats
[22:30] Speaker 3: in the Middle East with sand Winter Levy of the
[22:32] Speaker 3: Carnegie Endowment for International Peace. That's next it's halftime. This
[22:36] Speaker 3: is Bloomberg Tech. Welcome back to Bloomberg Tech. This is
[22:48] Speaker 3: what markets look like on Friday to end the week.
[22:51] Speaker 3: There's a lot going on in the context of Bloomberg's
[22:53] Speaker 3: reporting on the tech sector and the war in Iran.
[22:56] Speaker 3: And as that one hundred is off session lows but
[22:58] Speaker 3: down eight tens of percent, and bonds have kind of
[23:00] Speaker 3: been wavering. You see the US tenure yield at four
[23:03] Speaker 3: point four one four percent. Interestingly, like a lot of
[23:07] Speaker 3: pressure on the Nasdaq one hundred in the moment is
[23:09] Speaker 3: coming from some of the megacap names. I was talking
[23:11] Speaker 3: with the team Caro on the equities desk, like what's
[23:14] Speaker 3: going on. There is some kind of stabilization in memory pricing,
[23:18] Speaker 3: for example. We know that there's a recalculation in recent
[23:23] Speaker 3: days of capital expenditures that impact some of the hyperscalers.
[23:26] Speaker 3: But you look at Apple, Meta, Tesla, Amazon, those are
[23:29] Speaker 3: on a points basis the biggest drags right now, and
[23:31] Speaker 3: not that much in the news flow beyond. It's just,
[23:34] Speaker 3: I guess a volatile week politically speaking, and as we
[23:38] Speaker 3: went through a ted there's a lot that the market
[23:40] Speaker 3: needs clarity on.
[23:41] Speaker 4: Yeah, let's talk about one of those stories. Politically speaking,
[23:44] Speaker 4: Anthropic it's fouled to legally contest a Pentagon decision that
[23:48] Speaker 4: declared the company a US supply chain threat, a designation
[23:51] Speaker 4: typically reserved for foreign anversaries. Capping off a very turbulent
[23:55] Speaker 4: week in the AI race, Blue most Shrinkafari joins us
[23:58] Speaker 4: now for more on the unfolding AI saga that you
[24:01] Speaker 4: have been at the forefront of reporting with for at
[24:03] Speaker 4: least a week now. It was last week when we
[24:05] Speaker 4: first saw the explosive conversation around being deemed a supply
[24:09] Speaker 4: chain risk, and now it's come in writing.
[24:14] Speaker 2: That's right, and you know.
[24:15] Speaker 15: Dario m and A, the CEO of Aanthropic, also put
[24:18] Speaker 15: out a memo yesterday saying that he believes that the
[24:23] Speaker 15: order is narrow in scope to the effect that it
[24:27] Speaker 15: won't actually impact hopefully business with customers who are not
[24:31] Speaker 15: at the government or working directly with the Pentagon. I
[24:35] Speaker 15: think it's going to take a while to figure out
[24:37] Speaker 15: exactly how customers react to this.
[24:42] Speaker 3: A few places that we've looked over recent days interesting
[24:45] Speaker 3: there's the Censor Tower data on uninstalls and downloads, and
[24:49] Speaker 3: there's the app stores and like it's part of the
[24:51] Speaker 3: Q and AI newsletter. Right, we're going to look at
[24:54] Speaker 3: this idea that there have been changes in chat, GPT
[24:58] Speaker 3: open aies tool in terms of downloads and DOUN stores, and
[25:01] Speaker 3: then Claude Anthropics what are you seeing.
[25:05] Speaker 15: Yeah, So it's really interesting because Anthropics claud business has
[25:08] Speaker 15: historically been more enterprise, especially in the past year, and
[25:13] Speaker 15: now we're seeing this shift the sort of consumer bump
[25:15] Speaker 15: for the Claud app, and a lot of that is
[25:18] Speaker 15: because some users agree with the company stands on pushing
[25:23] Speaker 15: for restrictions on surveillance and autonomous weaponry in their negotiations
[25:29] Speaker 15: with the Department of War.
[25:32] Speaker 10: So now those.
[25:33] Speaker 15: Gains are up to a million downloads a day. Yesterday
[25:37] Speaker 15: the company was saying that. Being said, Chatchapt by far
[25:40] Speaker 15: has a large majority of the market share here. It's
[25:45] Speaker 15: going to take a lot of sustain growth at that
[25:47] Speaker 15: level for Anthropic to even come close to Chatchapt on
[25:49] Speaker 15: the consumer side. But consumer support, any kind of support,
[25:53] Speaker 15: is much needed for Anthropic right now.
[25:54] Speaker 10: So that is an interesting bump.
[25:57] Speaker 4: In terms of just where they stand and consume penetration.
[26:01] Speaker 4: It is much smaller than Opening Eye. More broadly, how
[26:04] Speaker 4: much of a concern is it to Anthropic that the
[26:09] Speaker 4: enterprise part of the business could be impacted if they
[26:12] Speaker 4: are asked on install As you said, there is clarification
[26:15] Speaker 4: that really a wouldn't extend too many parts of the
[26:19] Speaker 4: other companies that use clawed only with Pentagon contracts.
[26:24] Speaker 7: That's right.
[26:24] Speaker 15: This is the million billion, multi billion dollar question, right.
[26:28] Speaker 15: Anthropic has grown very rapidly in the past year. It's
[26:33] Speaker 15: now at a reported nineteen billion in run rate revenue. However,
[26:38] Speaker 15: a majority of that does come from enterprise contracts. So
[26:40] Speaker 15: if they are our enterprise business, so if they are
[26:42] Speaker 15: losing marquee customers over this, who may also have dealing
[26:47] Speaker 15: themselves with the Pentagon and are worried about the implications
[26:51] Speaker 15: of the supplier risk designation, that could be devastating to Anthropics.
[26:55] Speaker 15: So that is a big question is how why does
[26:58] Speaker 15: the supplier risk designation really? Will Anthropics end up legally
[27:02] Speaker 15: challenging it? And is this issue resolved? I think that's
[27:06] Speaker 15: going to be the question to watch for customers.
[27:08] Speaker 3: Bloombo Sharen Gafari, Thank you very much. And as tensions
[27:12] Speaker 3: rise between Washington and leading AI firms, the global stakes
[27:16] Speaker 3: are growing as well. Data centers, the backbone of artificial
[27:19] Speaker 3: intelligence have now merged as potential targets in the escalating
[27:23] Speaker 3: conflict in the Middle East. Let's bring in Sam winter Levy,
[27:26] Speaker 3: Carnegie Endowment for International Peace Fellow. Twenty four hours ago,
[27:30] Speaker 3: we went to our correspondent in the emir region and
[27:34] Speaker 3: went through the specific sites that have been targeted, the
[27:38] Speaker 3: reason why, and just for the audience, Amazon owns and
[27:42] Speaker 3: operates some of those facilities. You've had time to do
[27:46] Speaker 3: your own research and study of the data. How does
[27:49] Speaker 3: this fit in this idea that data centers are a
[27:51] Speaker 3: target in the context of a war in Iran and
[27:54] Speaker 3: a conflict between the United States and Iran.
[27:58] Speaker 16: Sure, yeah, so, thanks so much having me. Datata centers
[28:01] Speaker 16: are becoming increasingly central to a very broad range of
[28:04] Speaker 16: economic life, national security activities, and they are fundamentally soft targets.
[28:08] Speaker 16: So for the Onians looking for targets, so they could
[28:11] Speaker 16: they could cause disruption, could try to broaden the conflict,
[28:14] Speaker 16: could bring the conflict home to the United States. Targeting
[28:16] Speaker 16: a data center that is operated by a US company
[28:19] Speaker 16: that's a symbol or of US Gulf cooperation, it becomes
[28:22] Speaker 16: quite an appealing target given how easy they are to
[28:24] Speaker 16: hit with a drone with a barage and missile strikes.
[28:28] Speaker 16: Given how fragile and vulnerable these facilities.
[28:30] Speaker 3: Are vulnerable, how do you protect a data center.
[28:35] Speaker 2: In this In this case, sure so it's hard.
[28:38] Speaker 16: So fundamentally, these data centers are filled with GPUs, they're
[28:42] Speaker 16: very fragile. It's quite easy to take out a chiller,
[28:45] Speaker 16: which are these things that keep the data centers cool,
[28:47] Speaker 16: keep the servers cool, or to take out a generator
[28:50] Speaker 16: or the transformers that provide electricity to the data center.
[28:53] Speaker 16: There are things you can do. You can use more
[28:55] Speaker 16: reinforced concrete, you can try to harden the facilities. You
[28:57] Speaker 16: can use air defense systems, anti own devices, jamming systems,
[29:02] Speaker 16: an anti anti missile systems. But all of this drives
[29:05] Speaker 16: up the costs of building a data center, and none
[29:07] Speaker 16: of this will guarantee security against a kind of swarm
[29:10] Speaker 16: of cheap drones. And it's very very expensive to protect
[29:12] Speaker 16: against these sorts of cheap strikes. So there are things
[29:15] Speaker 16: you can do to protect them. I think historically these
[29:17] Speaker 16: companies and the governments be much more focused on protecting
[29:19] Speaker 16: data centers from cyber attacks, protecting them from trespasses, but
[29:23] Speaker 16: much less focused on strikes from drones or from missiles,
[29:27] Speaker 16: which you know is an inevitable risk, I think to
[29:30] Speaker 16: building expensive infrastructure in a region where drones and missile
[29:34] Speaker 16: strikes are unfortunately a real risk to worry about.
[29:37] Speaker 5: And you saw this risk.
[29:38] Speaker 4: You wrote about it in July in an opinion column,
[29:41] Speaker 4: really saying that the Golf is not a place to
[29:43] Speaker 4: be building out these data centers, Sam, but also the
[29:46] Speaker 4: Golf is the place longer term with more abundant energy,
[29:50] Speaker 4: and the Golf is the place that also needs AI
[29:53] Speaker 4: for its own citizens. So what do they do?
[29:56] Speaker 16: Yeah, So the Golf has huge ambitions in the AI race,
[29:59] Speaker 16: and they're very appealing partners for the United States for
[30:01] Speaker 16: many reasons, energy capital or the things you mentioned. So
[30:05] Speaker 16: I think it's inevitable that the data centers are going
[30:07] Speaker 16: to be built out in the region. I think this
[30:10] Speaker 16: underscores the risks of building the biggest computing clusters there
[30:13] Speaker 16: the most critical infrastructure. I think that the United States
[30:16] Speaker 16: should have been very carefully about where it wants to
[30:19] Speaker 16: cite that sort of infrastructure. But some degree of data
[30:21] Speaker 16: center construction is inevitable in the growth. But I think
[30:24] Speaker 16: governments need to think. Governments and the companies that operate.
[30:26] Speaker 16: The data centers need to think much more carefully about
[30:28] Speaker 16: their resilience plans, their redundancy plans, how to protect them
[30:31] Speaker 16: from these sorts of physical risks, because as you know,
[30:33] Speaker 16: as you said, this was a foreseeable risk.
[30:35] Speaker 7: These targets are soft.
[30:37] Speaker 16: The more they become symbols of technological might, the more
[30:40] Speaker 16: they become central to economic life, the more appealing they
[30:42] Speaker 16: become as targets for nonstate actors or for hostile states
[30:45] Speaker 16: that want to cause disruption.
[30:46] Speaker 4: Okay, there is a lot that goes into a decision
[30:49] Speaker 4: as to where to put a data center, not only
[30:51] Speaker 4: about the power that's accessible or the water, but the
[30:54] Speaker 4: people and the labor and the costs. SAM Therefore, though
[30:57] Speaker 4: with your Carnegie in talment intational international.
[31:00] Speaker 5: Piece hat on where should these be built?
[31:03] Speaker 16: From a US perspective, so, I think these data centers
[31:06] Speaker 16: are going to be built all over the world.
[31:08] Speaker 7: I think we should be building them all over the world.
[31:10] Speaker 16: I think for the biggest, most important, most critical computing clusters,
[31:15] Speaker 16: the United States should be thinking carefully about siding them
[31:18] Speaker 16: in close US allies, NAO states, within the United States itself,
[31:23] Speaker 16: areas where they can be kind of protective from attacks
[31:25] Speaker 16: or whether it can be more easily protected from attacks.
[31:28] Speaker 7: That's not to say that we shouldn't.
[31:29] Speaker 16: Be building data facilities in the Gulf too, but probably
[31:32] Speaker 16: not our most critical computing clusters. And I would I
[31:35] Speaker 16: would also say that these strone strikes are also underscore.
[31:38] Speaker 16: The data centers all over the world could be a threat.
[31:40] Speaker 16: So even a data center in Europe could well be
[31:42] Speaker 16: targeted by Russian sabotage for example. You know, these these
[31:46] Speaker 16: are fragile sites wherever they're saved.
[31:48] Speaker 3: You know, the argument the Gulf is not the place
[31:51] Speaker 3: to build the world's AI infrastructure. The reality is that
[31:54] Speaker 3: is exactly where they're building the world's AI infrastructure, and
[31:58] Speaker 3: it's where the capital is coming from, and it's where
[32:01] Speaker 3: in VIDEO and AMD have deals to send their chips,
[32:04] Speaker 3: and you have capital flowing from the Gulf into the
[32:07] Speaker 3: United States to fund infrastructure products projects.
[32:10] Speaker 2: In the United States.
[32:12] Speaker 3: I mean, I welcome any response to that reality that
[32:17] Speaker 3: you have.
[32:18] Speaker 16: Yeah, so I think, you know, as I said, I
[32:20] Speaker 16: think that is the reality to some extent, the customer
[32:22] Speaker 16: is toward extent we go forward with these types of projects.
[32:27] Speaker 16: You know, it remains to be seen what this will
[32:29] Speaker 16: look like after this conflict. Because it'sertainly going to drive
[32:31] Speaker 16: up insurance premiums. It may make it harder to attract
[32:33] Speaker 16: engineering talent to build out these dead des centers. I
[32:36] Speaker 16: think this will't make video AE or cider Adia rethink
[32:39] Speaker 16: their plans to any great extent, but I think it
[32:41] Speaker 16: will complicate those plans. I think, you know, it should
[32:45] Speaker 16: make the US government much more cautious about where it's
[32:48] Speaker 16: like its most critical data centers. But again, as I said,
[32:51] Speaker 16: data centers are going to be built everywhere. I think
[32:53] Speaker 16: this attack should be a wake up called that all
[32:55] Speaker 16: of these data centers need to think much more seriously
[32:57] Speaker 16: about physical attacks and not just cyber attacks.
[32:59] Speaker 7: Digitalis are natural disasters.
[33:01] Speaker 5: I wake up.
[33:02] Speaker 16: Cool.
[33:02] Speaker 4: You've been trying to give for at least six months now,
[33:05] Speaker 4: sam Win to Levy, we appreciate your time of Carnegie
[33:07] Speaker 4: and Dowmon for international peace. Coming up, How are supply
[33:11] Speaker 4: chain risks impacting the world's largest automotive supply. I'm a
[33:15] Speaker 4: justly going to hear from the BOSH CEO now, Stephan Herton,
[33:19] Speaker 4: that's next to the Bloomberg Tech.
[33:30] Speaker 3: The conflict in the Middle East is in its seventh
[33:32] Speaker 3: day and with President Trump insisting on unconditional surrender from Iran.
[33:36] Speaker 3: Concerns about the impact on global supply chains arising. I
[33:39] Speaker 3: spoke with BOSH CEO Stephan Horton about what that's meant
[33:42] Speaker 3: for the automotive industry.
[33:44] Speaker 17: So far, we've seen that evs have heard the first wave,
[33:49] Speaker 17: and that was not as big as we all expected,
[33:52] Speaker 17: especially in the US. We saw a less demand than
[33:55] Speaker 17: we actually were hoping for. So now a second wave
[33:58] Speaker 17: comes in with but also pluck in hybrid vehicles or
[34:02] Speaker 17: range extended vehicles. That's a very interesting technology which you
[34:05] Speaker 17: also see in China. So very much all work is
[34:08] Speaker 17: now focused on this combination of electric vehicle technology and
[34:12] Speaker 17: engine technology, and this engine technology is a modern engine
[34:15] Speaker 17: technology with a differentiated engine. So BOSH is doing this
[34:19] Speaker 17: work because we have kept our capability in engine work,
[34:23] Speaker 17: I sees and we can combine these two and do
[34:26] Speaker 17: the plug and hybrid vehicles very important. So I expect
[34:29] Speaker 17: actually demand for these kind of vehicles to be rising
[34:32] Speaker 17: quickly because they are more fuel efficient and same time
[34:35] Speaker 17: they have range, which you want to have and you
[34:37] Speaker 17: can still pull a boat or have a heavy vehicle,
[34:40] Speaker 17: which is practically of use for many people.
[34:43] Speaker 3: But the situation with the war in Iran has any
[34:47] Speaker 3: impact or bearing on that.
[34:49] Speaker 17: Well, I think people will be sensitive always now a
[34:52] Speaker 17: bit more on what gas prices are, how much they
[34:55] Speaker 17: have to spend for the vehicle, what's the vehicle used,
[34:57] Speaker 17: and probably this pluck in hybrid vas or hybrid vehicles
[35:01] Speaker 17: in total with a certain electrical range, which is really useful,
[35:05] Speaker 17: So you can drive electrically for a certain range and
[35:07] Speaker 17: then you'll have to we will switch over to gasoline.
[35:10] Speaker 17: This is something which is for the market the optimal
[35:13] Speaker 17: solution as of now it looks like it. And obviously
[35:17] Speaker 17: people becoming more sensitive about gasoline prices means they will
[35:20] Speaker 17: tend more or not go from a pure ic because
[35:23] Speaker 17: more to a combined so hybrid vehicle or even a
[35:26] Speaker 17: full electric vehicle, that's for sure.
[35:29] Speaker 3: We started this conversation with semiconductors. You talked about the
[35:33] Speaker 3: situation with the start of the year, but also some
[35:37] Speaker 3: parallels with the COVID era and the chip shock, the
[35:40] Speaker 3: chip crisis.
[35:42] Speaker 2: Is this war in Iran different?
[35:45] Speaker 3: Do you see there being a similar level of impact
[35:50] Speaker 3: on semiconductor supply or bottleneck and where are you most
[35:56] Speaker 3: focused within the realm of silicon silicon.
[36:00] Speaker 17: I'm more watching the AI build up because the I
[36:03] Speaker 17: build up, for example, demands high amount of ramps, right, So,
[36:07] Speaker 17: and I also observe that because the same time we
[36:10] Speaker 17: are looking at the ship supply for all applications we do,
[36:15] Speaker 17: but at the same time we're also applying AI in
[36:17] Speaker 17: a massive volume, So that's a very important thing to monitor.
[36:22] Speaker 17: These chips are used in AI data centers to build
[36:26] Speaker 17: the we need to implement solutions like accident prevention in cars,
[36:30] Speaker 17: and that's the business cases which also AI need. So yes,
[36:35] Speaker 17: it is a balance, and yes, there will be probably
[36:38] Speaker 17: a shortage in certain components because the ramp up of
[36:41] Speaker 17: AI infrastructure is massive. On the other hand, these infrastructure
[36:45] Speaker 17: is used also on product we make, so it's used
[36:48] Speaker 17: in Hossald appliances, empower tools in cars, and that's an
[36:51] Speaker 17: important story to go forward.
[36:54] Speaker 3: Is there some panic right now in trying to secure
[36:58] Speaker 3: memory in particular?
[37:00] Speaker 17: No, I wouldn't say there is panic.
[37:01] Speaker 2: It's something we know.
[37:02] Speaker 17: We have seen more severe chip crisis some years ago, right,
[37:07] Speaker 17: But it's something obviously you need to look at because
[37:10] Speaker 17: you also want the functionality, right, you want the I functionality,
[37:14] Speaker 17: You want the better automational vehicles. You want the better
[37:16] Speaker 17: assisted driving because that's safer, it's better than just human driving.
[37:21] Speaker 17: So also in the automation you see probably both. You
[37:24] Speaker 17: see the full automation like in the EV, the full
[37:26] Speaker 17: y V, right, but you see also the hybrid, which
[37:29] Speaker 17: is the massive assisted automation making a driver just the
[37:32] Speaker 17: better driver. And that's based on AI.
[37:34] Speaker 5: Bosh THEO there Se van hearton coming up.
[37:38] Speaker 4: A massive boom in AI data centers is driving demand
[37:41] Speaker 4: for housing and amenities in remote parts of the country.
[37:45] Speaker 4: We'll talk about the state that's on demand, simulating golf.
[37:49] Speaker 5: This is Blueberg tech.
[38:00] Speaker 3: Construction of data centers has led to increase competition for
[38:03] Speaker 3: workers in areas with limited housing and amenities q the
[38:08] Speaker 3: rise of so called AI man camps take a listen.
[38:12] Speaker 2: To power the boom and data center building companies are
[38:15] Speaker 2: heading far outside of Silicon Valley and cities in general.
[38:19] Speaker 18: We're starting to see some pushback from cities where residents
[38:22] Speaker 18: are kind of concerned about what living near a data
[38:25] Speaker 18: center might mean for their property values and for their
[38:27] Speaker 18: way of life. We're seeing projects push out further into
[38:30] Speaker 18: rural stretches of places like Texas and Louisiana because you
[38:33] Speaker 18: can get much cheaper land out there, and you can
[38:35] Speaker 18: also find municipalities that are more willing to work with
[38:38] Speaker 18: you when you need to negotiate things like power arrangements
[38:41] Speaker 18: or zoning changes. While the AI companies are dooking it
[38:45] Speaker 18: out to figure out whose model will win, there are
[38:47] Speaker 18: these other companies that see opportunity just in the build
[38:49] Speaker 18: out itself. There is a huge demand to build data
[38:52] Speaker 18: centers as quickly as they can, so housing becomes a
[38:55] Speaker 18: big bottleneck. Most of these places that these projects are
[38:58] Speaker 18: getting built now just do not have the infrastructure and
[39:00] Speaker 18: that they need for this. Hotels fill right up, our
[39:03] Speaker 18: V parks are completely full with workers. All of a sudden,
[39:06] Speaker 18: there's a line at the grocery store or at the
[39:08] Speaker 18: gas pump. We have a lot of these sort of
[39:09] Speaker 18: single lane county roads and they're not built for rows
[39:13] Speaker 18: and rows of heavy duty trucks driving on them all day,
[39:16] Speaker 18: and so you get roads kind of deteriorating. People are
[39:19] Speaker 18: paying thousands of dollars a month just to rent r
[39:21] Speaker 18: V spaces near construction sites. We're seeing a wide range
[39:24] Speaker 18: of companies get into this, all the way from Wiley
[39:27] Speaker 18: locals that pitch together a little bit of land and
[39:29] Speaker 18: sell the RV spots all the way up to major
[39:31] Speaker 18: publicly traded corporations helping house workers. They've helped with oil
[39:35] Speaker 18: and gas in the shale boom. If you look at
[39:37] Speaker 18: it from above, it looks like an XL spreadsheet cut
[39:40] Speaker 18: out of the middle of farmland. You get this paved
[39:42] Speaker 18: surface where you have rows and rows of gray roofs
[39:45] Speaker 18: and temporary housing. They'll have their own toilet or they'll
[39:49] Speaker 18: have their own sink. They always have Wi Fi so
[39:51] Speaker 18: that you can make video calls to family back home,
[39:53] Speaker 18: and then there'll be a shared communal kind of dining setup.
[39:56] Speaker 18: They place so much empass on the food, not only
[39:58] Speaker 18: because these guys are working so hard all day that
[40:00] Speaker 18: they need the calories, but also because it's an easy
[40:03] Speaker 18: way to attract workers when there's such competition for them.
[40:06] Speaker 18: We'll also see a lot of amenity space. These are
[40:09] Speaker 18: workers that are spending fifteen sixteen hours a day doing
[40:12] Speaker 18: manual labor. You know, I figured when they got home
[40:14] Speaker 18: they'd want to go to sleep, but apparently there are
[40:16] Speaker 18: often gyms so they can get workouts in.
[40:20] Speaker 5: And you can see the.
[40:21] Speaker 4: Rest of the full story from Blomberg Originals on our website,
[40:23] Speaker 4: on YouTube or the terminal, and we want to get
[40:25] Speaker 4: more on this story now, the driving force and the
[40:28] Speaker 4: impact of these data center camps. Bloomberg's Joe Reporter Bloomberg's reporter,
[40:32] Speaker 4: Show Love and we were just seeing you there in
[40:34] Speaker 4: the original's piece and Joe talk to us about some
[40:37] Speaker 4: of the companies that are actually going to be thriving
[40:38] Speaker 4: in this moment, the ones that are actually doing the
[40:40] Speaker 4: building of because people, there's a great quote in your story,
[40:43] Speaker 4: people don't want to be spasically sleeping in their trailer.
[40:47] Speaker 2: That's right, Thank you for having me.
[40:49] Speaker 18: Well, we're seeing companies ranging all the way from smaller
[40:53] Speaker 18: local companies that have been working on you know, oil
[40:57] Speaker 18: fields all the way up to publicly traded corporation where
[41:01] Speaker 18: they're using the same assets that they were for immigration
[41:04] Speaker 18: facilities or for major drilling projects during the sale boom,
[41:09] Speaker 18: and they're using these assets now to build out these
[41:12] Speaker 18: man camps. And so these are companies like Target Hospitality
[41:16] Speaker 18: or Civio, and what they're able to do is reallocate
[41:21] Speaker 18: their resources. You know, until recently, oil was not priced
[41:25] Speaker 18: very very lucatively, so drilling had kind of declined. So
[41:28] Speaker 18: this presented a new opportunity with all of these projects,
[41:32] Speaker 18: especially as they're pushing out further into rural Texas, rural Louisiana,
[41:36] Speaker 18: places that just you know, you show up in a
[41:38] Speaker 18: place like Dickens County, which I wrote about in the story,
[41:41] Speaker 18: you know, you'll find a couple of restaurants, you might
[41:44] Speaker 18: find a motel, but it's just not nearly enough for
[41:47] Speaker 18: a fifteen hundred man workforce that's being brought in to
[41:50] Speaker 18: build these Darta centers.
[41:52] Speaker 2: At a more macro level, the date is quite compelling.
[41:56] Speaker 3: What the story is here, right, jobs growth and the
[41:58] Speaker 3: data center build out? Well, zoom out a bit, you
[42:01] Speaker 3: know what what is direction is this heading in at scale?
[42:07] Speaker 18: Right, We've seen a state like Texas for example. I mean,
[42:10] Speaker 18: if you talk to the leaders, if you talk to
[42:12] Speaker 18: the hyperscalers, they're very bullish on the opportunity for building
[42:15] Speaker 18: data centers in Texas. There's abundant land here, there's abundant energy,
[42:19] Speaker 18: and Texas leaders see this as an opportunity to overtake
[42:22] Speaker 18: Virginia even in the data center space. So you know,
[42:27] Speaker 18: Bloomberg Intelligence estimates that there's seven hundred billion dollars worth
[42:31] Speaker 18: of projects that are just in the planning stage right now.
[42:33] Speaker 18: And I don't need to tell you artificial intelligence it's
[42:37] Speaker 18: growing at a pace that really is hard to keep,
[42:39] Speaker 18: you know, keep tabs on. So these companies that I've
[42:43] Speaker 18: spoken with, you know, they see a very, very high
[42:46] Speaker 18: ceiling for where this buildout can go.
[42:49] Speaker 4: I mean, the number of Bloomberg Intelligent is saying seven
[42:51] Speaker 4: hundred billion dollars worth of projects in the planning stage,
[42:53] Speaker 4: one hundred and sixty billion already underway. But Joe, the
[42:56] Speaker 4: word camps kind of says it all in that these
[43:00] Speaker 4: aren't permanent housing.
[43:01] Speaker 5: Briefly, it's not long term jobs.
[43:05] Speaker 2: That's right.
[43:05] Speaker 18: And when you speak with local officials, even in a
[43:09] Speaker 18: place like Dickens County that has not seen a project
[43:12] Speaker 18: nearly of this scale, they're aware that these jobs are
[43:16] Speaker 18: not going to stick around forever and that while you
[43:18] Speaker 18: might have some people that stick around to work on
[43:20] Speaker 18: the data center once it's built, most of them are
[43:24] Speaker 18: going to have to get out of town. So you know,
[43:26] Speaker 18: it's temporary.
[43:28] Speaker 3: Bloomberg, Jo Lovenger, great reporting in video, imprint.
[43:32] Speaker 2: Great to have you on the program.
[43:33] Speaker 4: Charroc, Yeah, that does it for an extraordinary week on
[43:36] Speaker 4: Bloomberg chech Head.
[43:37] Speaker 3: Yeah, a lot to recap. Do it on the podcast.
[43:40] Speaker 3: You know where to find it. It's all on the
[43:41] Speaker 3: screen there but Spotify, iHeart and Apple.
[43:44] Speaker 2: 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.
