Bloomberg Audio Studios, Podcasts, radio news. This is Bloomberg business Week Inside, from the reporters and editors who bring you America's most trusted business magazine, plus global business, finance and tech news. The Bloomberg Business Week Podcast with Carol Messer and Tim Stenebek from Bloomberg Radio.
Going through the Paige Book, the anecdotal summary of what is going on when it comes to overall economic activity in the Fed's districts. Economic activity grew slightly in three districts, while the number of districts that reported flat or declining activity rose from five in the prior period to nine in the current period. Employment levels were study overall, though there were isolated reports that firms filled only necessary positions,
reduced hours in shifts, or lowered overall employment levels through attrition. Still, reports of layoffs remained rare. On balance, wage growth was modest well Increases in non labor input costs and selling prices range from slight to moderate. From all on this, we turned to Stephen Blitze's chief US economist at Global Data TS Lombard. He joins us here in the Bloomberg
BusinessWeek Studio. Important to keep in mind this is anecdotal evidence from the FEDS districts, but it does give a good idea of changes from different periods of time and how employers are feeling about the economy. What are your takeaways?
Well, I wouldn't downgrade the fact that it's anecdotal information, because if you go back to the time leading into eighth nine, they valued and they learned in their sort of you know, Monday morning quarterbacking of going in the value of anecdotal information. And I'm really not surprised by what this said because if you look at the economic data that's coming out, right, and I expect the Friday employment number not to be particularly surprising either direction.
So in line with expectations.
Yeah, in line with expectations in other ways. They don't expect minus one hundred thousand or something like that. But what this number tells you, This tells you why Powell said what he did at Jackson Hole making a for a FED chair, a very blanket statement saying you know it, the easing of labor markets stops here, all right, And anecdotal information is very important to them at turns in the cycle because they lead the data.
So how are you thinking about all the data that's coming in right now, because it really seems like we're in a situation where the market at least is very sensitive. They don't want to see, for example, a labor market that's too hot. They don't want to see when that is too weak. How are you kind of putting all of this together?
Well, I think on the edges, right, there's no recession, right, and so on the edges you see weakening economic growth. The July personal income data is actually pretty good, especially on real spending for discretionary items in other words, not
counting food, energy, rent. So it's actually pretty good, right, So this is really about this is unusual for most cycles, and that this is really about the FED easing before recessionary data takes place as opposed to the usual, which is catching up, right, And they don't want to be in catch up mode, right, And by anyone's model, the funds raise at least one hundred basis points too high. So the real question is how fast do you get
down there? And the data will determine that the market is figuring that once they do the twenty five, the data will soften enough to scare the market to do like a fifty in December, and then another twenty five in December, and that's how you get your one hundred right. The moves after that, right is really in anticipation of the funds rate going down to adjust for an inflation rate by the next year it's two percent instead of three.
I'm looking at labor markets. In the beij book, it says that employment levels STEVE were generally flat to up slightly in recent weeks. A few districts reported that firms reduced shifts and hours, left advertised positions unfilled, or reduced headcount through attrition, though accounts of layoffs remained rare. Employers were more selective with their hires and less likely to expand their workforces, citing concerns about demand and an uncertain
economic outlook. Is this coupled with the Jolts data that we got earlier today, Is this a return to normalcy?
Well, that's that's the question, right, and that's always that's been the question all along as to whether or not this is just a return to a normal economy or this is a precursor to recession.
What do you think?
I think that in what I'm reading here, this is typical precursor of recession.
He said, we're not in a recession right now.
No, we're not. We're not in a recession now to be very clear. Right, but before firms start to lay people off, right, what do they do, just in our own experience, right, they slow wage increases, maybe they don't give wage increases. They cut travel. Business travel is an important you know, ask you know, hotels what they're seeing in business travel, not vacations, but business travel.
Uh.
And then they cut hours, and then there's unfilled positions and they say, you know what, we're not going to fill this position.
We've been hearing for two or three years that a recession is just around the corner. Why is what? And it hasn't it hasn't arrived.
Well, we haven't gone around enough corners yet.
Okay, maybe we haven't. What's different about this environment?
Well, the difference is that this is not a normal cycle because there was a mandated shutdown and then a mandated reopening which was COVID, right, and you had a tremendous amount of fiscal stimulus along with that, and so you had this unreal, unsustainable six percent growth whatever it was in twenty twenty two. So the economy is going from six down to two two and a half, so
which is its normal cycle. As the impact of all these stuff kind of washes through, so it's very hard to tell what is just returning to normal and when are you going past that? Now, the Fed really didn't get tight until earlier late last year, earlier this year.
And my favorite indicator is the inversion of the real yeald curve, so real three month yels versus real ten yure yelds, and that inverted last November, and that's about a twelve month till, you know, on average, which is always a dangerous number, but on average is about twelve months, which puts you into coming into recession at the end
of this year. But of course, you know, it's not necessarily faded if the Fed does what it wants to do in terms of easing financial conditions, and the key one is to actually cut the cost of short.
Term financing, which they're set to do.
Which they will do twenty five in September. Unless, as a unless there's a unless there's a big negative surprise, right, there's no upside number like let's say it's two undred and fifty thousand that's gonna keep them, especially given this anecdotal information that will keep them from cutting twenty five. So the question is twenty five or fifty, and you would need a really poor employment number to get fifty.
Steve, We're gonna have to leave it there. Steven Blitz, Chief US Economists at Global Data TS Lumbar joining us here in the Bloomberg at Business Week studio. Always good to see you.
You're listening to the Bloomberg Business Week podcast. Catch US Live weekday afternoons from two to five pm Eastern. Listen on Apple Cardplay and then broud Auto with a Bloomberg Business app, or watch US Live on YouTube.
Well Suprilla. It is a bound when talking about what happened with the video. Yesterday, shares fell more than nine point five percentator Race nearly two hundred and eighty billion dollars in market cap. It lost more than the market cap of Adobe, which is one of the thirtiest thirty biggest stocks in the US market. Shares hired by the
three tenths of one percent as we speak. Then, aftermarkets closed, Ian King and Leah Nyland exclusively reported that the US Justice Department sent subpoenas to Innvidia and other companies as it seeks evidence that the chip maker violated anti trust laws, an escalation of its investigation into the dominant provider of AI processors. We got with us Ian King, Bloomberg News US Semiconductor and networking reporter, joining us from our San
Francisco bureau. Ian. The dust has settled a little bit in the last eighteen hours twenty hours or so since you and Leah reported this story. As the dust settles, remind us what is that issue here and what the Justice Department is alleging.
I mean, the important thing is that this is a subpoena. They've been surveying in Vidia's customers and competitors. Now they've moved to a subpoena, which, as you know, is a legally enforceable requirement to provide information. So what we think in terms of what this signifies is that they are moving closer to wards launching a formal complaint alleging that in Video has try and gross rules here.
Remind us what Invidia has said thus far about these subpoenas.
Yeah, in response to our story, they came back and they were quite feisty. They didn't deny the facts. What they said was, look, people buy our chips because they're good. If they don't, they can go elsewhere. So really, you know, showing a degree of irritation, I would say with this, they believe in truth and engineering. They believe that they are where they are because they are the best engineers for this type of product.
So what is at the core of the DOJ probe? What part of Nvidia's supply chain are they really looking at here?
Yeah, I mean, we won't really know that until as a formal complaint filed, but our reporting basically shows that what they're looking at is the idea. And this is something that's happened in the chip industry before with companies like Intel, where if you make multiple products for a data center, such as what in VideA does you know
networking software? They supply software, models, services, and the chips say, if you only want to buy the chips, maybe you're paying more, maybe you're not getting the kind of supply that you want, compared to those that are buying everything. That's one of the examples that's that's sort of been thrown around. Maybe there's some kind of pricing differentiation going on that the customers don't like or think is unfair.
All of these kind of ideas when you have a company like in Video, which is so utterly dominant in one area.
Okay, well you bring up a good point in that it is utterly dominant in this in this one area, and you said, well, essentially the response has been if you don't like our product, you can go elsewhere? Can consumers? Can customers really go elsewhere?
Sort of? I think the problem. I mean, obviously, competitors such as Intel and AMD are really you know, it's taken them a while, and AMD is closer but still not really at the level that in Video is that, and Intel, by its own admission, it is kind of trailing far behind. So in that sense, no, but there are other aspects to this. A company like AWS, Amazon's web services, they do their own networking, they do their own servers, they do their own software. So maybe they
just want the chips. Maybe it's hot. I'm giving you a hypothetical here. Maybe it's difficult for them because they don't want to buy everything. So that's the kind of argument that exists that maybe some of these companies that have the expertise and the ability to do a lot of themselves want just the chips and want those chips on for fair terms.
What do you make of the move in the stock we saw obviously in video fall almost ten percent yesterday, but it's actually in positive territory right now. Is that move justified with the news that you're reporting.
I mean, I think you would have to ask my market's colleagues a little bit more about that. I mean, looking at what was happening yesterday, and this is before US went out, there was you know, obviously all of the chips related stocks sold off, and those that had been tied to this kind of AI frenzy were amongst the leaders, and obviously in video led everybody because of its market happ it had a huge impact on the indexes, and particularly chip related index So that had happened pre
to this. I mean, what we should do is probably separate the short term fluctuations between the long term impact of something like an antitrust investigation, which will take a long time and which will likely be appealed if there is an investigation, if there is a formal complaint, and if the courts decide in favor of the government's case whenever that is brought.
Okay, whenever that is Broughty in help us understand and help our audience understand what the next piece of this puzzle would be. What are you watching for?
Yeah, I mean there would be a formal complaint would be the next step we have to say in that regard these things typically have taken years. Actions against Microsoft, actions in the past against Intel, and quite a lot of them have you know, very little material impact on a company's fortunes because I guess what technology moves quicker
than the legal system. But at the same time, it's clear, not only in the US but worldwide that AI and who controls it, who controls these vital passways, is a clear, you know, strategic geopolitical security focus for governments around the world. So I think we could take this as part of a you know, a broader look at these things, and obviously an influence that governments feel they need to bring to bear upon this market.
Hey, in what we have you, we'd be remiss if we didn't ask you for your view on what happened yesterday when it came to the socks. When it came to the chip stocks, we saw all thirty components of the socks down yesterday. The socks had its worst day going all the way back to twenty twenty. There was this UBS note that came out yesterday. But other than that, why did we see such a decline in chip stocks yesterday?
I mean, it's look at the multiples, look at the run ups that we've had this year, last year. It's clear that a huge amount of money has been shifted into this sector, and it's happened very very quickly, in video being the prime example of that. When you get that kind of concentration, you obviously attract a lot of retail investors and a lot of people who don't potentially have the knowledge. So you have a lot of you know, what people will call fast money coming into a sector
which traditionally didn't really have it. So that's going to make any rollover, any sense that perhaps we've got ahead of ourselves a little bit even more precipitous. When people get nervous and when they pull back, that is one explanation of what happened yesterday, and we've seen that happen recently. And then guess what the money came rolling back in again.
So you would have to ask, you know, long term investors and short term investors what their viewpoint is and what their biases to me, that is the kind of explanation that I've heard that makes most sense to describe what's going on.
Well, we certainly appreciate that explanation, and we certainly appreciate all the coverage that you and the team covering Nvidia and other chip companies. Ian King, Bloomberg News US Semiconductor and Networking reporter, joining us from our San Francisco bureau. Now we're going to talk Intel because the Biden Harris Administration's big bet on Intel to lead a US chip making renaissance is in grave trouble. This as a result
of the company's mounting financial struggles. It creates a potentially damaging setback for the country's most ambitious industrial policy in decades, so writes Mackenzie Hawkins, Bloomberg News US Economic and Industrial policy reporter. She joined us from Washington, d C. Mackenzie Intel is supposed to receive eight and a half billion dollars in grants and eleven billion dollars in loans. That's as a result of the Chips Act from twenty twenty two,
but only if the chip maker meets key milestones. Where is Intel right now in relation to those miles stones?
So President Joe Biden went to Intel's factory site in Arizona in March and announced this massive grant and loan package from the chips Stacked and what could be and
would be the program's largest award. And Intel, like more than a dozen other companies that are supposed to receive US government subsidies, is in the middle of a pretty intense due diligence process that comes after reaching that preliminary agreement earlier this year, and once the company signs a final term sheet and hits milestones on each of four projects that Intel's Chipsacked Award is supposed to support, they could start to receive funding from the government in the
form of a reimbursement for companies spending. But Intel has resisted certain due diligence requests from the Biden administration and had a pretty disastrous financial report early last month that calls into question whether the company will ever be able to reach the milestones that would trigger that disbursement in the first place.
So what is the US really looking here when it's examining these milestones, Like, what exactly are the milestones that Intel has to meet?
So the government has been actually pretty quiet on the milestones. They're very bespoke arrangements that go on not just a company by company basis, but a project by project one. So in Intel's portfolio, for example, you have everything from an R and D facility in organ to a small packaging project in New Mexico, and each of those facilities have different timelines, different milestones, and different portions of Intel's
grant that would fund those projects. But across the board, one of the key metrics for the Biden administration is general commercial viability, and the best way for a company to demonstrate that is by lining up major customers. So if you compare Intel to a company like Taiwan's TSMC, which is widely regarded to have the world's best semiconductor technology and has orders from pretty much every major tech name in the books, Intel doesn't actually have any major
customers lined up for its US facilities. It has some interest from companies including Media Tech, Broadcom, and Microsoft, but none of them have reached full production yet. The Commerce Secretary Hers Steff actually reached out to executives out in Vidia and AMD and asked them to consider manufacturing it Intel's Ohio facility, which could become the world's largest chip making factory if it's fully built out, and neither company plans to do so.
Hey mackenzie, how did this story change just in the last month, Because at the beginning of August, Intel reported earnings and it was the worst single day in decades. As you note, two major credit raterers downgraded the firm's debt to just a few notches above junk. The chip maker also slashing about fifteen thousand jobs. How did this story change just in the last month.
So the process that Intel's going through right now, which is the preliminary agreement followed by months of due diligence ahead of a final award, was always the process that the company was going to go through. But of the entire chipsack portfolio, which includes companies like TSMC, Samsung, sk Heinex, Micron, Intel has perhaps seen the biggest change in the company position in the industry from when that preliminary agreement was announced to now. And so we're in a bit of
a holding pattern. The Biden administration and just about everybody else are waiting to see what the company will do, and they're set to evaluate options at a mid September board meeting. So if the company does something like split off its manufacturing division or scale back its global factory plans. That could have an impact on its negotiations with the
Biden administration. And to put a finer point on it, if and there is an indication yet that this is happening, but if the company does decide to pair back any of its US plans specifically, that would almost certainly cause a reduction in its overall chip sacked toward.
I'm wondering, if maybe we can zoom out, just for people who are just tuning in now, why has Intel not been able to kind of capture that frenzy that there is, at least in the market for semiconductors. You know in videas the top performing stock in the S and P five hundred. It's been a huge theme this year, and yet we have Intel, which is a very well known name, and it just hasn't been able to ride that wave. The stock is down something like sixty percent
year to date. Just get us up to speed on exactly what they're missing.
So, you know, Intel is an American technological giant. This is the company that gave Silicon Valley its name. But there have been several years of strategic and technological blunders that CEO pat Glsinger came into leadership of the company aiming to turn around with a massive bet on Intel's factory division for factories that are known as foundries in
chips peak. But the foundry business has struggled. They posted a seven billion dollar loss last year, and they missed out on a big AI wave that companies like TSMC
have been able to ride. And so now Intel is sort of playing catch up against foreign competitors and their technology still lags several generations behind the cutting edge, and they've struggled to convince major other players in the industry, who get the bulk of their orders for AI chips from the industry leader TSMC, to take a bet on
their factory processes. So it's a bit of a spiral for Intel right now as they're trying to convince investors, convince the government, and convince industry that their technology is worth taking a bet on McKenzie.
On the US side of this, the US government side of this, the support that Intel was set to get from the US government is set to get from the US government. What does the US government want to see? Is this about jobs? Is about security? Is it about both.
So if we go thirty thousand foot and look at the Chips Act overall, this was a landmark law that was that President Joe Biden signed just about two years ago, two years ago last month, and the goal is basically to bring manufacturing of these critical electronic components back to the US. Chips are in everything you could think of, from a microwave to a phone to a nuclear missile, and for decades the US has ceded manufacturing leadership to Asia.
The pandemic revealed to everybody how important ships are. It was a shortage of even less advanced chips that caused auto supply lines to shut down during COVID, and governments around the world, not just the US, but from the EU to South Korea, to Japan, to Taiwan to of course China has decades of industrial policy in this area are really doubling down and trying to make sure that they have a stable and secure domestic manufacturing base for
both economic and national security. The jobs, of course, are an added bonus, but this isn't a jobs program. The goal is to get the world's top chip makers, and the US has gotten commitments from all of them, including Intel, to build significant manufacturing presence on US soil so that the country can make a fifth of the world's most advanced processors by the end of the decade.
But you say, you write in your piece that this is a long process. I mean you go back to the Obama administration and Intel was talking with the Obama administration for something that they ended up building and touting during the Trump administration.
That's right. And you know, one of the big difficulties facing the bidening minist is you know, they're looking at point in time company financials and market conditions. But this is an intensely cyclical industry. And as much as companies would like to make promises to politicians, you know, they'd love to stand with Biden at a bill signing or go to the Oval Office, the reality is that they're
beholden to their shareholders and to their customers. And so you have a strategy employed not just by Intel but by other major companies as well, to build factory shells, do what has cost billions of dollars but is actually the cheap part of the process, and only outfit them with actual equipment when they can see, you know, six months down the line, nine months down the line, exactly
what their demand picture is going to look like. And so right now you're starting to see a lot of the signs that you would want to see if the country were going to meet it's, you know, a decade end goal of making all of these chips at home, but that's in factory construction, not in actual equipment purchases, and so there are a lot of other indicators to watch to see whether companies will actually make good on their production plans in the times that they have set out with the administration.
I wanted to talk about that timeline. What do we know about what's next for Intel and how much time they do have before they would be set to receive those billions of dollars in grants.
Intel would love to know the answer to that question. Very active negotiation. You know, there are companies that are quite close, but we haven't actually seen any chipsacked money
go out the door. All of the announcements of which there are more than thirty billion dollars committed, have been in the form of these preliminary agreements like the one that I described Biden announcing for Intel in March, and so all of the companies are going through due diligence, actively hashing out the specific dollars and cents that are correlated with you know, X production, milestone, X customer commitment, and it's you know, a very company by company basis.
And so really what Intel is feeling right now is not necessarily like, oh, we're worried our chips ACKed award is going away. That would only happen if their US factories went the way, and Intel hasn't made any decision that effect. But it's about the timing. Intel has a balance sheet that's in pretty dire straits and they want to get this money as quickly as possible. But that'll
only happen if they cooperate with due diligence. They have resisted some requests from the Biden administration so far, and of course if they actually get their factories off the ground, which requires the company to invest billions of dollars itself.
To what extent does the US government need Intel to succeed? I'm not asking if it's like too big to fail, but is it too important to fail?
Intel's crucial to the Overall Ships Act effort. There are only three companies that make the types of advanced processors that the US is really really focused on in this effort, and that's Intel, Samsung, and TSMC, and both of the latter companies are building factories on US soil. But to get to a fifth of the world's production, especially when governments around the globe are trying to do the same thing, you really need these large scale commitments like the ones
that Intel is making. I mean, they're planning multiple fabs in Ohio and multiple fab expansions in Arizona. So to potentially lose those projects or to see them delayed, which again we don't really know what's going to happen on that front, would be a pretty massive blow to the overall ambition of the program. And tells also the only American maker of these chips, which the Pentagon has decided
as really crucial for Department of Defense supply chains. And it tells the sole intended beneficiary of a three and a half billion dollar program to stand up a secure facility to make semiconductors for military and intelligence purposes. So will the Chips Act fail if Intel fails? No, there are hundreds of companies that were interested in the money.
More than a dozen awards have already been announced, but it's impossible to overstate how critical Intel's investments are to not just the company's turnaround plan, but also the Biden administration's goals.
Mackenzie Hawkins, Bloomberg News US Economic and Industrial Policy reporter.
You're listening to the Bloomberg Business Week podcast. Listen live each weekday. He's starting a two pm Eastern on Applecar Play Android Auto with the Bloomberg Business Ad. You can also listen live on Amazon Alexa from our flagship New York station Just say Alexa Play Bloomberg.
Eleven thirty data centers and transmission networks each account for up to one point five percent of global consumption. That's according to the International Energy Agency. Together, they're responsible for emitting about as much carbon dioxide as Brazil annually. So over the summer, we learn the new data center's plan for Silicon Valley of the potential that a three point five gigawatts of demand for electricity. This, according to PGNE,
it's more than the output of three nuclear power plants. Emily. Needless to say, all of this computing power is using a lot of energy. Josh Patterson is trying to do more with less at Voltron Data, where he's CEO and co founder. It's an analytics software startup backed by firms including black Rock, Alphabet's GV light Speed, and more. Josh joins us from Charleston, South Carolina. Josh, good to have
you with us. Welcome to Bloomberg BusinessWeek. You are adding VideA for about four years before you co founded Vultron Data. What was the problem that you saw when you were at Invidia that you set out to solve?
Absolutely?
At Nvidia, I joined specifically to drive GPU acceleration in data analytics. Prior to in Nvidia, I was at Accenture Tech Labs and we were working on cybersecurity challenges and what we realized is we had the algorithmic and systems technology to solve really large scale threat detection, but it was preheitively expensive moving you know, parts of the software
stack to in video. GPUs allowed us to shrink different parts of it down tremendously, moving from forty servers to a single server and being much more energy efficient and faster, moving you know, graph analytics from thirty minutes to thirty seconds. And so after that we really caught this this bug. How can we gp you accelerate all of data analytics how can we help bring GPUs and data systems. And that's what I joined in video for there, I created
the rapids dot AI ecosystem. It's all the modular and composed of building blocks for building software for data analytics, machine learning, graph analytics, et cetera.
In video GPUs.
And our whole goal was data is going to keep growing, and so we know we have to be able to analyze that data faster, cheaper and with less energy.
Okay, so help us understand maybe in I mean I get it. Come on, maybe simple terms. If you have to explain this to you know, someone your elevator pitch, or you know, explaining this to your grandma, what exactly is a GPU and what exactly is voltron? I guess producing here that is reducing the energy consumption that these big data centers take.
It absolutely, you know, and so the first thing I would do is always do it. You know, kind of an analogy. We we know for a fact that in video GPUs are the leaders in AI. AI would be you know, prohibitively expensive, It would be borderline impossible without the innovation. Uh, you know, from B one hundred to A one hundred h one hundred and now the B
one hundreds coming out soon. And you know, GPUs are essentially these multi core chifts that allow you to do parallel data processing a lot faster, a lot more energy efficient, and it's been the building blocks of graphics, high performance computing, autonomous vehicles and ai as.
We all know.
What we're doing is we're taking that exact same speed up and we're applying it to databases. And when you think about a database, it has three layers. You have how you store data, how you manage data, how you actually do compute on the data, and then how you talk to that data system, your ap eyes, how do you send instructions and so what we do at Voltron Data that is pretty unique is we run that database compute.
So not the data warehousing, not how you store data, and not how you index and sort data, but the actual how do you join tables together, how you do group Yes, think about all the things that you do.
In Excel, but just at a much much larger level.
Instead of you know, megabytes of data, we're talking about terabytes of data, and we do that running on in video GPUs. We primarily focus on that middle layer that's just the compute, and we allow people to bring their own APIs to our engine, and that allows that integration
to be a lot faster and more seamless. And so whether you're using you know, one of the myriad of SQL dialects, a Python data frame API, a Java data frame API like Spark, you can send your existing code to our query engine that then grabs data from your
data lake and analyzes that data really really fast. And so you know, essentially we can take workloads that used to take two hundred servers and we can run them on two servers and it's faster, and that's how we get to these you know pretty you know, amazing energy reductions today.
Where are the servers that you're running these queries.
On, both in the cloud and on premise. We have customers who do air gap on premise installations. So think about, uh, you know, your Department of Defense and your intelligence community agencies as well as you know, high frequency trading firms where you know, for a myriad of privacy reasons, they don't want you know, their data centers having free access
to the Internet. Traditional data centers that you would find in you know, Fortune five hundred organizations, your top banks, retailers, et cetera, as well as in the cloud, and so we can run our software across all major clouds.
The only prerequisite.
Is in video GPUs that is the backbone of what theseus runs on. And then from a from a kind of a big data how do you schedule and orchestrate it. We use Kubernetes, which is, you know, the foundations of cloud native technology. So we're we're cloud native and we're accelerated native. We're built from the ground up to fully leverage in video GPUs, accelerated networking and storage.
So how's business so far? What kind of customers do you have? And have you seen I guess an increase in them because I know you're a new company.
We're a new company.
We were found in twenty twenty one, but we didn't really launch theseus until this March at Nvidia GtC and that's where we release our benchmarks showing that we can scale to hundreds of terabytes of data with you know, single digit number of servers.
You know, business is doing well.
We've we've we've signed some really great deals with a few of the intelligence agencies, a very large commercial bank, one of your fortune one hundred, uh, you know, top five banks, and we also have about a dozen customers that we're doing enterprise support and advisory for around open soare software and so think about your you know, your high frequency trading firms, Snowflake UH and others or customers of our enterprise support and these two things work very
graciously together. So, first, we help people use existing open source software as efficiently as possible, and that is things like Apache, Arrow and IBS and substrate and park and so they can build these better data lakes, they can build interoperability, they can have these you know, more UH.
Efficient composable systems. But it also.
Allows them to more graciously upgrade to these cheap accelerated query engines when they're ready. And so both sides of our business really is revolving around how do people do more with less? How do they you know, continue to use their existing code, how can they leverage best practices? Is an open source to drive enterprise value today?
Josh, your company you're based in South Carolina? Is your company in South Carolina?
We are fully geographically distributed.
We have engineers from Japan to Australia the long way, probably about two thirds in the US, one third global, and you know pretty distributed across the US.
Is that pretty high con Do you see that as an advantage? I mean, when we talk about companies such as your Silicon Valley is the geographic location that certainly comes to mind for for a lot of these companies.
You know, So when our company was was born, it was born out of a kind of an amalgamation of some of the world's best talent in hardware engineering, software engineering, database design, high performance computing as well as you know in video GPU know how, and that talent is very rare, and so if we wanted to move everyone to one place, it would have been you know, it would have been expensive.
And so this was one of the ways that we we could really keep our costs as low as possible while building the world's best engineering team and allowed us to build this really amazing engine I would say, faster and more capital efficiently than you know, ever before Historically typically, you know, people spend four to five times as much
building a data system than we did. And that's really because you know, even from our time in and video, we were geographically distributed some of our leadership team, and we've been you know, fully remote, probably some of us since twenty sixteen, so even pre COVID, and so this was a working model that we were very used to. Is very common in open source, and we leverage you to our advantage.
Josh, come back and visit us again. Really appreciate you joining us. Josh Patterson. He is co founder and CEO at Voltron Data, a journal.
Now about you let me drive Oh no, alright, please, I'll do.
I want to.
It's good question.
This is the Drive to the Clothes on Bloomberg Radio.
All right, it's time for Drive to the Clothes with Thomas Raymond, investment management partner at Callan Family Office. He joins us here in the Bloomberg Interactive Brokers studio. Welcome, thank you for joining us. For me, so in video was down today, but the notes that you sent us had a lot to do with AI. Where exactly you're seeing opportunities in that part of the market.
Yeah, well, first off, perspective matters a great deal when the markets are chopping like this and you have a cell off, especially with the magnitude of Navidia's declined the other day. But Navidia is not another Netscape. There is a clear difference between that company and what we saw in the tech craise and the tech bubble, all right, I mean the earnings in revenue numbers that we've seen
are one hundred percent growth year every year. It's not the same, right, so kind of put everyone's mind at ease. It is not the same. But what we are seeing. One of the the theories that we've kind of had was that there's going to be a shift from AI producer to AI consumer and various benefits from that, and that could unlock a huge productivity innovation wave.
Where we are we seeing that yet with S and P favend companies. I mean every quarter, I'm like listening on these calls to hear about how a company that's not an Nvidia, for example, that's not actually producing the products or you know, building the services, that is doing this analysis, are actually using these tools to improve their bottom line.
Yes, you are in some unlikely places. Okay, so Walmart, let me give you an example. Walmart is not typically considered a tech company, but you look at what they said, their CEO said in their most recent earnings announcement, by the way, they outperformed on an earnings basis. They said that they were able to catalog eight hundred and fifty million of their items through the use of large language
models AI. It would have taken them one hundred times their workforce to replicate that work in just a fraction of a time. That's just one example. That is Walmart. Visa was another one. A couple of years ago. They saved twenty five billion in fraud detection from the use of AI. There are a lot of examples out there of companies they don't think are tech companies, but man, they start to look like them, and there is this push to be more asset light and more clear evidence
of this is earnings growth. Right in the first quarter, the MAG seven names at earnings growth of just those seven companies of fifty percent plus earnings growth year. Youer the rest the other four hundred and ninety three companies down two percent. In the second quarter, there's been a convergence, there's been a normalization of the earnings growth from the
MAG seven names and growth on the bottom line. The number of the delta now is a twenty some odd percent because the MAG seven areas growth came down, but the rest of there's other's other feign or ninety three companies started to go up. So it is an earning story. Right, the cost savings that you get immediately cruise to your bottom line when you embrace these you know, embrace AI.
So where does that leave big tech then, because you're still seeing a big appetite to buy those apples, Alphabets, Microsofts and when you look at you know how much of Nvidia's customer bases those companies, they do make up a large portion. So do you hold do you sell that area?
Well, there's time. Horizon is the controlling factor in this right for those that have the ability to embrace the volatility along the way tech is just it's not going away. It will be a bumpy ride. But let's just fast forward into a couple of years and all these upstream effects from AI. Take the ROBOTAXI. We're here in New York. I don't see any robotaxis just yet. Yeah, but they're in San Francisco.
You can't you can't walk a step, We can't a block in San Francisco without seeing a way mo.
Yeah, but that's not possible without AI and not we're not necessarily endorsing the tesla's and the way moves and the bodies of the world, but having staple positions in tech into those AI and those companies that are embracing it. It's it's something you need to have, but you got to have the perspective and the time horison to support it.
Okay, speaking of perspective, everybody's favorite topic politics, Uh, the election coming up. We've got a couple of months until. Wow, that's crazy, actually when I say it like that, a couple months until yeah, literally a couple months. How do you see that as a risk out there?
So I've heard there's there's going to be an election as well. Yeah, so and I and I'm in a swing state. I'm in Pennsylvania, So I hear every other every other ad is a is it as a political ad? How we're approaching it is? I mean, there's going to be it. It feels like a toss up. But to make an investment bet predicated upon a certain election outcome
never really works. Well. What we would reshape the narrative to is some of the sure things, right, some of the more certain items that are about to come about, one being interest rates, right. We know injuries rates are going to come down. The path is clear, whether it's twenty five or fifty, you know that that's up for discussion, but we know it's going to come down. We see FED fun futures. It's it's going and we have had a FED that's committed to transparency and open communications and
proactive communications. We know interest rates are going down. That's one certain that will play out irrespective of who wins in November. The other certainty is we're still going to see a lot of unspent government dollars right there's a trillion dollars. There's a trillion of unspent federal dollars that's still out there from the Inflation Reduction Act and things like that that's still coming. There's still going to be a federal spend that's going to influence the economy. So
there's some sure things that are out there. We would always advise not to have political ideology influence a investment decision. Tends not to work out well. Try to kind of reorient and make decisions based off those more certain things that we see coming.
We have a big payroll report coming on Friday, and we also got labor market data today that showed a weakening labor market. How are you thinking about jobs in the US right now and how it impacts markets?
Well, it's it's that derivative effect into into FED policy. And the reality is from from the economy. From an economic standpoint, what we saw when rates went up, We saw that the economy isn't as sensitive to rates as we once saw. There was twenty five percent of homeowners refinanced when we had ultra low rates. Now the opposite effect will hold true. We might not get that juice
when the economy when rates start to come down. From an economic standpoint, we will get it from a lot of you know, from a financial market standpoint, a lot of companies that at floating rate debt, typically small caps, smaller companies as compared to larger companies, they'll benefit somewhat more directly from that. So it's the economy from a rates perspective. And then there's the financial market effect.
You're saying, we might not see pent up demand because a lot of people already have low mortgages.
Correct, Yeah, okay, yeah, I'm one of them.
Yeah, a lot of folks out there. I mean, I think I forget the statistic. It's like something like eighty percent of people have sub four percent or something like sub five percent. I don't know it. It was I don't have the stat off the top of my head, but it is a large number of people and remember that's fixed.
So yeah, it's over twenty million homeowners refinanced to the ultra A low rate. So it's it kind of you know, that dynamic has desensitized rates to the broader economy at large and spending. So the jobs report it's meaningful, but in terms of you know where rates are going and that interplay not you know, I think we're there might be an overstatement of where things can go.
Thomas Raymond, investment and management partner at Callent Family Office, joining us here in our Bloomberg Interactive Broker's studio.
This is the Bloomberg Business Week Podcast, available Apple, Spotify, and anywhere else you get your podcast. Listen live week day afternoons from two to five pm Eastern Bloomberg, 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 Jerminal