# Nvidia Earnings This Week; Biggest Power Deal in History

An episode of [Bloomberg Tech](https://invisiblenetwork.ai/podcast/1611767434), 18 May 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/dd0a325c-da9c-4ba1-b83f-b44e01476634/audio.mp3?utm_source=Podcast&in_playlist=9caafc41-289c-4115-995d-ae7c0030ab75

Bloomberg’s Caroline Hyde discusses what to expect from Nvidia this week as investors anxiously await the company's earnings results. Plus, the AI data center boom triggers the largest power deal in history, the filing for SpaceX's IPO could come as soon as this week, and Ed Ludlow brings the latest from Dell Technologies World in Las Vegas. See omnystudio.com/listener for privacy information.

## Transcript

[0:02] Speaker 1: Bloomberg Audio Studios, Podcasts, radio news. Bloomberg Tech is live
[0:13] Speaker 1: from coast to coast with Caroline Hide in New York
[0:17] Speaker 1: and Eva Loo in San Francisco.
[0:21] Speaker 2: This is Bloomberg Tech coming up.
[0:23] Speaker 3: All eyes on Nvidia this week as investors anxiously await
[0:26] Speaker 3: the company's earnings results, plus the AI data center Boom
[0:29] Speaker 3: triggers the largest power deal in history with next Era
[0:33] Speaker 3: acquiring Dominion Energy for sixty seven billion dollars, and Elon
[0:37] Speaker 3: Musk says he's in Texas working on plans for the
[0:39] Speaker 3: space XIPO.
[0:40] Speaker 2: The filing come as soon as this week.
[0:43] Speaker 3: But first we check in on these markets, which again
[0:46] Speaker 3: are sinking for a second straight day. Like the NASA
[0:48] Speaker 3: one hundred under pressure on Friday, it was around anxiety
[0:51] Speaker 3: or underwhelmed between the US China relations. Now it's front
[0:54] Speaker 3: and center back on geopolitics and the US around conflict.
[0:57] Speaker 3: We're off by five tenths of a percent as all
[0:59] Speaker 3: pushes higher, but we're seeing the Semiconductor index once again
[1:02] Speaker 3: taking a pullback.
[1:03] Speaker 2: We're down by two point two percent.
[1:04] Speaker 3: In fact, we've lost about six percent on this index
[1:07] Speaker 3: in the last two training days as we saw a
[1:09] Speaker 3: significant sell off on Friday, it ramps up into another day.
[1:13] Speaker 4: Now.
[1:13] Speaker 3: All of this is why we still try to digest
[1:15] Speaker 3: the reality of the AI demand and spend, and all
[1:18] Speaker 3: eyes are going to be over in Las Vegas on
[1:21] Speaker 3: Dell World today. This is as we think about how
[1:24] Speaker 3: Dell is putting on its annual conference with developers, and
[1:27] Speaker 3: we're trying to understand what new tech, new innovations, what
[1:30] Speaker 3: new offerings are on board. But there's a very special
[1:32] Speaker 3: conversation we had a bit later. Ed Ludlow's standing by
[1:35] Speaker 3: and look, you're going to speaking to the men of
[1:37] Speaker 3: the moment a little bit later.
[1:41] Speaker 5: Yeah, And there's a lot that needs to be reconciled.
[1:44] Speaker 6: Right.
[1:44] Speaker 5: This is the first time that the Gensen one will
[1:46] Speaker 5: be speaking, having comeback from China, and what needs to
[1:49] Speaker 5: be reconciled is that in March, at in Vidia's own conference,
[1:53] Speaker 5: Gensen one said that they were ramping up their supply
[1:56] Speaker 5: chain for H two hundred. The US had approved the
[1:59] Speaker 5: licenses they could do business in that market. The president
[2:03] Speaker 5: on Friday at board Air Force one said that he
[2:05] Speaker 5: did discuss H two hundreds was ujingping, but China does
[2:09] Speaker 5: not want them, and so now we're in this holding
[2:11] Speaker 5: pattern where we need to explain to the market what's
[2:15] Speaker 5: going on, because a big part of the trade Friday
[2:17] Speaker 5: and a little bit into this morning's trade has been
[2:20] Speaker 5: this idea that nothing concrete came out of President Trump's
[2:23] Speaker 5: visits to China with a video kind of the center
[2:26] Speaker 5: of it. But generally speaking, so many supply chain questions
[2:29] Speaker 5: post melt up in the AI infrastructure trade.
[2:32] Speaker 3: Yeah, maybe we take a breather in that melt up
[2:35] Speaker 3: just on the day ed or a second straight day,
[2:37] Speaker 3: But meanwhile, the enthusiasm doesn't seem to be particularly being
[2:40] Speaker 3: curbed for in video. If you're looking, some of the
[2:41] Speaker 3: analysts knows today.
[2:46] Speaker 5: Yeah, it's fun to just hit fa go on the
[2:48] Speaker 5: Bloomberg terminal and just go back to basics right when
[2:50] Speaker 5: they had their print on Wednesday. EPs growth above eighty percent,
[2:55] Speaker 5: top line growth just below eighty percent. And what's so
[2:58] Speaker 5: fascinating about being here at Deltagies World is that if
[3:01] Speaker 5: you cook back at history of how these companies used
[3:04] Speaker 5: to do business, right, Nvidio never really sort of had
[3:07] Speaker 5: Gensen one fly around the world and deliver chips directly
[3:10] Speaker 5: to the end user. Dell was the kind of key
[3:13] Speaker 5: sales channel and so in the era of more agenta
[3:17] Speaker 5: KI where more enterprises are investing to have the compute
[3:21] Speaker 5: capacity they need. It's a really interesting lens. The timing
[3:26] Speaker 5: is curious, of course, right speaking Monday in video reports
[3:29] Speaker 5: earnings Wednesday, but they are very key partners in the
[3:33] Speaker 5: state of play for AI servers what the AI factory
[3:36] Speaker 5: but data center for running AI workloads. And you know,
[3:41] Speaker 5: this is kind of his playbook, right Carol, Let's be
[3:43] Speaker 5: honest about it. You know, Jensen does tend to appear
[3:46] Speaker 5: alongside his partners often. But I think we'll learn a
[3:49] Speaker 5: lot today about the state of the world, the demand levers,
[3:51] Speaker 5: and also the issues with supply.
[3:53] Speaker 2: So much to learn ed.
[3:54] Speaker 3: Ludlow is going to be with us a little bit
[3:56] Speaker 3: later in the show, but tune in later today. It's
[3:58] Speaker 3: going to be an exclusive intero with the CEOs, as
[4:01] Speaker 3: he says, of Dell of Nvidia, live from Dell Technologies World.
[4:05] Speaker 3: That's a two thirty pm Eastern eleven thirty am Pacific time.
[4:08] Speaker 3: But as we think ahead to Wednesday and videos earnings
[4:11] Speaker 3: out after the Bell, I want to bring you what
[4:12] Speaker 3: to expect with Manday seeing Bloomberg Intelligence and Mandy. We
[4:15] Speaker 3: just heard a little bit about the drivers the trends
[4:18] Speaker 3: from ED, but get into the nitty gritty on Nvidia.
[4:21] Speaker 3: Is there as much enthusiasm as to really what the
[4:24] Speaker 3: CPUs are bringing to bear as well as the GPUs
[4:26] Speaker 3: and the infrastructure.
[4:28] Speaker 7: Yeah. Look, I mean when a company is growing top
[4:31] Speaker 7: line the way in Vidio is growing, and that top
[4:33] Speaker 7: line growth seems to be accelerating.
[4:36] Speaker 8: It just there's no doubt that, you know, they are
[4:39] Speaker 8: capitalizing on the demand and the chips remain in short supply.
[4:45] Speaker 8: The only thing I would caution on is you know,
[4:48] Speaker 8: the rising memory prices, so we still have to see
[4:52] Speaker 8: the impact of that on in videos margins, and they
[4:55] Speaker 8: have talked about, you know, maintaining those mid seventy percent
[4:57] Speaker 8: growth margins and the other the thing is there's been
[5:01] Speaker 8: that big shift from training to more inferencing and reasoning.
[5:06] Speaker 8: And we know Nvidia almost had a monopoly when it
[5:10] Speaker 8: comes to training workloads, but when it comes to infrincing
[5:14] Speaker 8: and now with the rise of Entropic, you know, and
[5:18] Speaker 8: how much traction Entropic has had with coding agents, it
[5:21] Speaker 8: remains to be seen if Nvidia still has that bulk
[5:25] Speaker 8: of the share when it comes to the use of
[5:27] Speaker 8: their chips, and I think that's something we'll try to
[5:30] Speaker 8: ascertain with the earnings call.
[5:33] Speaker 3: I mean, it may move is moves in inference. I
[5:36] Speaker 3: think about the acquisition of Grock with a queue not
[5:38] Speaker 3: like k but Mandie talk to us about what we
[5:40] Speaker 3: might hear about that partnership, about what they're building to
[5:43] Speaker 3: serve the inference market.
[5:45] Speaker 7: Absolutely, and you bring up a very good point that
[5:47] Speaker 7: they have talked about, you know, Grock being a very
[5:50] Speaker 7: big driver on the infrincing side to the tune, you know,
[5:54] Speaker 7: north of twenty billion dollars just from that acquisition. What
[5:59] Speaker 7: they're doing integrating that acquisition. And look, I mean we
[6:03] Speaker 7: know Entropic has also partnered with Nvidia now and Anthropic
[6:08] Speaker 7: is the one that is had the most traction you know,
[6:11] Speaker 7: this year with coding agents. So from that perspective, some
[6:16] Speaker 7: of those influencing workloads will be deployed on Brock. And
[6:19] Speaker 7: to your point, you know, AGENTAI does require a lot
[6:23] Speaker 7: more CPUs, which is why we've seen you know the
[6:26] Speaker 7: likes of Intel and AMD do very well. So it
[6:30] Speaker 7: will be interesting to see how Nvidia is positioned in
[6:34] Speaker 7: all of this with you know, CPUs also having their
[6:37] Speaker 7: moment when it comes to the chip DEMND.
[6:39] Speaker 3: That's looking towards Wednesday, after the bell, all after that
[6:42] Speaker 3: key conversation that Ed has Later today, blumbg Intelligence is
[6:45] Speaker 3: man deep saying thank you very much, Elo Musk. He
[6:49] Speaker 3: says he's back in Texas working on plans for that
[6:52] Speaker 3: SpaceX ipo as the rocket and AI company prepares what
[6:55] Speaker 3: could be the largest listing ever. Now Mask made the
[6:58] Speaker 3: comments virtually the sum International Smart Mobility Summer in Tel Aviv.
[7:02] Speaker 9: Take listen IPO. You know we've going to get the
[7:05] Speaker 9: ipo SpaceX ipo going pretty.
[7:07] Speaker 10: Zoon, I think.
[7:08] Speaker 3: Meanwhile, SpaceX has notified investors it's executing a five for
[7:12] Speaker 3: one stock split. It's lowering the per share price for investors.
[7:16] Speaker 3: In an email, shareholders were told the fair market value
[7:19] Speaker 3: are of each share was adjusted to about one hundred
[7:21] Speaker 3: and five dollars roughly five dred and twenty seven dollars
[7:24] Speaker 3: following the split.
[7:25] Speaker 2: According to sources.
[7:26] Speaker 3: Now coming up CME and Silicon Data, they are teaming
[7:30] Speaker 3: up to create a futures market for computing power, a
[7:32] Speaker 3: key resource driving the AI boom. Silicon Data CEO Carmen
[7:37] Speaker 3: Lee is going to be joining us. This is Rumberg Tech.
[7:52] Speaker 3: Let's take a look at today's big number, sixty seven
[7:55] Speaker 3: million dollars. This is the largest power deal in history
[7:58] Speaker 3: triggered by guess what.
[8:00] Speaker 2: AI data center?
[8:00] Speaker 3: Boom next Era is acquiring Dominion Energy for sixty seven
[8:04] Speaker 3: billion dollars in stock forging a giant utility.
[8:07] Speaker 2: It's going to span from Florida to Virginia's.
[8:09] Speaker 3: Tech hubs at Bluemmegg's Leanna Baker, who leads the M
[8:12] Speaker 3: and A team is here with us. Leanna, what is
[8:15] Speaker 3: it that the driving force here? Is it really all
[8:17] Speaker 3: about the AI compute?
[8:18] Speaker 4: Need?
[8:19] Speaker 11: You said it, and I'm surprised it took this long
[8:21] Speaker 11: to see a major deal in utility since demand for
[8:25] Speaker 11: power is at levels not seen since, you know, after
[8:28] Speaker 11: World War Two. So it was really a matter of
[8:30] Speaker 11: time for some of these giants to come together in
[8:33] Speaker 11: utilities because they need to team up to be able
[8:36] Speaker 11: to pay for the build out that's necessary for AI.
[8:39] Speaker 3: They say it's about scale, but with scale comes ayes
[8:43] Speaker 3: of regulators.
[8:44] Speaker 2: How easy is this going to be to get through?
[8:46] Speaker 11: From what we're hearing, It's not impossible. Now's the time
[8:49] Speaker 11: under you know, the Trump administration to go bold and
[8:52] Speaker 11: go big on deal making. But also where Dominion the
[8:56] Speaker 11: target is based. It's only in three states apparently. You know,
[9:00] Speaker 11: Virginia's data center ally is a really key part of this.
[9:04] Speaker 11: But because only three states might be needed for regulatory approval,
[9:08] Speaker 11: you know, this could could really happen within a year
[9:11] Speaker 11: to eighteen months. Of course, there's other regulatory bodies that
[9:14] Speaker 11: also regulate power, nuclear, et cetera. But it's not an
[9:19] Speaker 11: uphill battle. Analysts are saying, you know, it's not a
[9:21] Speaker 11: trivial regulatory review, but looks like it's going to be okay.
[9:25] Speaker 3: The timing is going to be key though, in terms
[9:28] Speaker 3: of there's a little of political pressure, a little bit
[9:31] Speaker 3: political focus on the cost of energy right now in
[9:34] Speaker 3: large pot because of the AI build out, So how
[9:36] Speaker 3: is that going to serve a customer base? Do we
[9:38] Speaker 3: have any idea what impact that happen pricing?
[9:40] Speaker 11: In the press release, they hit on this idea that
[9:43] Speaker 11: there's going to be these credits going to Dominion customers.
[9:46] Speaker 11: So they're very aware that they have to focus on
[9:49] Speaker 11: this idea of cost and price to make sure that
[9:52] Speaker 11: consumers don't think their utility bills are going to go
[9:55] Speaker 11: up because of this deal, so that will be key.
[9:58] Speaker 11: You're right, and you know time will tell if prices
[10:01] Speaker 11: will go up.
[10:02] Speaker 3: Leanna Baker an extraordinary m and a Monday. We thank
[10:05] Speaker 3: you for talking us through it. Look, we talk about
[10:07] Speaker 3: the need for energy within AI, but compute, of course
[10:10] Speaker 3: is officially becoming a financial asset class that we all
[10:13] Speaker 3: care about too. CME Group, Silicon Data then launching the
[10:16] Speaker 3: world's first futures market for AI computing power later this
[10:20] Speaker 3: year Penning regulatory Review now joining us, and please to
[10:22] Speaker 3: say in the studio is Silicon Data CEO. Commonly you're
[10:25] Speaker 3: also the CEO of Compute Exchange, So talk to us
[10:29] Speaker 3: first and foremost what the relationship with CM is going
[10:31] Speaker 3: to be bringing us. How important is it to have
[10:34] Speaker 3: a futures market and compute.
[10:36] Speaker 12: It's incredible important. I'm glad that like came in. This
[10:39] Speaker 12: week is exciting week for the semi industry as a whole.
[10:43] Speaker 12: As you mentioned energy spend, we're pubably going to spend
[10:46] Speaker 12: over ten trillion dollars this year. Think about the SPA market,
[10:49] Speaker 12: fund energy to call, location, to the server themselves. Everything
[10:53] Speaker 12: is at a commodity level. You need finance your infrastructure
[10:57] Speaker 12: to head to manage our risk exposures them While the
[10:59] Speaker 12: bank right they're underwriting trillion's dollars loans, how they manage
[11:04] Speaker 12: their exposure? How do they hatch your future volatility as online?
[11:09] Speaker 12: So from a commodity point of view, think about how
[11:12] Speaker 12: oil futures the wonderful job for people hatching oil fluctuation,
[11:17] Speaker 12: especially during the round wartime. So it's incredibly important to
[11:20] Speaker 12: have futures.
[11:21] Speaker 3: You mentioned the banks, but who else would be the
[11:23] Speaker 3: players involved in such a market that need to be
[11:26] Speaker 3: hatching right now?
[11:27] Speaker 12: So think about people who naturally long GPUs, so they
[11:30] Speaker 12: listen to themselves, new clouds, design houses, fabs, Everyone long
[11:34] Speaker 12: gp exposure. They want a way to hatch it through
[11:37] Speaker 12: short futures or produptions whoever needs compute just like you
[11:41] Speaker 12: and me, Like astartups, they spend billions of dollars on GPUs.
[11:45] Speaker 12: They need to hatch their costs as a complete component
[11:48] Speaker 12: to a line item. Traditionally, sales companies they do not
[11:52] Speaker 12: really have a cost line item, as I found people
[11:54] Speaker 12: now they do either found GPU level or token level
[11:57] Speaker 12: and it's a big line item coming up.
[12:00] Speaker 3: So the whole push is about transparency and an ability
[12:03] Speaker 3: to be able to put these sorts of assets on your.
[12:07] Speaker 2: Overall balance sheet.
[12:09] Speaker 3: But then when it comes transparency just about the cost
[12:11] Speaker 3: of compute, you've already been doing that with the computer exchange.
[12:14] Speaker 2: What is it thing you built there and who's using that?
[12:16] Speaker 12: It's great question. So compute changes farm market for GPO resources.
[12:20] Speaker 12: So we do four contracts as well as reserve contracts
[12:22] Speaker 12: for GPUs resources. So a lot of participants. I would
[12:26] Speaker 12: say all participants so far are the underlying consumption layer,
[12:29] Speaker 12: so AI startups all the way to the neon clouds,
[12:31] Speaker 12: so two really big sides of companies.
[12:34] Speaker 2: We work with European Neo.
[12:36] Speaker 12: Clouds, American New Clouds Sales Eastern Union Clouds. On flip side,
[12:40] Speaker 12: we have startups and enterprises globally. It's really fascinating you
[12:43] Speaker 12: see how people try to do longer term contracts because
[12:46] Speaker 12: they want to lock here in the rate. They don't
[12:48] Speaker 12: want the forty percent daily volatility comes on demand GPO pricing.
[12:51] Speaker 12: There's other way to hadg that fuel futureon options. So
[12:54] Speaker 12: we want to make sure we have the right venue
[12:56] Speaker 12: for the right use space. If you want to reserve GPUs,
[12:59] Speaker 12: you want to face called deliveries go to comput Change.
[13:01] Speaker 12: You want to hedge financial exposures, go to CME in
[13:04] Speaker 12: the future.
[13:06] Speaker 3: Tell us about what sparked this whole idea, I mean
[13:08] Speaker 3: for full transparency for audience. You used to be part
[13:10] Speaker 3: of the Bloomberg team, you used to work in this
[13:13] Speaker 3: very building. What was it that you felt was missing
[13:15] Speaker 3: in the market, Whether it was there literally no transparency
[13:17] Speaker 3: around around access to compute from one AWS versus a
[13:21] Speaker 3: Microsoft offering or a neocloud offering from Corway for example.
[13:25] Speaker 12: I won't say it's no transparency, but very little. So
[13:28] Speaker 12: unless you're in the ecosystem, you're actually in the neocloud transactions.
[13:32] Speaker 12: You prop do not know what's a resource price of
[13:34] Speaker 12: paying people paying for longer term contracts. The about term
[13:37] Speaker 12: structure risk today and I thinking about this lack of
[13:41] Speaker 12: visibility on demand pricing all the way to reserve which
[13:44] Speaker 12: is three years, five yearson deal. And also GPU is
[13:47] Speaker 12: not homogenius products. You need to normalize that. You can
[13:50] Speaker 12: just say, hey, I'm paying the HIT one hundred with
[13:52] Speaker 12: certain HVM, You're paying different HVM devents your location.
[13:56] Speaker 2: We can just do a simple average.
[13:57] Speaker 12: You need indices to normalize the online assets. So producers
[14:01] Speaker 12: as a data that's trackable a representative to the price
[14:04] Speaker 12: we are paying today.
[14:05] Speaker 3: You know, we're just talking to man thing of BlueBag inteligence,
[14:07] Speaker 3: who are saying that we need to get to the
[14:09] Speaker 3: bottom of invidious numbers that come after the bell on
[14:12] Speaker 3: Wednesday about how much HBM or memory costs really starting
[14:16] Speaker 3: to impact their margin. What is it you're saying in
[14:18] Speaker 3: your data about just the cost of computer at the.
[14:20] Speaker 12: Moment with the price that going up, the price going
[14:23] Speaker 12: up since December last year, it was pretty interesting last
[14:26] Speaker 12: year everyone's talking about, Hey, shouldn't the computer price always
[14:29] Speaker 12: come down and come down to zero by a year
[14:31] Speaker 12: A It's not really happening, supplied. The man shift every
[14:35] Speaker 12: single day, the term structure to forward pattern shift every
[14:38] Speaker 12: single day because people have these expectations, they do different
[14:41] Speaker 12: bournets come out all the time. You stood at last week.
[14:44] Speaker 12: Last year is a fab and this year is memory
[14:46] Speaker 12: and going forward maybe you call location space. So every
[14:50] Speaker 12: time it's different.
[14:52] Speaker 2: So pricing pressure is there for you.
[14:54] Speaker 3: The pressure is on to get this deal solidified with
[14:56] Speaker 3: the CME.
[14:56] Speaker 2: You've announced it, but we didn't say there's regulatory oversight.
[14:59] Speaker 2: What sort of he do you need to go through
[15:01] Speaker 2: what sort of questions we asked.
[15:03] Speaker 12: So we are actively working with CME and then review
[15:05] Speaker 12: with CFTC and other regulatory bodies.
[15:08] Speaker 2: It is a very.
[15:09] Speaker 12: Traditional data future products. We're not talking anything that's exotic, right.
[15:14] Speaker 12: We have done that for oil, for gas, for any
[15:16] Speaker 12: other as a classes, so I really don't see a
[15:18] Speaker 12: major concern farm for anyone.
[15:21] Speaker 3: And going global, how global is the offering that you
[15:24] Speaker 3: currently have with CME or indeed the compute exchange. Where
[15:27] Speaker 3: else do you likely to push in two more broad
[15:29] Speaker 3: offerings worldwide?
[15:31] Speaker 12: So we launched GPU Industries at Bloomberg Terminal last year,
[15:34] Speaker 12: the world's first GPU indices. We cover all prices x China,
[15:39] Speaker 12: so think about the European from the US. We normalize
[15:42] Speaker 12: it your location to win data points, so anything x
[15:45] Speaker 12: China is capture our indices. So ideally around the world
[15:48] Speaker 12: you can hatch that you're exposure with our indices.
[15:51] Speaker 3: Well, have to see how this progresses come back on.
[15:54] Speaker 3: I hope Silicon Data CEO commonly there on all things
[15:58] Speaker 3: futures market for compute becoming up. We're talking Apple's revamped Siri,
[16:02] Speaker 3: but it was expected to introduce a new privacy focused
[16:04] Speaker 3: features aimed at selling it apart from rival chatbelts or
[16:08] Speaker 3: on that. Next, this is Blueberg Tech time now for
[16:23] Speaker 3: Talking tech and first up by Jus. One percent drop
[16:25] Speaker 3: in sales was less than analysts had expected, void by
[16:28] Speaker 3: its new AI businesses, but the tech giant in China
[16:31] Speaker 3: is making a strategic pivot, shifting towards automated agents and
[16:35] Speaker 3: aipower cloud.
[16:36] Speaker 2: Units and away from search plus.
[16:39] Speaker 3: Publicis is acquiring online data broker live Ramp for about
[16:43] Speaker 3: two point five billion.
[16:44] Speaker 2: Dollars in cash.
[16:45] Speaker 3: Says it's a paying a thirty percent premium per share
[16:47] Speaker 3: to supercharge its data assets and accelerate its AI driven
[16:51] Speaker 3: marketing capabilities and.
[16:53] Speaker 2: Shares of Japanese chip maker Kyosha.
[16:56] Speaker 3: They were flooded with by orders after forecasting a massive
[16:58] Speaker 3: eight point two billion dollar quarterly operating report profit. It
[17:02] Speaker 3: is latest sign of meteoric demand hardware for AI services
[17:06] Speaker 3: and the resulting shortage and conventional memory chips. Let's talk
[17:10] Speaker 3: elsewhere in the world of AI, because Apple is set
[17:13] Speaker 3: to debut is iOS twenty seven and this year is WWDC,
[17:16] Speaker 3: featuring a revamped Siri experience.
[17:18] Speaker 7: Now.
[17:18] Speaker 3: The update is expected to introduce new privacy focus features.
[17:22] Speaker 2: That could help.
[17:22] Speaker 3: Differentiate SyRI in the crowded chatbots space. For more, Bloomberg
[17:26] Speaker 3: Apple and Consumer Tech Managing editor Mark gum And brings
[17:28] Speaker 3: us what is always the bread and butter you bring
[17:30] Speaker 3: us to scoops and push us towards For many reasons,
[17:34] Speaker 3: Apple wants to lean in on its privacy advantages. That
[17:36] Speaker 3: what is going to make it distinct versus the other
[17:39] Speaker 3: offerings already out there.
[17:41] Speaker 13: Yeah, for Apple, privacy has been a differentiator now for
[17:44] Speaker 13: going over a decade. They talk about privacy when they
[17:48] Speaker 13: release any new product, or any new operating system or
[17:51] Speaker 13: any new software service feature. For them, it's, you know,
[17:55] Speaker 13: their north star that they like to say, everything is
[17:58] Speaker 13: built in from that perspective, and they release see it
[18:00] Speaker 13: is a selling point that people want devices that have
[18:02] Speaker 13: better encryption, better security, better privacy than the competition. And
[18:06] Speaker 13: obviously this is really important for everything you're doing on
[18:09] Speaker 13: Apple devices. You have your financial data on there, your communications,
[18:14] Speaker 13: you have your smart home information, all sorts of different features.
[18:17] Speaker 13: You want the key private and that's going to extend
[18:20] Speaker 13: to the new SERI. We're in an age right now
[18:22] Speaker 13: where everyone's got a chatbot, everyone's got AI features, and
[18:26] Speaker 13: just like you've seen Apple do with features and comparing
[18:29] Speaker 13: themselves to Meta and Google and whatnot. You'll see them
[18:33] Speaker 13: do that with this chatbot that they're going to announce
[18:35] Speaker 13: on June eighth and developer conference, talking about THEIRS retaining
[18:39] Speaker 13: less information and less so called memory than other chatbots.
[18:43] Speaker 13: And one feature that a report over the weekend that
[18:45] Speaker 13: they're planning for this is the auto deletion of chatbot conversation.
[18:49] Speaker 13: So you can set your chats just like you can
[18:52] Speaker 13: with the Messages app, to save in your phone and
[18:54] Speaker 13: your iCloud forever, or you can have them auto delete
[18:57] Speaker 13: after a year, or you can have them auto delete
[19:00] Speaker 13: after thirty days.
[19:02] Speaker 9: So that's going to be a bit of a unique feature.
[19:04] Speaker 13: Other chatbots have what they call on incognito mode, which
[19:07] Speaker 13: is basically, you know, private browsing mode and a web browser,
[19:10] Speaker 13: but private chatting in your chatbot.
[19:13] Speaker 9: But this is going to be a little different.
[19:16] Speaker 3: Talk us about just how technologically that makes it stand out.
[19:19] Speaker 3: If I have an auto delete for thirty days, what
[19:21] Speaker 3: does that actually mean to the memory that than me
[19:24] Speaker 3: having had that discussion with the chatbot and it being
[19:27] Speaker 3: useful going forward.
[19:29] Speaker 9: Yeah, Well, it's two separate things.
[19:30] Speaker 8: Right.
[19:30] Speaker 13: The auto delete of conversations means that that conversation, that
[19:34] Speaker 13: chat that's just gone forever. In terms of memory, right,
[19:38] Speaker 13: you have chatbots that keep your memory forever, keep your
[19:41] Speaker 13: memory for a finite amount of time. Apple is going
[19:44] Speaker 13: to retain memory a shorter amount of time than a
[19:48] Speaker 13: lot of the competition, which may make it useful in
[19:51] Speaker 13: the short term less useful in the long term. They're
[19:54] Speaker 13: using techniques. There's one technique they have called differential privacy
[19:58] Speaker 13: to be able to retain some memory but have it
[20:00] Speaker 13: glued exactly to the user, right, And when it's stored,
[20:04] Speaker 13: not glued towards the user, but you're able to stort
[20:06] Speaker 13: on device, so it's able to come back and tell
[20:09] Speaker 13: you what you've talked about or give you that context.
[20:12] Speaker 9: So I think it is going to work fairly well.
[20:15] Speaker 13: It is going to be fairly competitive, but Apple really
[20:17] Speaker 13: going to market it around privacy and security.
[20:20] Speaker 3: As always A muster read with power on Bloombo's Mark German,
[20:22] Speaker 3: Thank you very much. Indeed, let's talk about Miami International
[20:25] Speaker 3: Holdings or MIAX. It has seen tremendous growth since it
[20:29] Speaker 3: was founded over a decade ago, operating eight exchanges, cross options,
[20:33] Speaker 3: Secretaries and futures for disclosure purposes as part of a
[20:37] Speaker 3: collaboration with Bloomberg and MIAX have launch pad Futures and Options,
[20:41] Speaker 3: and Bloomberg has a dedicated studio on the exchange floor
[20:44] Speaker 3: where Bloomberg's Norah Melinda is joining us. Now, heart off
[20:47] Speaker 3: your interview with MIAX CEO Tom Nus Gallaghan. Talk to
[20:50] Speaker 3: us about what you've been discussing over there.
[20:54] Speaker 14: Well, Carolina, have the opportunity to chat with Tom Gallagher
[20:57] Speaker 14: earlier this morning. Of course, we're here celebrating the launch
[21:00] Speaker 14: the Bloomberg one hundred equity Index Futures. Now, this is
[21:03] Speaker 14: a product that the company is launching, hopefully geared at
[21:06] Speaker 14: retail traders. We've been seeing an emphasis on a lot
[21:08] Speaker 14: of companies, a lot of brokerages really being interested in
[21:11] Speaker 14: bringing in the retail participation. This company offering the opportunity
[21:16] Speaker 14: of commission free for retail traders as they are launching
[21:19] Speaker 14: this product, the first of a series of products to
[21:22] Speaker 14: also come. But what was really interesting from my conversation
[21:25] Speaker 14: earlier with Tom was the idea of Miami as the
[21:28] Speaker 14: Wall Street South. We've been seeing so many big business leaders,
[21:33] Speaker 14: Ken Griffin, a number of other names talking about the
[21:35] Speaker 14: tax environment being more favorable in states like Florida, southern states,
[21:40] Speaker 14: and so you're certainly seeing my EX really leveraging that
[21:44] Speaker 14: coming to Miami. They're headquarters, funnily enough, were actually in Princeton,
[21:48] Speaker 14: New Jersey, but the trading floor just behind me. So
[21:51] Speaker 14: a lot of activity here in some of these southern states,
[21:54] Speaker 14: and certainly my AX looking toward further growth and development here.
[21:58] Speaker 2: Fantastic.
[22:00] Speaker 3: Thanks for bringing us and insights from your conversation. We
[22:02] Speaker 3: appreciate it. Coming up, well, we've got plenty to be
[22:05] Speaker 3: discussing with a recursive CEO, Richard Soshia. It's an AI
[22:09] Speaker 3: company coming out of Stealth and already he's been running
[22:13] Speaker 3: you dot com as the CEO. He's already also got
[22:16] Speaker 3: AIX which is a venture fund, but now he's starting
[22:19] Speaker 3: yet a new company all around the issues of AI
[22:23] Speaker 3: and where we take large language model creation in the
[22:26] Speaker 3: lab and the frontier lab lab area. We're looking though
[22:29] Speaker 3: right now as AI not being the savior to these markets.
[22:32] Speaker 3: Now's that one hundred off by seven ten percent. Look,
[22:34] Speaker 3: it's a second straight day of losses. We've seen the
[22:36] Speaker 3: semiconductor index as well, off by two point four percent.
[22:38] Speaker 3: That brings its losses in the last couple of days
[22:41] Speaker 3: to more than six percent. Of course, got a lot
[22:44] Speaker 3: of the eyes on a video for later this week Wednesday,
[22:46] Speaker 3: after the bell we get their numbers. We're off five
[22:48] Speaker 3: percentage point for the world's most valuable company. This is
[22:50] Speaker 3: Bloomberg Tech. Welcome back to Bloomberg Tech. We got a
[23:00] Speaker 3: little bit of risk aversion creeping into the market, and
[23:02] Speaker 3: it's largely geopolitically focused. We're worried once again about the
[23:06] Speaker 3: conflict to US Iran, and we're seeing all prices on
[23:09] Speaker 3: the rise. That is seeing yet a second day of pullback,
[23:11] Speaker 3: particularly on an ADSDAK one hundred or off by eight
[23:13] Speaker 3: nine ten percent, let's call it. We're seeing the chip
[23:16] Speaker 3: stocks under pressure and a particular Micron and the like
[23:18] Speaker 3: really dragging in points perspective. But I'm also looking what's
[23:20] Speaker 3: happening in Crippo. Look risk assets in general taking a
[23:24] Speaker 3: bit of a hammering. We're seeing down two point eight percent,
[23:26] Speaker 3: but we're sub seventy seven thousand now it's seventy six
[23:29] Speaker 3: thousand for bitcoin, and there's been about seventy million in
[23:32] Speaker 3: terms of the actual bearish bets coming back into the market,
[23:35] Speaker 3: bullish bets are being taken off. We've seen seven hundred
[23:38] Speaker 3: million dollars worth of those bullish bets being taken off
[23:40] Speaker 3: in the last twenty four hours alone, according to coin Glass.
[23:43] Speaker 3: So just keep an eye on risk sentiment. But look,
[23:45] Speaker 3: if we look across the Asia, not always bearish. I'm
[23:48] Speaker 3: looking at Samsung shares in South Korea up nearly we
[23:51] Speaker 3: call it three percent. Investors were actually encouraged by signs
[23:54] Speaker 3: at the company's union it's open to some negotiations, while
[23:57] Speaker 3: at local court moved to limit the threat of strike remotes.
[24:01] Speaker 3: Global Tech editor Peter Elstrom joins and just remind us
[24:04] Speaker 3: of the context here. This is Samsung that has done
[24:07] Speaker 3: phenomenally well in terms of share price, in terms of profitability,
[24:10] Speaker 3: and now labor wants to see some upside from it.
[24:15] Speaker 15: Yeah, these are very tense times in South Korea, make
[24:18] Speaker 15: no mistake about it. Samsung is squaring off against its union.
[24:22] Speaker 15: The union has said it may go on strike as
[24:24] Speaker 15: soon as Thursday for eighteen days. That's a big deal
[24:27] Speaker 15: for the company because, as you say, they're in the
[24:29] Speaker 15: middle of this boom and memory chip prices, which has
[24:31] Speaker 15: been great for their profits. They made a ton of
[24:33] Speaker 15: money because of that. The whole AI industry needs more
[24:36] Speaker 15: of the memory, both from Samsung and from s K
[24:38] Speaker 15: Heinez at this point, but this labor disruption threatens to
[24:42] Speaker 15: throw a wrench into that whole dynamic. And just to
[24:44] Speaker 15: set the scene, it is quite unusual for Samsung to
[24:47] Speaker 15: have this kind of conflict with their union. For decades,
[24:50] Speaker 15: they never had any conflict at all, and then just
[24:52] Speaker 15: a couple of years ago they had a small strike,
[24:54] Speaker 15: but it wasn't that serious. So this is unusual for
[24:57] Speaker 15: them to come this far. As you say, they've agreed
[24:59] Speaker 15: to go back to the table after setting this possible
[25:01] Speaker 15: strike date. They are talking again, but they haven't reached
[25:04] Speaker 15: a resolution yet. The court took a little bit of
[25:06] Speaker 15: the damage, the potential damage off the table by saying
[25:09] Speaker 15: there are some limitations about what the union can do
[25:12] Speaker 15: at this point, but there's still this possibility of a
[25:14] Speaker 15: strike coming up later on this week, which would affect
[25:17] Speaker 15: not just Samsung and the South Korean companies, but the
[25:20] Speaker 15: AI industry more broadly.
[25:21] Speaker 3: Yeah, everyone waiting with bated breath as we worry about
[25:24] Speaker 3: memory prices yet further if there's less of supply. But
[25:27] Speaker 3: Peter I mean, it was just last week that we
[25:30] Speaker 3: saw the South Korean market lose billions, hundreds of billions
[25:34] Speaker 3: of dollars in just ninety minutes, as people wondered how
[25:37] Speaker 3: politically speaking, we might see AI profits be shared with
[25:41] Speaker 3: a nation writ large. We've always seen s k Heinez
[25:44] Speaker 3: have to offer more of its profits to its workers.
[25:46] Speaker 3: Just remind us of the tensions there and if they'll
[25:49] Speaker 3: also spill out globally.
[25:51] Speaker 15: Yeah, the Korean society right now is having this debate
[25:54] Speaker 15: about what are you going to do with these enormous
[25:56] Speaker 15: profits coming out of AI? And also the government is
[25:59] Speaker 15: beginning to ask this very important question, how should governments
[26:02] Speaker 15: prepare for this world that's coming at us with AI,
[26:05] Speaker 15: with all these AI improvements, where perhaps you're not going
[26:08] Speaker 15: to have the same kind of jobs that you had
[26:09] Speaker 15: in the past. So, as you're referring to, one of
[26:12] Speaker 15: the policy advisors to the president posted something on his
[26:16] Speaker 15: Facebook page where he talked about the possibility of a
[26:18] Speaker 15: citizen dividend where they would take some of the profits
[26:21] Speaker 15: from these big companies and use those to establish a
[26:24] Speaker 15: more stable foundation for the society that could involve retraining.
[26:28] Speaker 15: He said it could involve support for people who want
[26:31] Speaker 15: to startup companies. It could involve a number of different things,
[26:34] Speaker 15: but his goal was really to kick off this conversation
[26:36] Speaker 15: and try to ask the question, how do we as
[26:38] Speaker 15: a government proactively start to think about these big issues
[26:41] Speaker 15: and if we're going to get this windfall of profits
[26:44] Speaker 15: because these companies are making so much money, how do
[26:46] Speaker 15: we use that cash to the best advantage for what's
[26:48] Speaker 15: ahead of us?
[26:49] Speaker 3: Thinks Peter Elstrom. Fascinating what's happening in Asia. We appreciate it.
[26:54] Speaker 3: Let's look at what's happening with Google's leadership in AI
[26:56] Speaker 3: right now has made computing power a really prized resources
[27:00] Speaker 3: inside the company, with priority increasingly going two major projects
[27:04] Speaker 3: like Gemini and paying customers.
[27:06] Speaker 2: The bottleneck. It's frustrated to some employees, prompting a number of.
[27:09] Speaker 3: AI researchers to leave for startups for Morell bloombergs Julia Love,
[27:13] Speaker 3: who covers Alphabet joins us now and Julia they're going
[27:16] Speaker 3: and founding their own companies because they can't get what
[27:19] Speaker 3: they want done internally at Google. How how much of
[27:23] Speaker 3: a change is that to be struggling to get access
[27:26] Speaker 3: to computing?
[27:28] Speaker 6: Yes, so I think it's a you know, just a
[27:30] Speaker 6: fact of the AI industry that researchers would always like
[27:34] Speaker 6: more compute, but the shortage has just become more acute
[27:37] Speaker 6: in this day and age because large language models require
[27:40] Speaker 6: such vast amounts of computing power. And so traditionally AI
[27:45] Speaker 6: research at Goodle, especially in Goodle Brain and one of
[27:49] Speaker 6: the labs, was very idalitarian and researchers not a certain
[27:52] Speaker 6: amount of computing power and they could team up to
[27:55] Speaker 6: try to do bigger projects. But now that fool still exists,
[28:00] Speaker 6: but it contracts in times of high demand, and researchers
[28:06] Speaker 6: just often can't get meaningful research done with their allotations.
[28:11] Speaker 6: And so if their work isn't a priority at Google,
[28:14] Speaker 6: that is the only you know, avenue they can pursue,
[28:17] Speaker 6: whereas in a startup they have more doors to knock on.
[28:21] Speaker 3: I mean, let's just talk about Andrew Dye, who've focused
[28:24] Speaker 3: on in this particular story. You've also focused on a
[28:27] Speaker 3: female founded company that's gone on to get VC backing
[28:30] Speaker 3: when they articulated.
[28:32] Speaker 2: To you what it used to be like what it
[28:33] Speaker 2: is now in terms of building businesses.
[28:35] Speaker 3: But is it because Alphabet is too busy selling out
[28:39] Speaker 3: its compute to its cloud customers or what is it
[28:42] Speaker 3: that's making it so much more of a tussle.
[28:45] Speaker 6: Yes, I think there's a few things happening at once.
[28:48] Speaker 6: One is that generative AI has made you know, Goodle's
[28:52] Speaker 6: consumer products so popular, and Goodle needs a lot of
[28:55] Speaker 6: computing power to serve those. And then the cloud business
[28:59] Speaker 6: is to peking off, so Brutal wants to meet that demand.
[29:03] Speaker 2: And then at the.
[29:03] Speaker 6: Same time, AI research is more computing and intensive than
[29:07] Speaker 6: it once was, and so researchers need more compute to
[29:11] Speaker 6: do a cutting edge experiment. And so there's a lot
[29:14] Speaker 6: of mouths to feed. Everyone wants more. And if people
[29:18] Speaker 6: run up against roadblocks at Google, that, oh.
[29:27] Speaker 2: What a tech issue.
[29:29] Speaker 3: On the Tech Show, Dan, we'll get back to Julia
[29:32] Speaker 3: Love when we can, but we appreciate her insights there. Meanwhile,
[29:34] Speaker 3: let's talk about where else computer is being used and
[29:37] Speaker 3: what we're seeing in the Frontier labs. Because Recursive is
[29:40] Speaker 3: a startup creating AI that conducts experiments on how to
[29:44] Speaker 3: safely improve.
[29:45] Speaker 2: Itself and it just came out of stealth. The company is.
[29:47] Speaker 3: Valued at four point six five billion dollars. Back is
[29:50] Speaker 3: including Greycroft Video, AMD Ventures and more. CEO Richard Susha
[29:55] Speaker 3: joins us now for more. Many will recognize you because Richard,
[29:58] Speaker 3: you've joined on another company that you're also leading, you
[30:01] Speaker 3: dot Com. You also happen to be busy in venture capital.
[30:03] Speaker 3: So talk to us about Recursive. What is it that
[30:06] Speaker 3: you felt needed to be fixed?
[30:09] Speaker 10: So we want to build recursive self improving superintelligence that
[30:14] Speaker 10: automates knowledge discovery. You know, AI now is code and
[30:18] Speaker 10: can code, and so you can have a closed loop
[30:22] Speaker 10: where you replace human ingenuity and trying to come up
[30:27] Speaker 10: with ideas, implement those ideas, and validate those ideas, and
[30:32] Speaker 10: you can actually allow AI to do that on itself.
[30:35] Speaker 3: The reason you're able to enter out of stealth with
[30:37] Speaker 3: a four billion dollar market capitalization is because you're a
[30:40] Speaker 3: pretty well known famed AI researcher, Richard. But there are
[30:44] Speaker 3: many well, shall we say, there are many very ambitious
[30:47] Speaker 3: labs out there who also want to reach superintelligence.
[30:49] Speaker 2: Why aren't they doing it right?
[30:50] Speaker 3: What is it that you feel Recursive is going to
[30:53] Speaker 3: gain the edge or rather than helping another lab already
[30:56] Speaker 3: do what they're doing.
[30:58] Speaker 10: Yeah, we're we collected a combined an incredible team of
[31:03] Speaker 10: co founders, folks from Google DeepMind, from open AIE from
[31:09] Speaker 10: Salesforce Research and many other incredible labs Meta and so on.
[31:14] Speaker 10: We are focused on open endedness and really have the
[31:18] Speaker 10: entire company try to allow AI to build the next
[31:22] Speaker 10: better version of AI. And whenever in AI research we
[31:25] Speaker 10: replace manual systems with learned systems, we see a massive improvement.
[31:31] Speaker 10: And then on top of that, we can sit now
[31:33] Speaker 10: on top of this whole new infrastructure where we built
[31:37] Speaker 10: a company from the ground up by empowering AI to
[31:40] Speaker 10: do its own experimentation, whereas most of the other labs
[31:43] Speaker 10: in the past basically started before that was possible.
[31:47] Speaker 3: Okay, so you've got this focus from the very birth
[31:50] Speaker 3: of the business, but you also need a lawful lot
[31:53] Speaker 3: of money, as we can see there raising for Compute to.
[31:55] Speaker 2: Take on some of the juggernauts. How much is it
[31:58] Speaker 2: a costly business?
[31:59] Speaker 3: How are you thinking of reframing the demand for GPUs
[32:02] Speaker 3: and the like.
[32:04] Speaker 4: You're absolutely right.
[32:05] Speaker 10: A big new scaling law we're trying to prove is
[32:09] Speaker 10: that the more compute results in more inventions, more improvements,
[32:13] Speaker 10: and so compute is one of the largest costs for us,
[32:17] Speaker 10: and that's where we're really excited to have the folks
[32:20] Speaker 10: like Nvidia and AMD be part of this round.
[32:23] Speaker 3: Just the open ended nature of driving in endless innovation.
[32:26] Speaker 3: You say you want to do that safely. What are
[32:28] Speaker 3: the worries and the limits of it not being done safely?
[32:31] Speaker 4: You're absolutely right. Safety is a huge concern.
[32:34] Speaker 10: That's why we have some of our co funders who've
[32:36] Speaker 10: done research in rainbow teaming, for instance, where I improve
[32:40] Speaker 10: the safety of large language models like Tim rock Teeschel
[32:44] Speaker 10: and his collaborators worked on. We think we take safety
[32:49] Speaker 10: very seriously and we believe though that this is the
[32:53] Speaker 10: best and fastest way to superintelligence, which will then allow
[32:57] Speaker 10: humanity to flourish much much more by inventing building ultimately
[33:01] Speaker 10: a Eureka machine, one that allows us to invent everything
[33:05] Speaker 10: else thereafter.
[33:07] Speaker 2: How are you going to divide your time? Richard?
[33:10] Speaker 3: Your CEO of you dot com, and you're also at AIX,
[33:14] Speaker 3: the venture firm that's just lost its co founder.
[33:16] Speaker 2: So how do you divv this all up?
[33:20] Speaker 10: I work all the time and have incredible teams around
[33:24] Speaker 10: me that enable this of AI agents, also AI agents, yeah,
[33:29] Speaker 10: but also some really incredible humans.
[33:31] Speaker 3: Okay, incredible humans cost a lot of money, so were
[33:33] Speaker 3: they incentivized by equity, by the sheer ambition of the task,
[33:40] Speaker 3: or how are you going to be hiring yet further
[33:41] Speaker 3: talent because they're also getting to a significant expense line
[33:45] Speaker 3: compared to GPUs.
[33:48] Speaker 10: That's absolutely right. Yeah, I think as a founder these days,
[33:51] Speaker 10: you want to think more about the overall potential of
[33:55] Speaker 10: the pie rather than your slice an amount of pie
[33:58] Speaker 10: that you have as a And we do share equity
[34:01] Speaker 10: with our employees and everyone, and I think that encourages
[34:05] Speaker 10: and excites everyone. I think in general, the more you
[34:08] Speaker 10: have ownership of what you're creating in this AI world,
[34:11] Speaker 10: the more you love AI, because then you just care
[34:13] Speaker 10: about the outputs. And so I think it's important to
[34:16] Speaker 10: share that equity and ultimately grow something amazing together.
[34:20] Speaker 2: Which is so sure.
[34:21] Speaker 3: Great to have some time with you, CEO, co founder
[34:23] Speaker 3: of Recursive, We appreciate your time today. Coming up, we're
[34:26] Speaker 3: going to go back to Las Vegas is Dell World
[34:29] Speaker 3: and we're speaking with Daniel Roberts from the AI cloud
[34:31] Speaker 3: provider Iron.
[34:32] Speaker 2: That's next.
[34:33] Speaker 3: That's a pretty big tech Let's get back over to
[34:46] Speaker 3: Las Vegas. Our own ed Ludlow is standing by with
[34:48] Speaker 3: Dell Technologies World.
[34:49] Speaker 4: Event upon you ed, Yeah, we've got a really great
[34:55] Speaker 4: case study.
[34:55] Speaker 5: Iron is an AI cloud that is building infrastructure all
[34:59] Speaker 5: over the world, not just the United States and all
[35:01] Speaker 5: these geographies.
[35:02] Speaker 4: Dell is its key server partner.
[35:05] Speaker 5: Right it wants the Nvidia GPUs. Couldn't think of a
[35:08] Speaker 5: better conversation to have to kind of kick off the day.
[35:10] Speaker 5: Dan Roberts, co founder co CEO, you've been busy guy
[35:14] Speaker 5: as well on the M and A side. You've acquired
[35:17] Speaker 5: a marketing or creative agency. I mean basically why. But
[35:23] Speaker 5: there's also this conversation Caroline and I have it all
[35:25] Speaker 5: the time about the public's perception of what's really happening
[35:29] Speaker 5: with data center and how it impacts them.
[35:31] Speaker 13: Yeah.
[35:31] Speaker 16: Look, brand awareness, we're going global, We're growing our customer base.
[35:36] Speaker 16: It's really important to us. But that's a small point
[35:38] Speaker 16: of the grand scheme of things. It's all about power
[35:41] Speaker 16: and time to compute, and the issue is you can't
[35:44] Speaker 16: employment a software update to bring on power. The tech
[35:48] Speaker 16: industry is very brilliant at going really fast, but you
[35:51] Speaker 16: can't code your way to an AI gig awad of compute.
[35:54] Speaker 4: And that's what we're focused on.
[35:55] Speaker 5: In the United States, at least there is some some
[35:58] Speaker 5: sort of public backlash in zip codes where data center's
[36:01] Speaker 5: getting built, the public feels the pressure of higher energy process.
[36:05] Speaker 5: What's that like for iron and what are you confronting
[36:08] Speaker 5: in that respect?
[36:09] Speaker 16: Yeah, eight years ago when we started this business, we
[36:11] Speaker 16: had a differentiated strategy. We never went to metropolitan areas.
[36:15] Speaker 16: We went to regional communities where there was an abundance
[36:18] Speaker 16: of renewable energy that creates a cost structural advantage over time.
[36:22] Speaker 16: And we've reinvigorated local communities across British Columbia, Canada, Texas.
[36:27] Speaker 16: And it's an advantage for us because we are located
[36:30] Speaker 16: where people want us.
[36:31] Speaker 4: To do business.
[36:32] Speaker 5: Dan, I think I'm saying, just as with everyone else,
[36:35] Speaker 5: you're in a position where demand for compute, your compute
[36:39] Speaker 5: is vastly outpacing your ability to supply it. What is
[36:43] Speaker 5: the biggest barrier to you getting more capacity online, getting
[36:47] Speaker 5: more stuff built?
[36:49] Speaker 4: Absolutely? I mean we're still in the dial up era
[36:51] Speaker 4: of AI.
[36:52] Speaker 16: When you step back the amount of friction you've got
[36:54] Speaker 16: waded in for prompts, You've got to craft your prompt
[36:56] Speaker 16: really carefully. Complex prompts take fifteen to forty seconds to
[37:00] Speaker 16: actually spit out an adser as. That speeds up demand accelerates,
[37:04] Speaker 16: which drives more compute and more use cases. And for us,
[37:07] Speaker 16: it's been a focus on what is the end choke point,
[37:10] Speaker 16: and that's power, it's feel, it's concrete, it's actually building
[37:14] Speaker 16: the physical infrastructure.
[37:15] Speaker 4: That takes years and years.
[37:16] Speaker 16: If you wanted to start today and build a gigawat
[37:19] Speaker 16: ai factory, you are looking twenty thirties before you get
[37:23] Speaker 16: the first compute online.
[37:24] Speaker 5: Let's get a bit deeper on that. So if you
[37:26] Speaker 5: wanted to say, in other words, you choose a site,
[37:29] Speaker 5: you get permissions, You're not going to get it online
[37:32] Speaker 5: for six years, four years.
[37:34] Speaker 16: It'll take you eighteen to twenty four months to get
[37:36] Speaker 16: the attention of a utility to work out whether your
[37:38] Speaker 16: block of dirt even has power or not.
[37:41] Speaker 4: That's the starting point. How we're getting to it. What
[37:43] Speaker 4: are the utilities not doing that you need them to do? Oh, utilities,
[37:47] Speaker 4: they're doing the best they can.
[37:48] Speaker 16: Like they're running complex networks with a lot of generation
[37:51] Speaker 16: on one side, a lot of power demand on the
[37:53] Speaker 16: other and at the end of the day, when you
[37:54] Speaker 16: connect to them. This is why grid connected capacity is
[37:58] Speaker 16: so valuable. When you connect to them, they're underwriting twenty
[38:01] Speaker 16: four to seven guaranteed power, so they're naturally risk adverse and.
[38:04] Speaker 4: They're naturally dealing with in the inundation of requests.
[38:07] Speaker 5: At the moment, I'd love to go back to basis
[38:09] Speaker 5: for seconds. You build data centers, and those data centers
[38:12] Speaker 5: have servers inside them, and those servers are for AI workloads.
[38:17] Speaker 5: What's it like working with a Dell? What is it
[38:20] Speaker 5: that you rely on Dell four? And how do you
[38:23] Speaker 5: get hold of in video GPUs?
[38:24] Speaker 6: Right?
[38:25] Speaker 5: There is this idea that you can just phone gents
[38:27] Speaker 5: and one and say hey, I need some of those,
[38:29] Speaker 5: but in practice it probably doesn't really work like that.
[38:31] Speaker 14: Yeah.
[38:32] Speaker 16: Look, we've got great relationship with both Dell and in Video.
[38:35] Speaker 16: We announced a partnership with n Video for five gigawats
[38:37] Speaker 16: of compute.
[38:38] Speaker 4: They're also asked Weed in an investment.
[38:40] Speaker 16: Yes, so they're exposed to upside in US as we
[38:44] Speaker 16: grow and deploy their GPUs, and we're working on their
[38:47] Speaker 16: reference architecture for a gigawat factory in Sweetwater. But the
[38:51] Speaker 16: opportunity with those partners is to work closely and build
[38:55] Speaker 16: up the whole ecosystem all the way from land Steel
[38:58] Speaker 16: concrete down to the end us of this compute.
[39:01] Speaker 5: Is there a geography somewhere in the world that is
[39:03] Speaker 5: a best place to build a data center right now?
[39:06] Speaker 16: Look, every geography is slightly different in terms of how
[39:08] Speaker 16: they handle things. But the majority of our capacities in Texas,
[39:12] Speaker 16: and they've been pretty good.
[39:14] Speaker 5: If you could ask mister Wog and mister Dell one
[39:17] Speaker 5: thing right now, or we ask something of them, what
[39:20] Speaker 5: would it be.
[39:22] Speaker 16: Oh, look, I think the whole industry is moving from.
[39:24] Speaker 16: If silicon is no longer the choke point, how do
[39:27] Speaker 16: we solve steel? How do we solve concrete? How do
[39:30] Speaker 16: we solve the killer watts?
[39:32] Speaker 5: And this president and this administration they have an answer
[39:35] Speaker 5: for that.
[39:36] Speaker 4: I think everyone's trying to work it out. It's been an.
[39:39] Speaker 5: Interesting evolution for your company right now. Public in the
[39:43] Speaker 5: United States, Sydney, Australia is your home. What's that like,
[39:47] Speaker 5: trying to work across borders at the moment and trying
[39:50] Speaker 5: to get projects done at the same time in different jurisdictions.
[39:54] Speaker 16: Oh, look, the reality is we've got projects in North America,
[39:56] Speaker 16: We've got projects in Europe, We've got projects in Australia. Yes,
[40:00] Speaker 16: my family is in Sydney and I get back for
[40:02] Speaker 16: the weekends. But you know when people ask where I leave,
[40:05] Speaker 16: plenty bulances on a plane.
[40:07] Speaker 5: Den Roberts, co founder co CEO of Iron really critical
[40:12] Speaker 5: AI cloud and like your previous guests, I think everyone's
[40:15] Speaker 5: just working all the time right now, character.
[40:17] Speaker 2: Largely on planes ahead. Thanks so much.
[40:20] Speaker 3: Look tune in later today as ed sits down with
[40:23] Speaker 3: the CEOs of Dell of Nvidia for an exclusive interview.
[40:26] Speaker 2: Live from der World.
[40:28] Speaker 3: You don't want to miss up to thirty pm Eastern
[40:29] Speaker 3: eleven thirtym Pacific.
[40:31] Speaker 2: Coming up new thirty F filings.
[40:33] Speaker 3: They're revealing how hedge funds are repositioning around big tech
[40:36] Speaker 3: and the AI trade. Will break down the BEGINUS moves next.
[40:39] Speaker 2: This is blue Beg Tech.
[40:52] Speaker 3: Fresh regulatory filings known as thirty F filings, giving investors
[40:56] Speaker 3: a close look at how hedge funds and major money
[40:58] Speaker 3: managers are positioning around tech and the AI trade. Joining
[41:01] Speaker 3: US now to break it all down. As Blomberg's Hemma
[41:03] Speaker 3: Palma is always so much happening each quarter at this time,
[41:07] Speaker 3: Hama passed through, for example, what's happening with Microsoft, because
[41:10] Speaker 3: certainly the bill Atman News quarter attention. Where else is
[41:12] Speaker 3: there been moves in that company?
[41:14] Speaker 17: Absolutely, Bill Ackman taking a pretty sizeable steak in Microsoft,
[41:17] Speaker 17: more than two billion dollars worth, making it his fourth
[41:20] Speaker 17: biggest holding. And this is a stock that fell a
[41:23] Speaker 17: great deal in the first quarter, So he may have
[41:25] Speaker 17: been buying up these shares that attractive prices, but he
[41:29] Speaker 17: believes that Microsoft is stronger and a lot more resilient
[41:31] Speaker 17: than investors are out there believing. He thinks Microsoft three
[41:35] Speaker 17: hundred and sixty five suite with Word and Excel or
[41:39] Speaker 17: two of the most valuable franchises in enterprise tech. So
[41:43] Speaker 17: really Bill a sentiment from Bill on Microsoft, but that's
[41:47] Speaker 17: not what everyone's thinking.
[41:47] Speaker 2: Yeah, is that at odds with some of the others
[41:49] Speaker 2: the Tiger cubs.
[41:50] Speaker 17: And that's absolutely So it's really interesting when you see
[41:52] Speaker 17: these divergences. So Tiger Global and Co. Two slash their
[41:56] Speaker 17: steak in Microsoft by more than half.
[41:59] Speaker 2: Wow, Loan Pine got out of it entirely.
[42:01] Speaker 17: So you see really different moves from these firms that
[42:04] Speaker 17: really do spend a lot of time investigating these companies. Now,
[42:08] Speaker 17: this stock fell twenty three percent in the first quarter.
[42:12] Speaker 17: It has rebounded some of that since.
[42:14] Speaker 3: What about just in general, whether Microsoft is the old
[42:18] Speaker 3: one out here in terms of everyone either taking one
[42:22] Speaker 3: one view or the other, or of the big moves
[42:24] Speaker 3: in other companies.
[42:25] Speaker 2: So we did see a.
[42:26] Speaker 17: Decent amount of selling in the first quarter from across
[42:29] Speaker 17: the Tiger cubs. We look at them because they tend
[42:31] Speaker 17: to focus on tech, they tend to be directional and intentional.
[42:35] Speaker 17: Keep in mind, the first quarter incredibly volatile. This March
[42:38] Speaker 17: was a wild month. A lot of funds lost a
[42:40] Speaker 17: lot of money. That could be why we saw some
[42:42] Speaker 17: firms trying to sell. We don't know when in the
[42:45] Speaker 17: quarter funds were selling, just because of the nature of
[42:48] Speaker 17: thirteen f's.
[42:49] Speaker 2: But we did see some other trends.
[42:50] Speaker 17: When we look at popular cells Amazon, this is a
[42:54] Speaker 17: stock that sunk about ten percent in the first quarter.
[42:57] Speaker 17: Tiger Global trimmed at stake so to cut Amazon by
[43:01] Speaker 17: about a fifth, and Loan Pine sold out of Amazon entirely.
[43:05] Speaker 17: Berkshire Hathaway also exited that company too, So you do
[43:09] Speaker 17: see some big moves in some of these localized companies.
[43:12] Speaker 3: For sure, I'm fascinating because some of them ever since
[43:14] Speaker 3: written up higher on the back of that exactly Hemma
[43:17] Speaker 3: Palmer always breaking it down for us. We so appreciate
[43:20] Speaker 3: all the moves from the regular refindings. Meanwhile, that does
[43:22] Speaker 3: it for this edition of Bloemberg Tech. Tune in later
[43:24] Speaker 3: today again, do not miss this exclusive interview. We can't
[43:27] Speaker 3: tell you about it enough CEOs of Dell of Nvidia.
[43:30] Speaker 3: They are live from Dell Worlds at to thirty pm
[43:33] Speaker 3: Eastern eleven thirty am. If you're over on the Pacific side,
[43:36] Speaker 3: don't forget to check out the podcast too. You can
[43:38] Speaker 3: find it on the terminal as well as online on Apple, Spotify.
[43:41] Speaker 2: And iHeart.
[43:42] Speaker 3: From New York and from Las Vegas today, this is
[43:44] Speaker 3: Blueberg 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.
