Latest / Bloomberg Intelligence / Nvidia’s $750 Billion in Deals Reignite Circular AI Fears

Transcript

Nvidia’s $750 Billion in Deals Reignite Circular AI Fears · Bloomberg Intelligence

• 0:02
  Speaker 1: Bloomberg Audio Studios, podcasts, radio news. You're listening to the
• 0:10
  Speaker 1: Bloomberg Intelligence Podcast. Catch us live weekdays at ten am
• 0:14
  Speaker 1: Eastern on Apple Coarcklay and Android Auto with the Bloomberg
• 0:17
  Speaker 1: Business App. Listen on demand wherever you get your podcasts,
• 0:21
  Speaker 1: or watch us live on YouTube.
• 0:24
  Speaker 2: In videos, working on a fresh round of AI infrastructure
• 0:26
  Speaker 2: deals potentially worth more than seven hundred and fifty billion dollars,
• 0:30
  Speaker 2: including an AI initiative with s k Heinez for almost
• 0:33
  Speaker 2: half a trillion dollars. The CEO of Nvidia, Jensen Hong,
• 0:36
  Speaker 2: spoke in an exclusive interview with Ed Ludlow about investing
• 0:40
  Speaker 2: in South Korea. Take a listen.
• 0:41
  Speaker 3: This is the golden ages for Korea, as you know,
• 0:44
  Speaker 3: their semiconductor businesses booming, their industrial businesses booming. You know,
• 0:49
  Speaker 3: this is a country that has the ability to help
• 0:51
  Speaker 3: the world build out the AI infrastructure that they're incredibly
• 0:55
  Speaker 3: adapt at adopting new technologies and as a real technologically
• 1:01
  Speaker 3: forward learning society.
• 1:03
  Speaker 4: All right, that was it.
• 1:04
  Speaker 5: In videos CEO Jensen one and joining us now is
• 1:07
  Speaker 5: Betech anchor Ed Ludlow ed. The numbers just keep coming,
• 1:11
  Speaker 5: the investments just keep coming. The circular deals keep coming.
• 1:17
  Speaker 5: What's the view of that type of structure as this
• 1:20
  Speaker 5: AI infrastructure evolves. What's it feeling out in Silicon Valley
• 1:23
  Speaker 5: about how this is evolving.
• 1:25
  Speaker 6: Part of the reason there is even a debate about
• 1:27
  Speaker 6: it being circular not is that there isn't a great
• 1:29
  Speaker 6: explanation of what the numbers actually represent. So you've done
• 1:33
  Speaker 6: the reporting about you know what Dean has written on
• 1:35
  Speaker 6: the terminal on open AI. If we just take SK
• 1:37
  Speaker 6: as a case study, it's five hundred billion dollars or
• 1:40
  Speaker 6: half a trillion dollars. And at some point in that
• 1:43
  Speaker 6: conversation with Jensen, I just said, could you just explain
• 1:45
  Speaker 6: what that number actually is and he was like, well, sure,
• 1:48
  Speaker 6: it's a number that represents what we and Vidia are
• 1:51
  Speaker 6: going to be buying from sk Heinex in memory chips,
• 1:54
  Speaker 6: but it also represents what SK as a group a conglomerate,
• 1:58
  Speaker 6: will be spending on their own AI infrastruc ructure and
• 2:00
  Speaker 6: then everything in between that we co invest in. And
• 2:02
  Speaker 6: there's like no real explanation of like, okay, so you're
• 2:05
  Speaker 6: giving them dollars or are they giving you dollars or
• 2:08
  Speaker 6: are you basically just tallying it up and calling it
• 2:10
  Speaker 6: even at the end of the day. And that's a
• 2:12
  Speaker 6: big part of it, right that people just don't know
• 2:14
  Speaker 6: the direction or travel of those dollars. In the open
• 2:17
  Speaker 6: AI case, what reporting is much more explicit, where you
• 2:20
  Speaker 6: know in videas both both guaranteeing and backstopping but also
• 2:23
  Speaker 6: providing financing for AI data centers that at the end
• 2:27
  Speaker 6: of the day they do use in Video's technology and
• 2:29
  Speaker 6: that is by definition circular.
• 2:32
  Speaker 2: Okay. So going back to that deal or agreement or
• 2:35
  Speaker 2: partnership with Skehiynix worth more than five hundred billion, do
• 2:39
  Speaker 2: we know if that ever actually happens. I mean, it's
• 2:41
  Speaker 2: one thing to announce that you have these grand plans,
• 2:43
  Speaker 2: it's another thing for the two companies to execute on it.
• 2:47
  Speaker 6: So it's over many years, it's over time, but there's
• 2:50
  Speaker 6: no set deadline or timeframe. It is literally bucket one
• 2:54
  Speaker 6: in video buying memory chips from sk Heinix. Sk Heinix
• 2:57
  Speaker 6: has sixty percent market share in high bandwidth memory, but
• 3:00
  Speaker 6: sk Group with SKA Telecom underneath it is a big
• 3:04
  Speaker 6: deployer of data center capacity. And where do they buy
• 3:08
  Speaker 6: their not just GPUs, but their their servers from now
• 3:13
  Speaker 6: from Nvidia. So that is it's a multi year thing.
• 3:16
  Speaker 6: It won't show up in any one company's one year financials.
• 3:20
  Speaker 6: But let me just be honest with you Scarlet about it.
• 3:23
  Speaker 6: Jensen and in video they like big round numbers, and
• 3:29
  Speaker 6: it's really normal for them to announce it that in
• 3:31
  Speaker 6: that way.
• 3:33
  Speaker 5: What do you think Silicon Valley is going to be
• 3:36
  Speaker 5: really listening for this week when we get some of
• 3:40
  Speaker 5: these big tech companies reporting on Wednesday and Thursday.
• 3:43
  Speaker 1: Yeah.
• 3:43
  Speaker 6: So it was a really simple arrangement between Wall Street
• 3:46
  Speaker 6: and the biggest technology companies this point, the biggest technology
• 3:50
  Speaker 6: companies had to say that they were going to spend
• 3:52
  Speaker 6: more every year, CAPEX would go up, and they needed
• 3:55
  Speaker 6: to show top line growth. And if they did that,
• 3:57
  Speaker 6: I would be on the show of View the next
• 3:59
  Speaker 6: morning and the big green on the screen and everything
• 4:01
  Speaker 6: would be rosy. Now there's a lot more scrutiny. There
• 4:04
  Speaker 6: needs to be much more evidence, not just top line growth,
• 4:07
  Speaker 6: that all of the rising CAPEX results in something material.
• 4:12
  Speaker 6: And we learned that lesson through alphabet right, so with Amazon,
• 4:15
  Speaker 6: Microsoft in particular, but also Meta to a lesser extent.
• 4:18
  Speaker 6: People want to say, like, okay, I understand this metric
• 4:22
  Speaker 6: about how people using the AI that you've developed, not
• 4:25
  Speaker 6: just that your sales are growing. You know, Alphabet showed
• 4:27
  Speaker 6: really strong growth, but the market was still worried about
• 4:31
  Speaker 6: the rising CAPEX number.
• 4:33
  Speaker 2: Before we let you go very quickly, there was a
• 4:35
  Speaker 2: Chinese chip maker that made a debut and in its debut,
• 4:38
  Speaker 2: cx empty jumped. I'm looking at this, it looks like
• 4:41
  Speaker 2: a type of four hundred and sixty six percent.
• 4:43
  Speaker 6: Yeah, and now is Mainland China's most valuable company overnight.
• 4:47
  Speaker 6: What you need to know is cx empt is the
• 4:49
  Speaker 6: number four maker of memory chips, but it is the
• 4:52
  Speaker 6: case study for China wanting its own domestic industry. It's
• 4:55
  Speaker 6: the flagship for China finding a domestic hero. It has
• 4:59
  Speaker 6: a long way to go to catch up, but the
• 5:01
  Speaker 6: proceeds of this IPO will help it to establish memory
• 5:04
  Speaker 6: manufacturing in China. And it's also an international success, so
• 5:09
  Speaker 6: those other memory names will be looking over their shoulder.
• 5:11
  Speaker 1: Stay with us.
• 5:12
  Speaker 2: More from Bloomberg Intelligence coming up after this.
• 5:18
  Speaker 1: You're listening to the Bloomberg Intelligence Podcast. Catch us live
• 5:22
  Speaker 1: weekdays at ten am Eastern on Apple, Cocklay and Android
• 5:25
  Speaker 1: Auto with the Bloomberg Business app. Listen on demand wherever
• 5:28
  Speaker 1: you get your podcasts, or watch us live on YouTube.
• 5:33
  Speaker 5: Next one to bring in Laura Martin. She's a senior
• 5:35
  Speaker 5: analyst that need them a company. She's out there in
• 5:37
  Speaker 5: la I've known Laura for decades. She's covered the media
• 5:39
  Speaker 5: industry for decades, and what is characterized her work from
• 5:42
  Speaker 5: my perspective is just a level of innovation in her research.
• 5:46
  Speaker 5: She looks at companies and industries very differently than most
• 5:49
  Speaker 5: of the street. And as an example, I'm going to
• 5:51
  Speaker 5: say twenty five years ago, maybe even more, she really
• 5:54
  Speaker 5: put return on invested capital front and center for her
• 5:57
  Speaker 5: research and I didn't know anybody else doing that back
• 5:59
  Speaker 5: in the day. That was really unique and that drove
• 6:02
  Speaker 5: a lot of really cool research back in the day. Now,
• 6:05
  Speaker 5: I think she takes it one step further.
• 6:07
  Speaker 3: She uses the.
• 6:08
  Speaker 5: Rating some glass door which kind of looks at the
• 6:12
  Speaker 5: culture and values of its employees across her research coverage,
• 6:17
  Speaker 5: and she says, you know what if happy employees are
• 6:20
  Speaker 5: value employees and the create shareholder value, and that's one
• 6:23
  Speaker 5: of the metrics you should look at. Lauren, thanks so
• 6:25
  Speaker 5: much for joining us here, talk to us about this research,
• 6:28
  Speaker 5: and you guys highlighted Meta recently using this type of research.
• 6:33
  Speaker 7: Yeah, So what we're doing is we're looking at the
• 6:35
  Speaker 7: fact that out of the twenty three stocks we cover,
• 6:38
  Speaker 7: Meta pays the most in stock based compensation per employee.
• 6:43
  Speaker 7: Both are metrics you can pull out of their public financials.
• 6:46
  Speaker 7: So they're paying every single employee a mat at which
• 6:49
  Speaker 7: is like seventy thousand people three hundred thousand dollars each.
• 6:53
  Speaker 7: So presumably some people like receptionists, don't get stock, which
• 6:56
  Speaker 7: means they're probably playing people who do get stock five
• 6:59
  Speaker 7: hundred thousand dollars in stock based comp So you say
• 7:02
  Speaker 7: to yourself, why are they paying more a lot more
• 7:05
  Speaker 7: than everybody else in big tech and in big cap.
• 7:08
  Speaker 7: And when you look, you look at the answer. Out
• 7:10
  Speaker 7: of the twenty three companies we cover, their ratings by
• 7:13
  Speaker 7: employees today is a two point eight out of five,
• 7:17
  Speaker 7: which is the lowest of all the companies we cover
• 7:19
  Speaker 7: by a lot, Like Netflix is four point four, Google
• 7:24
  Speaker 7: four point two, and you know, Metas down at two
• 7:28
  Speaker 7: point eight.
• 7:29
  Speaker 2: So that's a huge different of.
• 7:30
  Speaker 7: The reasons they have to pay so much is because
• 7:32
  Speaker 7: their culture and values are really low and horrible internally.
• 7:36
  Speaker 2: Okay, say more about this. What do we know about
• 7:39
  Speaker 2: the culture and the values. I mean, you dug into
• 7:41
  Speaker 2: all of this, what did you find?
• 7:43
  Speaker 7: So I think the problem is Mark. You know, the
• 7:46
  Speaker 7: CEO is unimpeachable. He controls the company absolutely, and you've
• 7:50
  Speaker 7: seen him pivot really rapidly from the metaverse. Oh, change
• 7:55
  Speaker 7: the company name, but now we're firing all those people.
• 7:57
  Speaker 7: Quest Goggles were really in you know, Vogue for a
• 8:00
  Speaker 7: year and now they're sort of getting left by the wayside.
• 8:03
  Speaker 7: So now we're doing Now he's doing generative AI. But
• 8:06
  Speaker 7: three weeks ago he said we're getting into the chip
• 8:08
  Speaker 7: business and we're going to compete with Navidia. So I
• 8:11
  Speaker 7: think just this strategic pivoting really demoralizes people and confuses
• 8:16
  Speaker 7: them about what the strategic priorities are of the enterprise.
• 8:20
  Speaker 5: So the net result is to I guess, attract and retain,
• 8:23
  Speaker 5: they have to pay more. And in Silicon Valley that
• 8:27
  Speaker 5: means stock. That's not like Laura and I grew up
• 8:30
  Speaker 5: in Wall Street where it was just cash cash cash out.
• 8:32
  Speaker 5: There it's all stock. What's the long term impact do
• 8:36
  Speaker 5: you think on their equity value their stock performance?
• 8:40
  Speaker 7: I think it means you have higher employee churn, and
• 8:44
  Speaker 7: the people who churn are the people who can get
• 8:45
  Speaker 7: the next job, which means they're good people. So they're
• 8:48
  Speaker 7: leaving for Open AI or Anthropic or Google, and so
• 8:52
  Speaker 7: that makes to get the new person to replace that
• 8:55
  Speaker 7: churned out person is just more and more expensive. And
• 8:59
  Speaker 7: that's bad for erald because it's deluded to shareholders if
• 9:01
  Speaker 7: you have to pay people more to attract them and
• 9:04
  Speaker 7: retain them. So I think it's bad for it. It's bad
• 9:07
  Speaker 7: for both. I mean, both employees and Wall Street need
• 9:11
  Speaker 7: stock prices to rise, so it's bad for both employees
• 9:14
  Speaker 7: and Wall Street to have them have to give so
• 9:17
  Speaker 7: much equity to attract and retain people.
• 9:20
  Speaker 2: I think, what does it mean for the company's ability
• 9:23
  Speaker 2: to deliver on these new strategic initiatives? I mean, if
• 9:25
  Speaker 2: Mark Zuckerberg is pivoting the company every six months every
• 9:29
  Speaker 2: year to something new, he's paying up for the talent,
• 9:32
  Speaker 2: But can they execute if the I mean, if there
• 9:35
  Speaker 2: isn't that allegiance between company and employee the way you
• 9:38
  Speaker 2: might find elsewhere.
• 9:40
  Speaker 7: Right exactly. So one of the things he just said
• 9:42
  Speaker 7: in a town hall that he's going to lease capacity.
• 9:45
  Speaker 7: He just did a ten billion dollar leasing deal where
• 9:47
  Speaker 7: he has built too much capacity, because he said his
• 9:50
  Speaker 7: AI initiatives aren't going as quickly is he had hoped. Well,
• 9:55
  Speaker 7: that's exactly the answer to your question that the employees
• 9:58
  Speaker 7: are either infighting or it's clear what priorities are. Things
• 10:01
  Speaker 7: are happening slower in it, regardless of despite the amount
• 10:06
  Speaker 7: of money he's spending on talent.
• 10:08
  Speaker 5: Lauri, you're one of there's seventy two buy ratings on
• 10:10
  Speaker 5: his company, only seven holds you're one of them. Why
• 10:13
  Speaker 5: is that?
• 10:15
  Speaker 7: Because I think he is destroying value by not having
• 10:19
  Speaker 7: strategic focus that I think he's got. Quest was about
• 10:23
  Speaker 7: trying to go into competition with Apple. He's in the
• 10:26
  Speaker 7: shopping business because he's trying to displace Amazon. Of course
• 10:29
  Speaker 7: he does search because right so that tries to he
• 10:32
  Speaker 7: really is next to search, social and search. He is
• 10:35
  Speaker 7: the number one competitor to Google. And now he's going
• 10:38
  Speaker 7: in the chip business with competes with Navidia. So I
• 10:40
  Speaker 7: feel like his strategy is reactionary and keeps going into business,
• 10:46
  Speaker 7: you know, sort of trying to compete with the incumbent
• 10:49
  Speaker 7: winner take all you know, participant, and I just think
• 10:52
  Speaker 7: that won't work in the end. He's much smaller many
• 10:54
  Speaker 7: of these companies he's going into competition against. I just
• 10:57
  Speaker 7: think it doesn't work in the end.
• 10:59
  Speaker 2: I think stay with us. More from Bloomberg Intelligence coming
• 11:02
  Speaker 2: up after this.
• 11:07
  Speaker 1: You're listening to the Bloomberg Intelligence podcast. Catch us live
• 11:11
  Speaker 1: weekdays at ten am Eastern on Apple, Cocklay and Android
• 11:14
  Speaker 1: Auto with the Bloomberg Business app. Listen on demand wherever
• 11:18
  Speaker 1: you get your podcasts, or watch us live on YouTube.
• 11:21
  Speaker 5: All right, let's get away from the AI talk. Let's
• 11:24
  Speaker 5: go to the big take story out there. And it's
• 11:27
  Speaker 5: a good one because it's a non AI big take
• 11:30
  Speaker 5: story here. It goes about the tin can industry here,
• 11:34
  Speaker 5: and it goes to tariffs and protectionism, the industry's struggle.
• 11:39
  Speaker 5: That's the tin can industry struggle is a cautionary tale
• 11:42
  Speaker 5: as the US President tries to rebuild his protectionist resume.
• 11:47
  Speaker 5: This story when deeply diving into the tin can business,
• 11:51
  Speaker 5: which I don't even think about the tin can business,
• 11:53
  Speaker 5: but apparently we make a lot of tin cans every year.
• 11:55
  Speaker 1: Yeah.
• 11:55
  Speaker 5: Sean Donna Joins is here Bloomberg News. He's a senior
• 11:58
  Speaker 5: economics writer. Sean talk to us about the tin can
• 12:02
  Speaker 5: business in this country and how is it being impacted
• 12:04
  Speaker 5: by terrffs.
• 12:06
  Speaker 4: Sure, well, thanks for having me. Look, the tin can
• 12:09
  Speaker 4: is this wonderful little case study that we can take
• 12:14
  Speaker 4: into what has happened actually with tariffs and the results.
• 12:18
  Speaker 4: I think when we talk about tariffs nowadays, we often
• 12:21
  Speaker 4: talk about the promise that this will mean a return
• 12:24
  Speaker 4: in manufacturing, the promise that lots of jobs will come
• 12:28
  Speaker 4: with it. Eventually that will be a huge amount of
• 12:30
  Speaker 4: investment in the United States. And the tin cans are
• 12:37
  Speaker 4: an example that you kind of give us a bit
• 12:38
  Speaker 4: of pause when we hear some of those promises. And
• 12:41
  Speaker 4: I dove into this because in March of twenty eighteen,
• 12:45
  Speaker 4: Wilbur Ross went on TV and this was right as
• 12:49
  Speaker 4: they announced really the first Trump tariffs, which were the
• 12:52
  Speaker 4: tariffs on steel, and the markets were going down. Investors
• 12:57
  Speaker 4: were nervous, and Wilbur Ross went on and he held
• 13:01
  Speaker 4: up a Campbell's soup can and he said, you know what,
• 13:05
  Speaker 4: there's only two point six cents worth of steel in this,
• 13:08
  Speaker 4: and these tariffs aren't really gonna affect it and the
• 13:12
  Speaker 4: price of this, and moreover, we're going to get tens
• 13:15
  Speaker 4: of thousands of jobs and hundreds of millions of dollars
• 13:18
  Speaker 4: in investment. Well, he went out that day and he
• 13:23
  Speaker 4: bought a can of Campbell's. It was chicken soup, chicken
• 13:26
  Speaker 4: noodle soup, and he said he went as Local seven
• 13:28
  Speaker 4: eleven bought it for a dollar ninety nine. So I thought,
• 13:30
  Speaker 4: I did the same thing as Wilbur Ross, and I
• 13:32
  Speaker 4: went out to my local seven to eleven and I
• 13:34
  Speaker 4: bought that can of chicken noodle soup and it would
• 13:36
  Speaker 4: cost me three forty nine. Right, So that's not a
• 13:38
  Speaker 4: surprise that in this kind of era of inflation, price
• 13:42
  Speaker 4: has gone up. But it's why the price has gone up.
• 13:45
  Speaker 4: And part of the reason for that is there is
• 13:47
  Speaker 4: a fifty percent tariff on the steel that is used
• 13:51
  Speaker 4: to make tin cans. And then we started looking at okay,
• 13:54
  Speaker 4: has there been investment in making that steel that is
• 13:58
  Speaker 4: used in tin cans? Well, the answer is no, Actually
• 14:01
  Speaker 4: it's gone the opposite way. There were a dozen mills
• 14:05
  Speaker 4: that made ten plate steel when will Barross held up
• 14:08
  Speaker 4: that can. There's now three running in the United States.
• 14:11
  Speaker 4: In twenty eighteen, the US imported fifty percent of the
• 14:16
  Speaker 4: steel that was used in to make ten cans. Now
• 14:19
  Speaker 4: it imports eighty percent of that stike just because domestic
• 14:22
  Speaker 4: production has not gone anyway, and we just haven't seen
• 14:26
  Speaker 4: the jobs they're now in all of the steel mills
• 14:28
  Speaker 4: in the United States. They are now only about thirteen
• 14:31
  Speaker 4: hundred more people working in those mills than there were
• 14:34
  Speaker 4: in twenty eighteen. So it's you know, it's one of
• 14:37
  Speaker 4: the stories of I think of it as a story
• 14:39
  Speaker 4: of kind of unfulfilled promises, unintended consequences, and that, you know,
• 14:47
  Speaker 4: the promise of tariffs. What what actually happens when they
• 14:50
  Speaker 4: got Now this is just ten cans, right, this is
• 14:52
  Speaker 4: just ten cans part of the can.
• 14:54
  Speaker 2: What is happening to those ten cans because they're made
• 14:56
  Speaker 2: somewhat differently now than they were before, aren't they?
• 15:00
  Speaker 4: Yeah, this is really interesting. So I went up to
• 15:01
  Speaker 4: the Can Corporation of America, which is just outside Allentown, Pennsylvania,
• 15:05
  Speaker 4: and if you go there on the production line, they
• 15:07
  Speaker 4: will tell you that they use roughly the same amount
• 15:09
  Speaker 4: of steel that they did in twenty eighteen, and they
• 15:15
  Speaker 4: turn out more cans. Why because that tin can that
• 15:18
  Speaker 4: you're buying in the store is actually thinner than it
• 15:21
  Speaker 4: used to be. They're actually reducing the amount of steel
• 15:24
  Speaker 4: in each tin can to try and save on costs.
• 15:27
  Speaker 4: So we've had some kind of innovation there. That steel
• 15:31
  Speaker 4: is also, as I said before, it's more likely to
• 15:35
  Speaker 4: be coming from overseas.
• 15:37
  Speaker 5: And I love this stet here. Each year Can Corporation
• 15:40
  Speaker 5: and it's three hundred and fifty employees turn out just
• 15:42
  Speaker 5: shy of one billion in cans and two hundred different
• 15:46
  Speaker 5: sizes that are destined to be filled with everything from
• 15:49
  Speaker 5: coffee to industrial adhesives.
• 15:50
  Speaker 2: You know, I think about this and like, I can't
• 15:52
  Speaker 2: remember the last time I used a can opener because
• 15:54
  Speaker 2: so much soup, for instance, broth is always in the
• 15:56
  Speaker 2: paper packaging. Now instead of cannon, I go Campbell's Soup,
• 15:59
  Speaker 2: big time New Jersey Push. That's why.
• 16:04
  Speaker 4: Sorry. And if you talk to people in the can industry,
• 16:07
  Speaker 4: they will kind of look at you very skeptically and say,
• 16:11
  Speaker 4: there's nothing quite like a tin can in terms of
• 16:14
  Speaker 4: holding up and the ability to stack it and and
• 16:18
  Speaker 4: so on. But yeah, that has been one of the
• 16:20
  Speaker 4: threats to the can industry has been kind of alternative
• 16:24
  Speaker 4: packaging and some of that which has been prompted by
• 16:27
  Speaker 4: the higher costs of cans.
• 16:29
  Speaker 1: And this is the Bloomberg Intelligence podcast, available on Apple, Spotify,
• 16:35
  Speaker 1: and anywhere else you get your podcasts. Listen live each weekday,
• 16:39
  Speaker 1: ten am to noon Eastern on Bloomberg dot com, the
• 16:42
  Speaker 1: iHeartRadio app tune In, and the Bloomberg Business app. You
• 16:46
  Speaker 1: can also watch us live every weekday on YouTube and
• 16:50
  Speaker 1: always on the Bloomberg terminal