Bloomberg Audio Studios, podcasts, radio news. This is Bloomberg Intelligence with Alex Steel and Paul Sweeney.
The real app performance has been in US corporate high yield.
Are the companies lean enough? Have they trimmed all the fats?
The semiconductor business is a really cyclical.
Business, breaking market headlines and corporate news from across the globe.
Do investors like the M and A that we've seen?
These are two big time blue chip companies.
Window between the peak and cunt changing super fast.
Bloomberg Intelligence with Alex Steel and Paul Sweeney on Bloomberg Radio.
On Today's Boomberg Intelligence Show, we dig inside the big business stories impacting Wall Street and the global markets.
Each and every week we provide in depth research and data on some of the two thousand companies and one hundred and thirty industries our analysts cover worldwide.
Today, we'll look at a big deal in the M and A space that will create a mortgage behemoth.
Plus we're going to discuss the rising costs of food and wow that's effecting the consumer.
But first we begin with research. Bloomberg Intelligence recently put out on ten companies to watch for in the second quarter of twenty twenty five and for more on this.
Liz Paul and I were joined by Tim Craig, head Bloomberg Intelligence Global Chief Content Officer.
We first asked him to give us a broad scope of the company he's looking at.
So just as a reminder, these are all based off of focus ideas, which for us in Bloomberg Intelligence are high conviction out of consensus views where we see catalysts ahead that can actually change the market's mindset around these companies. So you've talked about the weight on consumer sentiment and things along those lines. Well, Dollar Rama, which is a dollar store up in Canada has yeah, exactly, They're going
to face slower wage growth, higher inflation. We've not seen estimate cuts yet for Dollar Rama, like we've seen a dollar Tree and dollar stores and whatnot. We think that's coming. Another one that is interesting if you want to throw on as well the whole tariff concept is PDD. You might not have heard of this, but this is one of the big Chinese companies that trade tech companies that trades in New York. They own Timu, which is that
ultra low priced e commerce platform in the States. There's their bigger platform is one in China, they're going to have constraints on what they can bring in at the right price, and they're also investing pretty heavily outside the US to grow their non US platform beyond China. So again we think that there's estimate cuts there to come.
Hey, Tim, One of the more looking at this list, one of the more controversial calls, I think is Tesla. Steve Manner, the analyst for Bloomberg Intelligence covering the auto business, he's pretty positive on Tesla here despite the stock price in there. I guess the concern about Elon Muskin is focus.
Yeah, you know, you could throw this in with tariffs, with policy, et cetera, and clearly it's taken a shellacking from a stock price perspective. We see two things going on here. Number One, there's been a lot of talk about how their sales have been disappointing as of Leyden. Is that because of politics, We think, frankly it's because of the model transition with the new Model Y coming out, orders even in China have been good, and we would expect to see sales to pick up as we proceed
through the second quarter and on ind to midyear. You've also got a really underappreciated battery storage business that is also starting to ramp up. So we see two catalysts here to play out that people aren't focused on because of all the other noise.
And here's one for Paul here Cracker Barrel, Oh yeah, is on your list again. I've never been.
Now, Tim, he lives in London.
You go to New.
York City, man like, I don't know what I'm going to see.
Tim grew up in southwest Virginia. He knows.
I mean I'll eat it. If you buy it and eat it, I mean give it to me, I'll eat it. Let's put it that way, all right, talk to us about.
Really good I've crack. Cracker Barrel falls into the restructuring reorganization bucket. And there's another one on the list as well that actually may be good for you, Alex. I'll come back to But cracker Barrel three years of estimate cuts. There's next to no buys on this on this stock. And from an underlying business perspective, they've made some wholesale revamps, the menu, their approach to service. Uh, they're remodeling stores, and we think that there's an inflection to come from
the standpoint of earnings trends. Notwithstanding the economy, this one's already been beaten down, and we think that there's an up decided to come Carrying's the other one carrying? Yeah? So again three years of estimate cuts, you think who is carrying? Well, their biggest label is Gucci. There you go, Alex and if.
I only discount him, let's be clear.
Yeah, So new CEO at Gucci, a new head designer at Gucci. We think there's a rejuvenation at hand, and it's not in the estimates.
So again, Tim, let's just step back a little bit. European markets really performing well, certainly relative to the US market. What are your clients saying here?
So I think there are a couple of things coalescing. Three. In fact, you take tariffs and all the concern that's going on. You take deep seek and that raising the specter of g Is there a different dynamic in the world of tech that's been driving US markets and US exceptionalism. And then you have a big European defense initiative and imperative with Germany coming out with what people around here are calling the Bazooka. And you add these together, there's
a growth idea developing in Europe. There's concern about what's going on in the US. And if you look at our economists, they see as much of a hit on US growth and inflation as they see it anywhere outside of Mexico and Canada. So you know, all of that weighs on the US and Europe is an alternative, and obviously we've talked a lot about how it's valued relative to the States. I think that creates a money flow issue, a money flow opportunity, just like it does for China.
All right.
Thanks to Tim Craig had Bloomberg Intelligence, a global Chief Content Officer, Each week we look at research from Bloomberg n EF previously known as New Energy Finance. They're the team at Bloomberg that tracks and analyzes the energy transition from commodities to power, transport, industries, building, and ag sectors. This week we looked at the growing enthusiasm for nuclear power for more.
Guest hosts Isabelle Lee and I were joined by Chris Gaddomski, Bloomberg and EF lead nuclear analysts. We first to ask Chris what his thoughts are on nuclear power as a source to fuel AI expansion.
It's a great technology for supplying clean, carbon free twenty four to seven base load power. However, there's a mismatch between the demand for the electricity. Data centers say they would like to have the electricity right away tomorrow, if not sooner, and any new nuclear capacity in the US is not going to come online until after twenty thirty. I mean you may have one or two outliers come on beforehand, but before you have a reliable demonstration of
the technology. Because these are all new technologies being built, we'll be looking for after twenty thirty, So possibly we could miss the first wave of advanced reactor contribution to the AI developments.
Why is nuclear construction in the West barely budgeting, whereas, in for instance, Asia Pacific region it's booming.
There's two big reasons for that. One, the price at which the Chinese, for example, can build a nuclear power plant is much lower than we can in the US, and that's the function the fact that China right now is building twenty eight nuclear power plants. We have built two in the last ten or fifteen years, and so they will build six nuclear power plants at the same site, and the construction team moves from one reactor to the next reactor to the next reactive reactor h reactor subsequently
costs less than the first. So to China to build reactors for maybe one fifth to one quarter of what the US can build, it makes sense for them to build a lot of nuclear power plants. We've lost the x tease and the desire to build a lot of large nuclear power plants. To cite some advantages and significant benefits to the technology, talk to.
Us about small modular reactors. Is that a solution.
Yeah, it is a solution because the not dynamic has has has changed in the past. We've seen let's build large reactors to get economies of scale in size. The conversations talk, well, let's build a lot of small reactors and get down the learning curve as quickly as we can. So there are only two small modi reactors operating modern ones operating in the world, one in China, one in Russia.
China is building another one. Russia is planning to build another one, but they haven't yet broken ground on that. So it's a technology that presents a lot less risk to the utility or the hyperscaler who is buying it. There's no big deployment risk. There is technology risk because these are all first of a kind, and so with the first of a kind project, especially in the nucle power business, there's a lot of uncertainty with regards to schedule and cost, and so people are thinking about this.
A lot of the utilities of are making plans to go forward. The US and Canada the UK are leading the effort, and so it's going to take some time before those first reactors, small, major reactors, and advanced reactors across the finish line.
We know that new nuclear projects are coming online slower, But do you think this administration will change that or is it going to be status quo.
The Trump administration in its first time first term was very very pro developing uranium the equivalent of drill baby drill from the uranium sector, and to a certain extent that uranium has bounced back up. The current administration now is favorable towards nuclear. Chris Wright, the Secretary of Energy, has a pedigree from Berkeley and MIT to the leading academic institutions regarding nuclear, and I think that they are
very pro nuclear. However, the big question is how much government funding will be used to support nuclear reactor construction. That's something that we're still working out and trying to understand Chris.
I finished the first season of the TV show Landman, so I now consider myself an expert in the oil and gas business, and I think my takeaway is, you know, listening to energy providers and energy users, we're going to need everything. We're going to need fossil fuels, We're gonna need the wind, We're going to need solar, all that kind of stuff. Maybe nuclear. How do you think about that?
Yeah, my take is that we certainly do need nuclear. If you look on a geopolitical basis, you have China building twenty eight reactors, large reactors. Also, as soon as it passes are post twenty third, they'll have more nuclear
power plants operating than we do. The Russians are very, very aggressively building nuclear power plants not only in Russia but in Middle East and establishing one hundred year relations with that, and I think that has caused a lot of concern for the US government saying, hey, listen, we need to sort of offer a viable alternative option for new nuclear in the years ahead. And I think it's a very very solid ambition. Biden dedministration called for two
hundred gigawats of extra nuclear capacity. That's tripling what we have right now, and so there is a tremendous need for it, and I think there has a place. It's not a solution for all the problems, but it works very well for many types of applications.
Our thanks to Chris Gadomski, Bloomberg and EF lead nuclear analyst.
All right, coming up, we're going to break down the future of AI empowering data centers.
You're listening to Bloomberg Intelligence on Bloomberg Radio, providing in depth research and data on two thousand companies and one hundred and thirty industries. You can access Bloomberg Intelligence via bi go in the terminal Paul Sweeney and.
I'm Alex Steele, and this is Bloomberg.
You're listening to the Bloomberg Intelligence podcast. Catch us live weekdays at ten am Eastern on Apple, Cocklay and Android Auto with the Bloomberg Business App. Listen on demand wherever you get your podcasts, or watch us live on YouTube.
We move now to the M and A space.
This week we heard that the online mortgage provider Rocket is acquiring mister Cooper Group, the country's largest mortgage servicer it's.
An all stock deal valued about nine point four billion dollars, and this will create a mortgage behemoth that handles one in every six mortgages in the US.
For more, guest host Isabelle Lee and I were joined by Page Smith, Bloomberg News consumer finance reporter. We first asked Page to explain the latest deal and what mister Cooper is.
Mister Cooper is more is best known, I would say, in the mortgage servicing side of things, So it's sort of after the fact, you've got your mortgage, but who were you actually interacting with, you know, gears down the line,
it's actually going to be mister Cooper. But this is a big deal for Rocket because they've kind of they've been trying to position themselves as sort of a one stop shop for consumers finance offerings, so think credit cards in addition to your mortgage and you know, now theoretically your mortgage servicing rights. So it's it's a pretty big deal for this company based in Detroit. Dan Gilbert a pretty prominent person in the finance yep, pretty big in
sports as well. But nine point four billion dollars, it's a big deal for these folks.
We have Rocket also striking a deal to acquire Redfin, which is a real estate brokerage. How will that build into this one or if they're even connected at all.
Certainly it's it is basically sort of from a sort of a nose to nose to toe, if you will, of the home buying experience. Everything will sort of be under this Rocket umbrella. It you know Redfinn as you as I think a lot of folks will you know quite well, is this platform for buying and selling homes and it would be sort of a starting point for the home buying journey. Rocket is well known for originating, getting or creating mortgages, and now mister Cooper is well
known for sort of the servicing of mortgages. So it's it is kind of from start to finish of the home buying process.
Theoretically, they got a lot of debt on their bouncy because I being a former banker, I don't look at stock market values. I look at enterprise value to include the debt. It's eighteen point five billion dollars because that's what you get paid on. You get paid on enterprise a so that is a big, big number there. Talk to us about the mortgage business. I mean, is it doesn't seem like there's a lot of deals happening in the business.
Well, I mean, I think we can look at interest rates. It's certainly been a tough time for home buyers and sellers, and that also weighs on these home on these home lenders. It's a tough time for folks who kind of have put their eggs all in one basket, which is the home lending business. But that's kind of what Rocket has been aiming to do, is to really branch out and diversify their business so they're not so interest rate reliant or exposed rather which they have been in the past.
And frankly that's showed up in their earnings over the past couple of years. So I chatted with CEO U and Krishna over the summer. Actually we did kind of a deeper dive into Rocket and the company, and one thing that he told me then was that artificial intelligence is actually a big bet that they're making to really try to boost the business and make that home buying process truly as smooth as possible for the customers, to kind of amp up their offerings and we'll see how
it works out. With for them in the end, But so far numbers look pretty good.
What's this deal expected? Were you hearing chatter before? Did it's completely shock you?
I was not hearing chatter before, and it seems like the market is responding in kind of an interesting way. But we're going to continue to follow the story and see how this kind of fits into the broader rockets.
Who do they compete against? Do they compete against the banks that make the loans or the banks that make the loans usually just syndicate them away, right.
Well, I kind of take a different take on that since I cover consumer and finance kind of broadly. I think of folks who are kind of trying to be that one one stop shop for consumer the consumer finance experience. So for me, I think of you know, so FI technologies for example, that is like a they have a lot of lending opportunities for consumers, but they do a
heck of a lot of other business as well. Robinhood even coming at it from the investing side of things now offering banking products for folks, That's kind of how I think about it in terms of competitors. But that's just from a fintech consumer finance perspective when it comes to home lenders. You know, other big folks in the space,
the space. I was just looking at some data by Inside Mortgage Finance that placed United Wholesale Mortgagees number one, Pennymac is number two, and Rocket as number three.
Penny Mac is what I use. Anything that rings a bell? Okay, Dan Gilbert just for what it's worth. Kids on the rich top list of Bloomberg all the wealthy people, he comes in at number sixty two. Well they're not bad. Net worth of twenty eight point seven billion dollars up two point six billion.
Year to date.
Yeah, he's a force in Detroit. I would say some of our colleagues out of the Detroit Bureau did a great story on him last year. If you're interested in the terminal, All.
Right, thanks to Paige Smith, Bloomberg News Consumer Finance Reporter, we move next to the artificial intelligence sector.
Guest host Normal Linda and I were joined by John Lynn, chief business officer at Equinox.
Equinex is the largest global data center provider and is listed on the Nasdaq Stock Exchange under the ticker symbol.
E q i X.
John lenn joined to discuss the future of AI empowering data centers, and I began the conversation by asking John to explain what Equinox does.
We're really the fundamental digital infrastructure provider of the world, building two hundred and sixty eight data centers across seventy four market If you're the guys, we are the guys building and connecting all of these cloud provide and enterprises, making all of that data available for AI.
And it's really interesting because I was just speaking with him during the break saying that I cover US real estate stocks, I cover real estate investment trusts, and that's exactly what falls in that patch. This is my guy. So wonderful to have this conversation. But I think that there's often people don't really when you think about data centers, they don't often think about the people that are actually
providing the real estate for data centers. So can you talk a little bit about how equinics differs from maybe for thinking about a data center itself, but more or less the fact that you guys are acquiring properties and doing it that way.
Yeah, you can think about it as full scope development, where I mean we're going from raw land getting their entitlements and then building the entire data center and then operating that for perpetuity essentially. And our focus is around making sure we're getting as many customers as possible into the facilities and really interconnecting their data flows together, which is pretty unique in the data center space, which has also been a great opportunity for us to participate in the AI growth.
What are you guys seeing here? What are you seeing from your client and the people you talk to about kind of their needs going forward? Because right now, I think in the marketplace, if you look at like Nvidious stock and some of the other stocks that trade around the AI theme, twenty twenty five has not been a good year, after obviously phenomenal extraordinary growth in twenty three
twenty four, maybe before that. How are you viewing the growth here in AI and from your end of the business, the real estate side.
Yeah, First, I'd just say, you know, AI is a portion of the demand for data centers, but data center as a whole are powering everything that everybody is doing, right like listening to this broadcast, you know, ordering food for lunch, you know, trading, trading on the exchange, et cetera. I mean, you need computers for everything nowadays, and that's
still continuing. I think, you know, digital transformation is not in the early stages anymore, but we're far from done, and so that is just the secular driver that will continue. From the eye landscape, I'd say obviously a huge amount of interest and excitement and I think the it caught the imagination of everyone, and I'd say, right now what we're seeing is like exciting use cases that are really
providing durable value, right. And I think it's still early stages for many of that across the general business landscape, but that's what gets me fundamentally excited. You look at a company like a Bristol Meyers squib a customer of ours. They're doing drug discovery using videogpus and like being able
to increase and accelerate their time to therapeutics. That's fundamentally going to improve like human life, right, and I think that there's so many different aspects that AI can improve based on that.
So run us through some of your biggest customers, who do you all work with.
Certainly the cloud providers are some of our top customers. We've got over two thousand different network providers as well. The Genesis of the company was really around how do we help the Internet scale? And that ended up being, well, how do we help the globes telecommunications and data flows scale. And so when you think about all of the cloud providers, how do they connect to the end customers, that's through
our facilities. And then as we've built that landscape, we've ended up basically becoming the place where enterprises put their most trusted assets. When you think about then, whether they have some workloads that are in the public cloud, well, they're going to have some that are going to have ownership and control of themselves. When they put those in our facilities, it lets them glue that infrastructure together and become like one super powerful environment.
And folks, Equinics is a publicly traded company. Eqix is the ticker. It's got a market cap of eighty one billion dollars. And if you want some research on it and you're on the Bloomberg terminal, Jeffrey Langbaum was my reat analyst. He covers eqix. You can go big and that's where you find the research on Equinox. John talk to us about the global formfront and we know you guys are global here, where are you seeing growth stronger growth versus weaker growth.
Yeah, I'd say across the landscape, there's still quite a bit of demand for data center activity. You know, we're particularly excited about some of the emerging markets Southeast Asia for instance. It's certainly growing quite a bit. But you know, based off of a lot of the recent Surgeon AI and kind of the use cases set up for that, just a tremendous amount of growth in the US over the course of the last two years.
So, I mean, when we think about your creditors in this broader landscape, there's obviously digital realty trust when we're thinking about publicly traded routes here in the data center space, and if you look over the past five years, we have equinic shares that have risen forty percent, but that's compared to digital realty that's risen about eleven percent. What do you think that you all are doing differently than your competitors.
I think one, it's our focus around really driving diversity of customer and like kind of having an ecosystem that we've built around the value that we're doing, and so
that's incredibly important for us. Like for the AI trade, for instance, we're focusing not just on capturing some of these large training footprints, but really, how do we make sure we're getting all of these AI players and exposing them to the rest of our customer base and really, again that fuel becomes additional growth across our entire portfolio.
Do you develop and build data centers or do you just buy existing we develop and build? Where are you developing and building these days? And if you say Texas.
Or Florida, Well, we're building all around the world. I think we've got sixty eight current like major construction projects across it. Yeah, so it's we're very active.
Wow, how about it?
It's really it's a big company in this. I mean, yeah, people think about you guys have to come to my path.
It's a beautiful space.
So what is your you know, what are your thoughts for people who are saying that, you know, the tech rally has run too far? You know, maybe we have CAPEC spend that's just you know, bloated. There's so much spending in this space. Is this a place to be investing right now? When we think about AI and places of that.
Regard, I think the long term trend around this is going to be inevitable, right, I think it's certainly we're creating durable value, not just for you know, kind of the planet and all of our customers, but but for shareholders. I think the the amount of investment in the space, and like the numbers are candidly like eyewatering right now. And so but a lot of that I think is
just capital accumulation rather than deployment. And you know, compared to a lot of other markets in the real estate side, it's actually a little hard to kind of overbuild just because there's so many natural like limiters in terms of the way we want to scale, from utility power availability, to supply chain to you know kind of just the amount of trades you need to be able to build
and operate these facilities. And so I think that that helps kind of provide more rationality than you know, in some some real estate markets where you know, you can throw up a shell pretty easily, you can you can kind of just like convert and overbuild. In this case, it's a very long development cycle, and so I think you'll you'll see kind of some self metering there.
Our thanks to John Lynn, chief business officer Equinics.
Coming up in the program, we're going to break down the rising cost of food and how that's affecting the consumer.
You're listening to Bloomberg Intelligence on Bloomberg Radio, providing in depth research and data on two thousand companies and one hundred and thirty industries. You can access Bloomberg Intelligence via b I go on the terminal. I'm Paul Sweeney and.
Am Alex Steele, and this is Bloomberg.
You're listening to the Bloomberg Intelligence Podcast. Catch the program live weekdays at ten a m. Eastern on app Cocklay and Android Auto with the Bloomberg Business App. You can also listen live on Amazon Alexa from our flagship New York station. Just say Alexa play Bloomberg eleven thirty.
We turned out to the restaurant industry this week. Guess so was normal? Lindon I were joined by Michael Halen, Bloomberg Intelligence senior restaurant and food service analysts.
He joined to discuss the rise and costs of food and how that's affecting the consumer.
First asked Michael to break down his most recent research on restaurant spending.
Restaurant space is unique. We were in a restaurant recession last year. Higher prices, Yeah, higher prices really kind of you know, the you know, the low income consumer kind of pushed back against higher prices.
Right.
QUSR has been raising their prices since twenty twenty. The rest of the restaurant industry has been raising prices since twenty twenty one. So last year was kind of that restaurant recession. This year, you know, we see you know, higher income consumers with a better balance sheet. Right, Crypto's up significantly. Home prices are still rising, right, and so you know, for that reason, we think that this is going to be a better year. And we still think so in the data that we've seen and from a
lot of the CEOs that we've spoken to. The weakness in February was broad based, and to me that's less concerning. I would be more concerned if they said, listen, low income consumer pulled back even harder, the middle income consumer started to pull back harder, that would be more concerning to me. Broad Based tells me that, well, it was really cold. We got snow all over the country, including Sarasota, right,
and New Orleans. It was the coldest January in thirteen years or so, right, and the flu was really bad. So for me, it really seems that people were just sick of the weather and sitting on the couch waiting for things to open up. We have some weekly data for early March, and data got better.
So what data are you looking at when it comes to consumer health? I know you said there's some that you don't really pay as much attention to.
Yeah.
For the restaurant stuff, we use black box intelligence. We get very good industry level and sub segment level seems source sales traffic and check data. When I'm looking at the consumer, you know, it's kind of dated right now because it comes out quarterly, But I'm looking very closely at credit card balances and so credit card balances, credit card delinquencies, autal on delinquencies. They are still rising, but at a much lower slower pace than they were early
last year, so it's a rate of change improvement. We're also looking at CPI declining, real income's rising, and savings rates rising, right, and so to me, those are all good things for the low income consumer, right, So yeah, we're not so concerned about that that consumer sentiment data.
How about terraffs, How does that impact the average restaurant if they're buying food and that type of stuff.
Yeah, Listen, if you're a mom and pop shop and you're importing you know, Italian or Japanese items, stuff like that, it could hurt.
Right.
For most our chains, they're sourcing a very large majority of their products in the United States. You know, Chipotle was one everyone was worried about their saying, it's gonna be like thirty basis points impact to their food.
So the block prices aren't going to go crazy.
Yeah, it's gonna be like thirty basis points for the Alvocado prices, right, And they've for years they've been you know, expanding beyond Mexico in terms of sourcing. One of the companies that we cover that has probably the most exposure overseas is Darden. They said they actually import about twenty percent of their items. Part of that is they have an Italian chain, right, but also it's just cheaper for them, and so they're working on sourcing domestically and in other
places to try to ease some of that pain. So even though it's a twenty percent number, it can be much lower than that, So the restaurants aren't going to be impacted that harshly.
How are consumers adjusting to the price increases. Are people just saying, Hey, I want to go out, I want to eat, so I'm just going to pay more even sell or are they a bit more resistant?
If so, witch groups, it's a case shaped recovery, right.
And so chains like Chilis, I mean, they just posted a thirty percent comp in the United States over a five percent comp. We've never seen that before for a chain that's been around as long as But they're bringing in younger consumers, wealthier consumers that are willing to spend, right. And so the top slant of the K, people with money, they're doing just fine, and they're still spending at restaurants, right.
It's the people on the bottom slant of the K. It's these chains that are catering to low income consumers, right, that are getting that pushback. Either they're you know, going to restaurants less frequently, they're doing more shopping at the grocery store, or and when they do go to the restaurants, oftentimes they're ordering off the menu, or they're ordering less drinks, apptizers, desserts and stuff like that.
Hey, Mike, how about labor real quick? Thirty seconds, bus boys, dishwashers, migrant labor migrant labors cut off here is that can be a problem for some of these restaurants.
Well, it's a concern for the restaurant and industry in general. Most of our companies are compliant, they are right, and so they're not really too worried about it.
Our thanks to Michael Halen, Bloomberg Intelligence scene restaurant and food service analyst.
This week, Bloomberg Intelligence hosted its fourth Generative Artificial Intelligence Conference and there were some great lineups in terms of how you apply and mag jen AI for more.
Alex and I were joined by Julie Choice, CMO of Sarah Bros. We first to ask Julie where she thinks we are with the changing views of AI.
Well, I think that AI is really kind of hitting the mainstream more. I think that you know, chatchipt surfaced at the end of was that twenty two? I mean technically, I think the first model came out in twenty twenty two, and it's been almost two and a half years since then, and chatchipt has found its way into so much of our lives, and so people are more comfortable with AI, right, and so now I think we're just shifting into okay, AI can be a part of my life. But it's still early days.
So where do you sit then, in the in the pyramid for AI?
So we are Cerebras makes this beautiful chip which is basically AI infrastructure. We are the equivalent of Nvidia, so in video GPUs, it's just an alternative type of processor for AI. So we are the underlying compute that's powering the training and running of models like CHET GPT.
I guess one of the questions now is people are trying to just get a better handle on what the compute needs are going forward. Gensen Wong at Nvidia remains extraordinary bullish about that, and he has been, you know, the voice of AI for many investors for the past couple of years. Where do you think we are in that compute need?
Oh my goodness, I completely agree with Jensen Hang. I've been following Jensen kind of unofficially as a mentor. I find him to be an incredibly inspiring leader. I agree. I think that the compute needs for AI are ever increasing now with models like GPT four O which and the whole one series. These are like these reasoning models. Right. I'm here in New York to go to the generative AI scaling event that Bloomberg is hosting, and it's all
about this like inference time. Compute. Inference time requires a tremendous amount of compute, and so we're still just scratching the surface of how much compute is really needed for this extra level of intelligence.
And do you play in the inference and the LM space like the training and the usage you do both?
Yes?
Okay, yes, so cerebversus in both training and inference.
What is has demand for your products changed at all since deep Seed came out in terms of pricing or in terms of demand.
Yeah, I mean deep Seak happened in January and it was such a big moment for the industry and we immediately within forty eight hours of the news that big day, we added deep Seek to our catalog of models. And so what we do is we offer Lama models, Deep Seek models, and other kinds of models to the developer community right in our cloud. And our differentiation from GPUs is that we're like twenty to seventy times faster in terms of like that response time when you put in
a query. Our responses are about twenty to seventy times faster because of the architecture being bigger, So deep Seek really created this moment of developers coming at us saying, oh my god, you guys have the fastest deep Seek. We want to you know, we want we want that, we want to see what that can do. And that's led to a lot of prototyping. But what we do see is that when it comes to developers that are building AI businesses, there's a lot more usage of Lama.
So Meta Lama is probably still the most popular open source AI model that is downloaded at least from the Cerebraus cloud. And then we have customers like Perplexity and mistral Alpha Sense based here in New York that are kind of serving their adapted Lama models as well as their own custom models.
So the deep Seak issue for the industry was, Hey, here's this Chinese company coming out with an AI solution at a much lower cost. Is that good or bad or neither good or bad for the AI evolution?
Oh?
I think it's very good because deep Seek was a state of the art model when it came out. It's still one of the best models in the world, especially I think with their at V three. And what it proves is that when you can open source this level of intelligence, You're bringing down the cost to developers and it increases developer creativity. And so one of the things that Cerebras is super passionate about is providing developers with the best models open source or custom you know, fastest
at the best price. And so we were just very excited when Deepseak open sourced, and clearly like the developers were very excited as well. It was definitely kind of an inflection moment.
In terms of what you're most excited about right now, Like, what's the coolest use case you've seen when it comes to training and when it comes to in fronts.
Fantastic question. So for training, I'm going to bring up a healthcare use case. So one of our favorite partners, and you know, we shouldn't have favorites, but I really appreciate the work that Mayo Clinic has done in Cerebras. They have been able to train world leading foundation model using genomic data, and the point of their genomic model
is to help patients of rheumatoid arthritis. I have a family member who's been struggling with this terrible disease for a decade, and this AI model can help her find the medicine that actually works much more quickly. I didn't realize that actually rheumatoid arthritis generally forty percent fail rate in terms of aligning the right medicine to the patient.
And so really proud of the work that the mail Clinic team has done in partnership with Cerebris, training that model super fast on our chip, releasing it in probably less than nine months of development time. Wow on the training side. On the inference side, I'm very excited to be working with companies like Perplexity as well as Mistral.
These are some very very AI forward companies leading the way in terms of disruptive search experiences for consumers and amazing chat assistance, especially in Europe, and we're powering their super fast results.
Your company, are you seeing any impacts of just the uncertainty surrounding how this TWERFF policy will evolve? Is it and your customers saying, you know, we're just going to wait a little bit before we place an order or something like that.
No, Actually, it's really all systems go Okay. Our customers want the fastest inference speeds, they want the smartest models in their applications. We're not waiting to find out. There's no time to wait. We have to deliver the fastest compute to the best Custos commers in the world.
Our thanks to Julie Choi, CMO of Sarah Bross.
This is the Bloomberg Intelligence Podcast, available on Apple, Spotify, and anywhere else you get your podcasts. Listen live each weekday, ten am to noon Eastern on Bloomberg dot com, the iHeartRadio app tune In, and the Bloomberg Business app. You can also watch us live every weekday on YouTube and always on the Bloomberg terminal