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 Stenebeck from Bloomberg Radio.
Hi, everyone, Welcome to the Bloomberg Business Wee Weekend Podcast. Well, this past week felt like NonStop news from the world of artificial intelligence. Among the headlines, TSMC posted a thirty four percent surge in November sales, reflecting a sustained growth from AI demand despite concerns that data center building will slow.
Reddit launched an AI powered chatbot. Amazon announced it's building a MEGAAI supercomputer with Anthropic Apple is working on an AI chip with Broadcom, and Donald Trump's appointment of venture capitalist David Sachs as the AI cryptos are is seen as a big win for big tech.
Are you done?
No?
Oh, You're not done? Okay, keep going please.
Trump's pick to run the FDC, the Federal Drake Commission, Andrew Ferguson, wants to, among other things, promote innovation in the AI market.
You're done now right?
No?
Oh?
Okay, and then there's Alphabet hitting new records in terms of a share price due to excitement over major development and quantum computing through the use of its Willow quantum ship.
Okay.
Yeah, there's an AI angle there.
Okay.
The bottom line here a lot of AI related stories and happenings. It's like top of mind. It's a huge theme of what we talked about this week. It's kind of a thread under everything right now. So for the next two hours, we highlight some of our recent chats. No, not with chat bots. Sometimes we talked about that, Yeah, speak for yourself. Conversations in and around AI, though, including one with the author of a new book about the company at the center of it all, in Vidia, that's in our next hour.
Coming up this hour. Some thoughts also on how AI is reshaping the tech sector and which companies stand to game. We're going to talk about that with the president of IDC, Crawford del Prett.
Plus, tech giants excited about AI are pouring ever bigger sums of money into energy intensive data centers to power it, and yet not everyone is so excited. That includes one Blackstone funded data center developer that's getting a major resistance in one Georgia neighborhood.
All of that to come. We begin, though, with Cisco Systems. It's the biggest maker of computer networking equipment, a bell weather for corporate technology spending. Last month when it reported earnings, Cisco's CFO saying Cisco is seeing a strong rebound in spending by corporations across all sectors and geographies. It's also benefiting from spending on the machinery needed to support AI computing.
In October, Cisco expanded its data center infrastructure portfolio to sustain GPU intensive AI workloads with Nvidia Accelerated Computing, with the hope it will help Cisco customers accelerate their AI adoption. After earning, some analyst raised price targets on Cisco's stock, citing AI traction. So there was a lot that we wanted to talk about with frankeutsutis, executive vice president and
chief People, Policy and Purpose Officer at Cisco Systems. We kicked it off asking about the impact on business dynamics following the US election.
What I find is that as a company, we have worked with all administrations and we build strong partnerships and relationships in the broader community with governments, and so actually we're meeting with a Trump advisor to just do some planning and to understand a little bit of the priorities
in the next year as well. It's so interesting because this year there were sixty elections around the globe, and so I think around the world you have people stepping into new administrations and we're absolutely doing that and planning for that as well.
What did you could I just ask you to follow? What did you feel like we're the priorities you got or what can you share with us?
Yeah, you know, I think right now we're learning, and of course we're waiting. I think there probably is some planning from a tariff perspective and to understand, okay, what does that really look like. I think there's some conversations from a tax perspective as well. And then, as you know, there are other really important policy issues, whether we're talking about security or AI and I think we'll have to wait to see where that sits from a priority perspective as well.
When you think about your supply chain and the potential for tariffs, what are the conversations that you're having around that, How could you be affected?
Yeah, I think always from a supply chain perspective, and this came out of COVID. There's a level of agility and planning that is going on now that is like ten x what all companies were doing before. And I think we've learned that. And on the tariff side, kind of understanding what do the numbers look like where and what impact does that have from a cost perspective to Cisco and potentially to our customers.
Is it going to cost more?
Because you could, it could, And I think at this moment we just have to wait and see and probably not get too excited or nervous one way or another. Let's wait and se and then we'll react.
But is it worth thinking about changing your supply chain in the sense of finding different suppliers, moving manufacturing facilities, or having suppliers move manufacturing facilities as a result of these tariffs.
You know what I would say, tim, is we have been building redundancy even before the worry about tariffs. And I think right now you have to be so fluid whether you're responding to a disaster that occurs in a particular location, whether there are suppliers shortages. I think this is the new name of the game. And so I would say there's a ton of planning going on against a ton of different scenarios.
I got to ask you only because it's so front and center for all of us, and we're looking at what happened to the United Healthcare CEO. You are in charge of overseeing your people almost you know, more than ninety thousand around the globe. I am curious about some of the conversations and you know, some of what we've had maybe with friends, family are peers here. People are frustrated and not justifying anything, but I think people are trying to figure out how to be heard. So I'm
just curious how you guys are thinking about that. Are your executives scared?
I think the first thing that I would say is that there was a tremendous amount of shock and sadness for Brian and the family, and I think we start there. But to your point, the next conversation with our employees is always worry about security, and I think those are conversations that we've also had a lot of. And then
the secondary piece. I do think it's important for us always to have conversations about those pain points, and I think we're always looking at what is going on from a broader health perspective, well being, healthcare, and so I think the most important thing that we can do is drive conversations so that when we talk to companies, we're really clear on the priorities and what we're hearing from our people.
When you say talk to the companies, you mean in your role as chief people officer correct overseeing benefits at the company, getting feedback from employees about their own healthcare. I'll give us some examples how you've made changes in recent years to plans in that sense and what employees have said.
Yeah, you know, it's so interesting when I go back about a decade, I think about the fact that at that moment we were realizing our employees have very different wants and priorities, and so as an example, that was the time where we started realizing, hey, there's a subset of our employees that want more benefits around elderly care. And we talk a lot about the Sandwich generation, right, Yeah, and so it's a real thing. Yeah, you know, our belief at Cisco is that you have to be assessing
your benefits on a regular basis. And the nice thing is with some of the breakthroughs around things like can't your treatments. It has allowed us to try different benefits for our people. One of the things that I think Tim was so important to us several years ago was just introducing elements like second opinion support for our people. When something comes up, when something happens, that's the moment that you need support.
Of course, how has the conversation around corporate and executive security changed in the wake of this tragedy. I've read reports that you have not you, but corporate security offices. Phones are just ringing off the hook, calling private security consultants and security officials and the like.
Yeah, I would say that on a pretty regular basis, we're always assessing security for employees around the globe. I just think that's a good practice to have. But when you have a tragedy happen like what occurred last week, I think it does remind people how important it is to think about this, to think about those executives that are a bit more visible and what needs to be done.
Of course, I want to ask you about the business environment. It was tap you. You guys did report about a month ago. We talked about it coming into it. Tell us about the business environment. Is there a little bit of like wait and see in terms of how the policies all play out or how are you guys seeing it. I think you gave, you know, an upbeat forecast or conservative annual forecast, but an upbeat outlook for the current period. So how much visibility do you feel comfortable around?
Yeah, you know, I think it's really natural that there's going to be an element from a policy perspective that is wait and see. Okay, But I do think that when I look at some of our technologies, when you're talking about security as an example, and against the backdrop of a very dynamic cybersecurity world, that's something that our customers continue to prioritize, and that's something that's going to
be separate from any change from an administration perspective. I also think as it relates to AI, our customers are getting more and more prepared. I would say on the web scale side, they're in it, they're working. On the enterprise side, I think it's probably the prioritization of use
cases for the future. I don't know if you saw this, but we just put out our new AI Readiness Report, and what was so interesting was that we did one last year and What was consistent across the two years was the pressure that companies and IT individuals are feeling around AI readiness.
But where we didn't.
See any improvement is around having the talent for the AI play.
Which is really fascinating.
No, and especially because eighty five percent of the participants said that they feel pressure to have a plan within the next eighteen months, and so it just tells us there's a lot of practical work that needs to be done. From a Cisco perspective, our focus is on AI infrastructure and that's something that will create a lot of demand. From a network perspective, just goodness as we move forward.
But I would say internally, there's a lot of work that we're doing with our people, and our hope is that we can share this more broadly with our customers as well.
Are you having any issues finding employees right now?
So I would say there's always a shortage for great talent.
That's how it feels.
And so whats the forward or there's certain categories. I would say there's certain categories, okay, So I would say absolutely, from an AI talent perspective, people are building skills right now, and so what's really fascinating is we start with a huge funnel of talent, and what we're finding is that a small percentage of that funnel has the AI skills
that we're looking for. What that means on a positive side is that companies like Cisco have to build that talent internally and it's going to create a lot of opportunities for our people.
Feels like we're companies are going to be very involved right and making sure their employees have the skills. Two quick things I want to get to. First of all, can we tell.
That you're getting ready to go?
Sure?
Okay, what you guys are?
Sorry?
Sorry that there you just but your global company and you are getting ready to go to Africa. We talked a little bit about offline about India. What is going on in those markets that you think is really important for the Bloomberg audience to know.
Yeah, I'll start with Africa for just a moment. I think what we see there is that the demographics and the number of young people embracing technology creates a tremendous opportunity. And so from a Cisco perspective, there's a lot of work that we're doing around key topics like cyber resilience. We did a report several months ago that basically showed that the number of cyber attacks on the African continent continue to increase and in fact it actually impacted GDP
by about ten percent. And so you see tremendous opportunities around policy around skills. We see a shortage of security skills around the world, but on the African continent opportunities too in that only ninety percent of cyber professionals on the continent or women, and so there's huge opportunities there. So a big part of my visit in addition to the digital projects, is looking at how we can drive
digital skilling on the continent. And then I would say from an Africa perspective, there's tremendous opportunity and growth and it's exciting to see. I think it gives us an opportunity from a Cisco perspective to be even bolder within the country to really look at the AI opportunity there as well. I'll probably be in India early in the new year and just kind of planning seconds. Yeah, we plan to do a lot more around side.
There's a lot going on, yes already already around the world. I know we ran out of time good luck, good luck with your travels. I know we'll be checking in again, but so appreciate it as always. Fran Casudas. She's executive vice president, chief People, Policy and Purpose Officer at Cisco Systems. Joining us right here.
You're listening to the Bloomberg Business Week Podcast. Catch us live weekday afternoons from two to five pm. Easter Listen on Apple car Play and then brout Auto with a Bloomberg Business app, or watch us live on YouTube.
This past week, Broadcom shares jumped after the Information reported that it's working with Apple on a server chip specially designed for AI. We also got earnings from the company. Check all of that out online or on the Bloomberg terminal. Meantime, Google debut Gemini two point zero, a new version of its flagship AI model that's set to be twice as fast. Shares of Google's parent Alphabet hit a record this past week, fueled in part by that news.
Earlier this month, HPE reached a record high after reporting better than expected quarterly revenue, driven by a sixteen percent increase in AI system sales and a thirty two percent jump in server sales used to power AI work demand for high powered computing to run AI workloads has led to a boom for hardware makers, including companies like HPE, Dell, and super micro Computer.
So to help give us an idea which companies stand to benefit from all this spending is Crawford del Pratt. He is president of the Global market intelligence firm IDC, of course short for the International Data Corporation, and he caught up with Tim and Bloomberg's Jess meant.
IDC overall we forecast, wait for it, thirty seven percent growth in servers and storage this year in dollars. Okay, So what's driving that? It's an architectural shift to AI, right, and we are seeing that service providers and then ultimately enterprises will need to upgrade their infrastructure to support these new what we call AI workloads.
So especially because when you hear AI, always think of maybe companies like an Nvidia, maybe not as sexy as a name quite like this, But when it comes to you, especially sort of the IT spending side of things with HPE as well as some of its other competitors, who do you think can really benefit from this? Beyonce, when we're talking about other type of chip type semiconductor type stocks.
Yeah, So when you start looking at what's happening with AI is that you have to upgrade your overall infrastructure. So that means that companies like HPE are benefiting, companies like Dell are benefiting. You're seeing also ultimately you'll see the software stack associated with managing your infrastructure. Ultimately your applications will benefit. But that's gonna that's going to take a longer period of time. But in the near term, you know, beyond the Nvidia, you're seeing that the entire
infrastructure stack needs to get grinded. Now for let's just get back to HPE for a second. What's also happening at HPE is that a number of years ago they went on a diversification strategy. They bought a company called Uruba Networks and a Ruben Networks expanded their footprint in the enterprise. So last quarter, HPE, you know, had a tougher reaction from Wall Street because they saw the gross margins dip. They saw them dip again, but that was really related to a slow down and a stall in
the enterprise infrastructure space. In a Ruba networks, what'll we call the edge of the network, the wireless and wireless infrastructure, and as that infrastructure now is working, it's inventory. HP saw a sequential increase in that business, even though it's still down twenty percent year over year. Sequentially, that business is starting to stabilize and their gross margins will start to have started the stabilized as well, although like I said,
they were down about four hundred BIPs. Now, I think going forward, what you're seeing at the company is and you're seeing this in other companies as well, like Dell, the demand is just sustained for infrastructure and the demand for nhp's case, hpe s case, the demand for infrastructure
around the network will likely increase. And don't forget they announced that they have the air intent required Juniper Networks that will add their ability to get the service provider side of the network and they expect to close that deal sometime in early twenty five. So it's a very interesting story. Is this company is really reinventing itself and is seeing the benefit of that reinvention in their numbers.
What's it trying to leave behind?
So what it's trying to leave behind is running on that X eighty six Intel treadmill of commoditization, right, I mean, so you know, HPE SAP back a number of years ago and said, where are the largest profit pools that we can prosecute in the enterprise? Networking hit the top of the list. That's why they went after Arouba Software. That's why they brought out things like es Morale Storage. That's why they and by the way, Dell Body MC Right in the case of HPE, they brought out Electra,
which is their new storage platform. Now they're looking at at networking and saying, huh, well, there's also this service provider networking, this really high capacity, high bandwidth networking. Juniper adds that that's so what they're trying to do is they're they're basically the gravitational force of commoditization is what they're leading. They're trying to leave behind Another stock.
I wanted to point out that a lot of tension is given to is super Micro SMCI. Now, at a certain point, it was one of the worst performing stocks in the S and P five hundred this year just because of those questions surrounding its accounting practices. Though if you look here to date, it's recovered some of those losses and it's actually help around forty five percent.
You were talking about.
Kind of the exposure when it comes to service providers, thinking about like Meta Salesforce, Microsoft SMCI doesn't exactly have the same kind of broad reach that some of its competitors do when it comes to a company like this. How do you make of this and its stock performance so far?
Yeah, So when you look at super Micro, it's a fascinating company because where an HPE or a Dell or Lenovo sells a broad range of products that address the needs of many many different kinds of enterprises and also service providers, Certainly super Micro has a range of products, but really what they do is a significant amount of engineering and innovation really for what we call ultra dense
service provider applications. And so you know you're a hyperscaler, so you're a Microsoft, or you're a or you're a Meta, or you're a Google. Super Micro is going to do that really really tight engineering between the component suppliers as well as the actual physical box and bring that box out and allow you to have very dense environment. The down side of that is that they don't have the same distribution channel, They don't have the same reach that you would see at a Lenovo we're in a Dell.
So the concentration of customers is very, very high.
Oh, it's incredible. You go to this, well, the concentration of customers, the concentration of who they're buying from both Okay, yeah, yeah, because they see that on both, it's like it's pretty remarkable. I mean, they're a huge customer of Nvidia. Yeah, oh absolutely, it's like they're one of the top customers there.
So so think about the AI and actually this this is a great metamor think about the AI kind of market right now as an upside down triangle. At the top of the triangle where the service providers and that's where the majority of your demand is coming. Okay, that's the animal. It's eating all the Nvidia processors right now.
So that's the Microsoft's and the Amazon's.
Hyperscarers, the Microsofts, the Amazons, the salesforces, the de metas, and as that triangle comes down, it gets narrow or narrow because the enterprise customers haven't really stepped up and bought into generative AI in the same way within their data center.
Is that but if they do it, wouldn't they just go to a hyperscaler. Well, that's just it rather than doing it on prem potentially.
But what we see is we see a lot of customers. I'm not saying this is a customer, but a customer like a a JP Morgan or someone like that, but significant an investment in custom applications, and those custom applications they want to be able to take those data stores
that they're creating in their custom applications. Those are in their own data centers, and they will ultimately want to train their own large language models and have their custom applications be able to use those large language models to be able to get generative AI answers to serve their customers and their employees in new ways. That is when the bottom of the pyramid starts to get wider, because then enterprises start buying into this technology. But we're not
there yet. Where we are is the top is what's consuming the vast majority of these processors. And what's interesting about generative AI is that usually with these compute architectures, it becomes a law of diminishing returns, where the performance starts to tap out no matter how much capacity you throw at it. We haven't seen that yet. With gener AI. You can keep throwing processing capacity at it and the system gets more and more and more capable, and you can also get significant speed gains.
Yeah, one of my favorite functions in the terminal is s PLC so you to see the supply chains. I'm sure this is what timmy have been looking at the customerlier, but just for our radio listeners so that they know so you can get engauge there. So super Micro has been in the S and P five hundred for less than a year. It got added, Yeah, it got added to turn the end of the first quarterback in March.
So I mean, what do you make of kind of like what's next here after it's trying to get over this hurdle with the accounting issues.
Yeah again, you know, and I d see you know, we we we don't pick stocks, but I but I will tell you that from a from a performance standpoint, there is such a significant amount of eng hearing and so much as a significant amount of demand within that company. I believe that in the long term they are going to benefit continue to benefit from this AI boom, particularly as service providers continue to invest. So you know, take that,
take that with what you will. But I don't see a world where super Micro is not continuing to participate in this market.
Okay, Crawford, I'm going to go off script a little bit, sure, because I want to get your take on this. You're in Boston, that's where you're absolutely okay. You spend a lot of time in the Bay Area. Have you taken a Weimo.
I have not taken a weaimo, but I have crossed the street in front of one.
And it didn't hit you didn't. It didn't hit You're That's what I like to see. Yeah, here's why I ask. I was out in San Francisco a few weeks ago. Now, anyone can take these weaymos. You don't have to be on a list anymore. I down the media further, right, I don't. I mean it was I was only able to go within San Francisco. That's all I needed to do. I know they're training them, Yeah, they're they're training them to go to the airport. Now you can't get on
the highway with them. Yet it blew my mind. It was incredible. I was a skeptic. I'm totally sold now for this one twenty eight minute trip.
What are the.
Companies that you cover play in that space? Like, if we're working to a world where these are all going to be these are going to be all over. How do the companies that you cover benefit.
Boy, there's number there's a number of ways. And it also blew my mind. One of my team members sent me a video of her in a wamo and the thing bananas. It's driving itself. It's driving itself around a lot of ways. So companies like so think of a driverless vehicle as a little data center, a little smart connected node on the network. It has its own network inside of it, and it's obviously consuming a lot of
processing power from the network. But what's different about a self driving vehicle is the amount of processing it needs to do. We call this compute at the edge of the network because it has to decide very very rapidly. Is that a human, is that a mailbox? Is that a dog? Is that a shadow? And it has to be able. It can't compute that at the core of
the network. It's got a computer at the edge. So the kinds of companies that benefit from that is surprise in Nvidia, where you know they've been a key partner to FSD so full self driving at Tesla for quite some time. Also, over time, maybe companies. People don't think as much about Samsung, so they are going to be making a lot of the semiconductor content as well as the LED technology where you see these cars are become big curve displays over time.
I do envision a world where we see auto accidents just plummet as a result of technology such as this. Am I am I dreaming? Or is that something that happens in our lifetime?
No, I don't think you're dreaming. I think it does fall in the category of the old Paul Sappho from Stanford Free so don't confuse a clear view with a short distance.
He's a futurist.
Yeah, he's he's he's a futurist. And I think we are going to get there, but I think we're going to go through kind of an icky period. And that period is where the self driving cars don't make mistakes, but the humans do. And I think that's going to have really tough situation, a tough impact on things like insurance, and it's going to make it, you know, potentially more expensive.
And that may actually get to a point where people start to say, do I want to be the person not using self driving because it's costs a lot for me to get in this vehicle and use this vehicle, because the risk that I may be distracted by doing something, looking at a phone, having a bad day, whatever, the
car never never has that. So I think eventually we will get there, but it's going to take us probably longer than people think, because roads are complicated, and they're complicated all over the world, and all I mean self driving in Rome. You know, it gives me, It gives me,
it gives me, gives me a headache. But I think that we are very close to being able to see it on highways, being able to see it on UH in a lot of US cities and in a lot of North American cities, and I think we're probably fat.
It's probably closer than people think. I think it's you know, probably in the next in the next five years, you're probably going to start to see more and more modes, more and more companies bringing out the kind of technology the Tesla can bring out, where you know, you've got one hand on the steering wheel, but really, if you want to go pretty full sell driving, you can.
I've actually never been in AWEIMO or.
Anything like that.
You're not the only one, are only a few cities where they do this. They do it in Austin, they do it in Phoenix, there was a story about Miami. Of course San Francisco, Yeah, in San Francisco.
So when you're thinking about legislation coming out of Washington, obviously we have a new administration coming in, but kind of realistically over the course of years, because this will take a while. As you were mentioning, what do you think would need to happen, because also you were talking about the complications of roads and depending on what type of city you're in.
Yeah, so there's going to be a lot of legislation that's going to need to be addressed here, which is why I think it's going to take a long time, much longer than the five years I mentioned for full self driving. But the kind of assist where you know, what Tesla has today is really only what Tesla has where you have won a little couple fingers on the steering wheel and the car is doing the vast majority of the driving. That's going to go to multiple manufacturers.
What I was in, what I was in was full self driving. I mean I wasn't even in the Yeah.
Yeah, So what you were in is going to take a lot longer from alegislation standpoint, where we're going to have to see again, this is going to get complicated because our state's comfortable with it is the federal government going across you know, if you want to go across state lines, states have to have agreements, and if states don't have agreements, then it becomes a federal issue. And I think that could take, you know, a lot longer than the technology could take it.
It's also hard from a rules perspective, and I mean rules in the sense of like rules of the road. For example, I had to drive recently and I was driving my car in New York City and there was a car double parked in front of me, so I couldn't get around it. In order to get around it, I had to actually go over a double yellow line. There were no cars coming right. But what I did was not considered following the rules in the sense of like if a computer, like I don't know if a
computer would actually do that. If I were in a self driving car, would I have just sat.
Would you have just sat there and been frustrated because you it's an obvious solution that's not going to harm any right, and you know, it's the kind of thing you could do potentially. And now those kinds of things, you know, can be those kinds of canations can be programmed in, but probably not for a long time. I'll give you another example. What about the example where people start to figure out where the vulnerabilities are in these vehicles.
So if you want to see that, you know, I'm making this up, but if you want to see the vehicle behave in maybe a strange way, come up very close to the back corner of it, or you know, you need to figure out sort of bad actors, not necessarily bad actors, but just people that try to figure out where the vulnerabilities are so that they can get around them faster potentially, or you know, potentially you know, try to gain the system in some way when you're
a driver and you're and you're driving sort of against if you will, a full sell driving vehicle.
Looks like Alphabet's robotaxi service is going to launch in Miami in twenty twenty six, So maybe that's more.
Of the cropping up.
I'm just waiting for New York.
That's right.
How how complicated is that?
Like will never see that? I mean, you start.
Getting outo some of those boroughs. You start getting downtown, it starts to get a little bit complicated. But I would say I don't think New York is meaningfully more complicated. San Francisco can be pretty complicated as well.
Yeah, I'm talking to you a friend who works Intact and he was saying, New York's tough because there aren't very many legal places to pull over where San Francisco there are a lot of driveways that people can pull into to get into one of these things. But yeah, it could be complicated here. See Crawford, always love it when you stop by. Thanks for sticking us.
For you both. Thanks so much for the time.
Crawford Del Pratt, President at IDC.
You're listening to the Bloomberg Business Week podcast. Listen live each weekday starting at two pm Eastern on applecar Play and 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.
The rise in artificial intelligence is turbo charging demand for bigger data centers. This is transforming the land that they're being built on and taxing a region's energy grids. In April, Goldman Sax estimated that data centers in the United States are projected to use eight percent of total power by twenty thirty, up from three percent in twenty twenty two.
These data centers have the capacity to consume it combined five hundred and eight tarawatt hours of electricity per year if they were to run constantly. Okay, so what does that mean. That's greater than the total annual electricity production in for Italy or Australia. Needless to say, data centers, well, yeah, they're power intensive.
That's some great context and perspective. Meantime, developers are meeting resistance at every turn from locals and property owners over how to accommodate the unprecedented power needs in their communities. One example is QTS. It's the data center developer that asset management giant Blackstone bankrolls. They are building QTS ten bunkers over six hundred acres of land in Georgia that'll house thousands of computers for companies such as Microsoft.
The complex is expected to consume as much electricity is about a million US households, leaving utility Georgia power rushing to build the infrastructure to meet demand. Some of that infrastructure falls on the land of homeowners upset with high voltage wires on their front lawns. Writing all about it is Josh Saul, Bloomberg News Energy and climate disasters reporter who joined Katie Griefeld in me.
What my colleague Dawn and I found was not that people were so upset about the data center itself. I mean, it's enormous. It's almost as big as Central Park, big as Disneyland. But what people are really angry about is the power lines necessary to electrify this data center, to run all those servers, to run the cooling, to run everything. Those are cutting through people's lawns. Basically, people feel like cutting through their lives. This is a quiet wooded area.
People love watching the birds come in and rest on the lakes and etc. So people are really upset that the power company has to build these lines in order to power the data center. And this is something we're seeing across the country.
And this is a simple question, but I assume that's a non negotiable that to power the data centers, there needs to be these big, ugly things. They can't put them underground or something I was going to ask the same thing. Okay, good run the same wavelength.
Without the transmission lines, it's just a big empty Walmart. You can't do anything without the power. Putting them underground is really expensive. That's basically it turned out to be a non starter. Some in the community, including the local church, asked, hey, can we bury these but it would raise the costs and created some other issues beyond what the power company and QTS were willing to deal with.
I have to imagine it would be disruptive too, if you have to, you know, dig it up.
Yeah, they do it here in New York City. That's a lot of the power lines and most neighborhoods are aried. Yeah, but yeah, what happens in New York City is not necessarily what happens around the country.
Right, So it's as much power as a million homes just for this one one data center.
That's one data center.
That's one huge data center. Yeah, one point four gigawatts for people watching this area. You want to get start to get used to the terms of megawats and gigawats. Okay, what's a giga gigawats a thousand megawa? What's a megaway megawatt is one megawhy let's let's move on here. Okay, a huge amount of power.
And what that.
Creates is issues with the local power generation. So you know the power company in the state Southern they have goals to shut off coal plants, to build fewer gas plants. Those are things that you do to get your carbon emissions down and hopefully have less climate change. But when you have when you keep adding more and more demand to the grid, you have to add more supply to the grid. And as one utility exec put it to me, that's just basic physics.
Where does Blackstone come in here.
Blackstone owns the data center developer QTS. They bought them a couple of years ago and QTS is now a major part of their major part of their portfolio. QTS builds data centers across the country. They have a footprint fifty billion dollars worth of data centers and they have plans for another fifty billion dollars of data centers in their pipeline.
That's really interesting, and I mean, Blackstone isn't alone in this, but Blackstone is going particularly big.
It sounds like that's right.
Blackstone is a big player in this and along with all the big tech firms which are also building their own data centers to run their own operations.
I mean, you think about this fight that's happening in Fayetteville, Georgia. It doesn't feel like this is going away, that this problem is only going to get bigger. You think about how Wall Street in particular is going all in on AI playing out in Georgia right now. But I have to imagine that something similar is happening across the US exactly.
It feels like, it feels to me like all of our systems are set up to build more and more of these things and to power them. You have the tech firms that are in this AI race and want the data centers because they need that compute power. You have private developers like QTS that are incentivized to build lots of these things. And then you have power companies which their mandate these are regulated utilities, but their mandate
is to provide power to customers in their area. So when a big new power demand shows up, their job is to provide power to that and build the power lines to it with a threat of eminent domain behind them. So it kind of feels like just everything is set up to get as many of these onto the ground as possible.
What do we know about what happens when a high power line like this goes on someone's property or goes through their property, Like, does it change the value of a home? Does it make noise constantly? Are you hearing buzzing those sorts of things?
It doesn't make noise. It certainly affects people's appreciation or love for their own home. You know, we talked to lots of people who are like, you know, I moved to I moved to Georgia because I wanted to take care of my mom, and we bought this place. It was going to be kind of our It felt like our little paradise, and now it feels like that paradise is crumbling. So people certainly don't like looking at power lines and then more in you know, more brass tacks here.
Appraisers we talked to said that big transmission lines can cut the value of a home by ten to thirty percent, depending on how big the line is, how close it is the house, how expensive the house is, or and the rural character is to the home and to the property. So to have you know, these are some of these are nicer homes. They're in an area with where that rural character is important. Some of them could be seeing big drops in their value.
To approach this like an economist, though, I mean you have to imagine that more people are going to move to some of these towns, that that's going to bring in more business, there will be more spending in the local economy. I mean, are there any benefits here.
It's not so much people because these data centers don't actually have many employees. You have some construction jobs on the front end, but you have sort of the stereotypical guy with the silver ponytail when by himself when the data center actually opens. But in terms of benefits, big tax benefits, a lot of you know, QTS is going to pay a lot, you know, different number, you know,
a huge, huge amount of money over time. Already the school local school district has separately received enough money and tax benefits to you know, retain seven teachers, half dozen teachers. So certainly, I mean it's economic development. That's why these local boards of commerce are saying like, please come here, and they sort of like actually sought out and wanted QTS to come there.
You've been You've been writing about and reporting on this space for quite a while at this point. Every time I see like a new big story from you, it has to do with AI and data centers. What is it about this area that makes it appealing to this company QTS? And what are the other areas that that people should think about when thinking, Okay, well is my community next for one of.
These Yeah, this area is attractive because they have a big stable power company. Georgia Power is owned by Southern and that's a you know, a great company for these developers to work with. And the developers were able to find a plot of land and talk to the power company about having having electricity provided. As far as where else they're coming, I mean it's almost like where else
are they not coming? You know, you have Northern Virginia's Data Center Alley, massive amounts of data center development there. There's also you know, Arizona, but they're really coming up in just large parts lots and lots of the US and also around the world.
I would think of like, you know, relatively cheap electricity, but that's tough to kind of come by in a stable way, right.
We actually heard more about We actually heard more about cheap electricity. When I was writing about bitcoin miners and their search for power, the hyperscalers have so much money and the race for AI is so big that they actually are more concerned about stable and large amounts of power, with the price being secondary.
That's interesting. One thing that I think about with a story like this, Josh, is how fast innovation happens, and the way that things change so quickly when computers become more powerful. And I wonder if they're overbuilding when it comes to this stuff, given that they're able to make these GPUs more and more powerful with every iteration, like do they need data centers that suck this much power and that are this big?
One complaint I hear is that the is not some kind of you know, more central or more organized planning, you know. So it's not that, you know, data centers aren't being built in one specific area for one specific purpose. But certainly the tech companies are pretty sure that they need a lot of these are building them rapidly for that reason.
It's today's big Take. It's one of the most read stories on the Bloomberg terminal. Josh Saul is at Bloomberg News Energy and Climate disasters reporter. He wrote the story along with Don Lim who covers private equity here at Bloomberg. Josh joining us here in the Bloomberg Interactive Brokers Studio. Check out Josh's story and more from Bloomberg The Big Take on the Bloomberg Terminal, and at Bloomberg dot com Slash Big Take.
You're listening to the Bloomberg Business Week podcast. Catch us live weekday afternoons from two to five pm Eastern. Listen on Applecarplay and then brout Auto with a Bloomberg Business act or watch us live on YouTube.
Plenty Ahead in our second hour of the weekend edition of Bloomberg Business Week, as our focus on artificial intelligence continues, including a conversation with the former US Navy fighter pilot who built Amazon's global air network. She joins us to talk about the new technologies Amazon is rolling out this holiday season, also the role of AI, and yes, what it was like flying fighter jets.
First up this hour, we begin with the one struggling upstart that's become the three trillion dollar king of AI. It's the world's second most valuable company at least currently it's like recently been neck and neck with Apple. Apple recently pulled ahead. But look in videos right there it is they just go back and forth, right, yeah, look safe to say no name is more associated with the aieuphoria over the last couple of years than video.
Absolutely, and video shares, by the way, are up about nine hundred percent since the start of twenty twenty three. It's come a long way from when it was founded out of Denny's in Silicon Valley more than thirty years ago.
Writing about in Video's journey is Take Kim, senior tech writer for Barons, a former columnist for Bloomberg Opinion, and he's the author of the new book The Nvidia Way, Jensen Wang and the Making of a Tech Giant.
It all started with three D graphics, and then program Mobile GPUs, and then Kuda, and now with the tensor cores and AI chips and Jensen Huang, who's the co founder and CEO, every single major computer shift in computer chips and three D graphics, he's learned to position in video to take advantage of it. I compare him in
the book to read Hastings with Netflix. Hastings had this intuitive sense as someday we're just going to watch video over the Internet, but the technology wasn't ready yet, so he started the DVB by rental business and just stayed around and invested in the technologies. And when the internet infrastructure was broad band got good enough for consumers and the video streaming technology was there, he jumped on it
and was the first mover and dominated that market. Jensen and Vidia did that like with three four different markets. Like every single time he positions the company, he doesn't know exactly when it's going to take off, like maybe it's five years from now, maybe seven years from now, but he positions a company in this tremendous amount of capital and R and D, and this the software ecosystem is actually huge competitive advantage that he invested in.
Well, let's talk about him as a leader, him as a visionary. What did you you spend some time with him during the writing of this book, you interviewed him earlier this year. What would you say it is about his personality, his leadership style, the way he conducts himself that has created this culture ad in video that has allowed it allowed it to not just succeed but dominate.
I would say there are three pillars, and the n video way is basically the Jensen a way, because I mean he's there from the beginning and create this.
Remember there's a Netflix way too.
Oh well, but this is like, this is Jenson's way. I would say three pillars are insane work ethic, talent, cultivation, and speed and velocity. So I'll just talk through each three work ethic. No one works harder than Jensen. I mean he is reading one hundred emails from his employees every day. He works all week. Maybe Elon Musk is up there. Sorry, sorry, but I'd even argue compared to
you On, maybe Jensen works more. Like he says he can't when he goes to the movie theater, he can't remember the movie because he's constantly thinking about work all day. He like lives in breezes work and work is relaxing. Like he drinks a Highland Parks Scotch Lane at night and Sunday and he's going through emails and like peppering his employees do this, do that, and that's his life. So even at the beginning, like I have these stories when I talk to these early employees, like they're working
past midnight, waking up at four or five in the morning. Yeah, and this is a time when Intel was about to crush them, and I had this one employee telling me, well, Jensen just inspired me to work even harder. So even though like I should go back to bed, and I really want to go back to bed, I'm going to get to work early and work even harder. So work ethic. No one, No one's going to be Gensen tumtrt worth. He just he's a work ofhol and will outwork anywhere
in terms of talent cultivation. He knows that the only way you win in this in technology is hiring the best people. So he's constantly recruiting, you know, he's every time he interviews or talks to someone, he goes, who have you worked with in the past that were rock stars? And he pockets that knowledge and then three years later he's like recruiting that person, constantly badgering them and sometimes they say no. In the beginning, it's like no, no, no,
I don't want to leave my cozy company. But he usually gets the people and then they stay for twenty five thirty years. And the reason is partly in video wins and everyone wants to stay on the winning team. But second it's just there's this great culture inside video that even though there's thirty thousand people. I write a
lot about this in the book. I think large parts of corporate America and technology industry suffer from dysfunction, from internal politics where you're spending half your day trying to please your boss's instead of doing what's right for the customer and for the company itself. And video costs this missions the boss. Make sure you're making decisions where you're fulfilling the company's mission, not you know, all the Game
of Thrones type internal in fighting. The other big part is when I talk to the video boys that leave in video, they can't adjust to the new companies when they go to Google, Microsoft, Apple, because because there is so much endless meetings, internal paralysis. Like one person says, it's like swimming through quicksand and I think they.
Should just retire and count their money.
Yeah.
Yeah, but a lot of it didn't just need to talk about we talk about company culture. We work with a company that's got a very distinct culture as well. And but it makes me think and Video wasn't just oh, we were at the right place at the right time. It sounds like has been constantly checking where does this market go? Because if you think about the timing of a like just how it all happened. This wasn't just accidental. Oh I got it.
Right, No, not at all. I mean he was investing in Kuda, which is the reason why in Vidia is kind of doing really well in AI, the ecosystem Kuda. He invested that in two thousand and six, and then when deep learning took off in twenty thirteen, Yeah, he was like, we need to invest in AI software libraries, and we need to go to campuses and expread the religion of Kuda.
Explain what Kuda is.
Kuda is a programming platform for parallel computing, which is
what Nvidia GPUs do basically. In videas, they call it accelerate computing, but it's basically just parallel computing where each GPU has thousands of tiny little processors, and most computer PCs run on CPUs which have four to eight processors normally, So in Vidia's kind of computing architecture is based on this parallel you know, tiny processor mentality, and Jensen from the beginning, even in nineteen ninety three, saw this as the future of computing, like you're going to break apart
all the workloads into tiny little processes and they're going to work on GPUs and he saw that from the beginning that this is the way the world of computing is going to go. So when AI took off.
In the video was sixteen years.
So that's amazing, Like, how.
Does you kind of keep going keep going because he.
Made like I said about Netflix, he knows where the future is going to go. Yeah, and he will keep in that.
I mean.
The amazing thing about Kuda is all these I talked about. I have a chapter in the book about Starboard Value, which is a very famous activist hed. Yeah, they were board members and Video were worried when Starboard took a big position that they were going to, you know, clamp down on Kuda because Wall Street was all over Jensen for investing so much in Kudah. It was taking die space on the chip, it was crushing their gross margins for years, and he kept doing it because he was
convicted this would be the future. And you know, luckily for Nvidia, Starboard didn't go activists. They they kind of pressured Nvidia. They did, and Video did a big stock buy back. The stock went up for twenty percent and the very peared and star wars like, I'll take that gain. Then Jess Smith told me that he was like, hmm, I should have never sold in video He met with Jensen. He was impressed by Jensen.
Seriously.
Yeah, he took the short term gain and you know that's history. But yeah, Jensen does that all the time. I mean the two key technologies for video game GPUs that the reason why gamers spend up to sixteen dollars to play video games on computer is DLSS, which is this AI upscaling technology, and raytracing, which helps the GPUs do do lighting better. Like it's amazing. He invested in ray tracing for ten years before it was commercialized, Like ten years. Most companies will invest it for two three
years and then give up. The LSS took six years, right, So this is this is a company where he tells me like he's paranoid about getting disrupted by small he's a student of history and all that stuff where coming for him. Yes, every company is coming for him, and he knows that they're going to come up from underneath.
And companies get complacent, and that's the thing. He is paranoi about being complacent, so he will he knows that if one thing he said, if you don't invest your debt in technology, he.
Knows that that's This is capitalism. This is like the way that efficient markets work. If something is incredibly successful, you're going to have other players try to enter the market. Who's coming after in Vidia? How big of a barrier or a moat do they have?
So most Wall Street analysts investors are worried about the big hyperscalers. So the Microsoft Azures, the Google Clouds, Amazon, so all of them are making their own versions of aichips.
Yeah, we had an Amazon story just a couple of weeks ago about the effort in Austin, Texas, right that they're doing there to try to take on in video.
And empirically, it makes sense. These are huge companies that can invest billions of dollars in R and D. But I'm not worried about the hyperscalers at all. The problem is why, because Google has amazing technology for TPUs, but no one uses it because everyone's afraid of getting locked in and the performance is as good. So when I talk to startup AI startups, they're like, I'm not going to lock myself into Google. I mean, Google has your reputation of like building stuff and then giving up and
moving on. I mean we've seen that time at Google. Studia is a great example with the cloud video game streaming, and no one wants to get locked into one vendor. And with Nvidia you get the best performance. So also the most important thing is Nvidia with Kuda has been optimized for over fifteen years, so if you ever have a problem, you could look up and say, oh, this is how to fix it. And they ironed out all the bugs. And if you're an AI startup, you have
one hundred problems. You're working on one. CEO told me, I don't want another problem where if you port your software to a different AI chip platform you always have problems. It's like going from XA six to power PC back in the day. So you don't want that problem. And then the other their problem is one of the biggest
competitive vantages is they accelerate product cycles. So back in the nineteen nineties, the way they drove all their comparison at business was the typical video game chick took eighteen months to design and ship, and they're like, forget that, We're gonna do every six months. And because they're so the work ethic and relentless execution. It drove all the comparison at business. So last October it took them two years to make a major GPU chip platform, and they're like,
forget that, We're gonna do that every year. So why would you switch to a different platform to save a little money on the hardware side when Jensen is gonna come out with a better AIGPU platform every twelve months. That's thirty times fast. Hard to compete with that, Yeah, you can't. And the technical risk is a huge thing because if you poort your stuff over it might not work, there might be bugs, and you don't want to risk that.
Like time to market if you're an AI startup right now is everything, and the fastest time to market is by using video.
And yet, as you're right, Jensen's always looking ahead, what's next? He sounds like he's constantly churning and looking at where, whether it's five years from now, ten years from now. As you talk about the things that he was investing in for a really long time, including these AI chips, so where is he looking next?
So I've listened to hundreds of hours of the speeches the last few years. The Internet's an amazing thing. I've talked to him directly, and I know the two most important things he's most excited about. Number one is digital biology. This area is just perfect for the AI computation power that's just going going up exponentially, so making drugs exactly. It's drug discovery, molecule.
Similar medicine, stuff like that.
All that stuff. And I've talked to a drug startup, a drug startup. I talked to the head of Bristol Meyers squib Digital and they're seeing vast productivity gains by using Nvidia AI service systems. And it makes sense. All these companies, especially to big companies, have decades of clinical data that they could put into these AI GPU systems.
And the thing that really is really critical that's leading to innovation is that you're getting these AIGPU clusters that are scaling from sixteen thousand GPUs and now to one hundred thousands, and now they're building one million. You know
that it's just a matter of time. So this vast amount of computing power can access and digest all that data, figure out all the commutations, and simulate all the biological interactions, and you could figure out drugs that could cure disease, be very specific.
It does seem like in Nvidia is inextricably bound to Jensen and Jensen too and video. What's the succession plan?
I cust First of all, he doesn't I him giving up in video anytime soon. This is his life, this is this is what he enjoys. So I'm complete, one hundred percent convicted that he's not planning to lead anytime soon. I mean, this is the love of his life, he says, I love the video all the time.
Everybody gets older, everybody gets older.
But how much of the company is him? Like we talk about this when you've got an individual so identified with a company that, if God forbid, you know, are there individuals I mean, after Tim Cook they were able to figure out their way forward after Steve Jobs. I mean, I'm just forgiving I'm sorry, sorry Tim after Steve Jobs.
But you know what I'm saying, Yeah, I mean, I think that's the critical question. He talks about how he kind of configured and structured in video to be a race car that he could steer perfectly. It's like perfectly matches his way of managing the company. I think that's the big that's a question for every big company right like this, this the assessing CEO. Who it is matters. I mean, you get a John Scully who almost drove people a business, or do you get a Satie Nadella
who just took Microsoft to new heights. I mean that's a question for every company.
One question we really want to know. Did you get to see the closet where all the leather jackets are?
No?
Come on, but you know it's such a part of him, right, yes, like yeah, no questions or anything came up.
I did not focus on the leather jackets. I probably should have. I did not ask about that.
But Part two.
Future column Future columns, great dive, deep dive about a company that we talk about I feel like on a daily basis and really has kind of changed certainly the investment narrative.
In a big way.
Take him senior technology writer for Barons, a former columnist for Bloomberg Opinion. The new book It is out today, The Video Way Jensen Wong and Making of a Tech Giant.
You're listening to the Bloomberg Business Week podcast. Listen live each weekday. He's starting a two PM's You're don Applecar Play and Android Auto with the Bloomberg Business Ad. You can also listen live on Amazon Alexa from our flagship New York station, Just Say Alexa playing Bloomberg eleven thirty.
This year, Amazon unveiled its next generation of fulfillment centers, ones that will be powered by AI and will have ten times more robotics. The company known for its automated warehouses and use of technology, has seven hundred and fifty thousand industrial mobile robots in use today.
It's not just robots. Amazon now using Jenerative AI to develop models for demand forecasting, and it's starting to roll out their AI powered vision assisted package retrieval to help delivery workers.
Sarah Roose is involved in all of this. She's vice president of Global Operations at Amazon and also a former US Navy Fighter jet pilot. And yeah, Tim and I wanted to start there with her.
Well, I've been with Amazon approaching fourteen years. Started out as an operations manager in Lexington, Kentucky in our fulfillment operations. Spent time in Kentucky, Texas, spent three years in the UK, which was an amazing experience. We turned to the US in twenty seventeen and led a team that built out our hair haul network. So we have a fleet of Boeing seven sixty seven seven three sevens.
You really do have your own airlines, well kind of.
That's something like that, and a lot of them in Kentucky around We have our main air hub at the Cincinnati Northern Kentucky Airport. They process over thirty flights a day, hundreds of thousands of packages daily.
I didn't hear any stories about fighter jets.
I'll get you, I'll get you after sure.
Yeah.
So spent six years building out the air haul network in the US, Europe, and India, and then a year and a half ago I took on the leadership of our global workplace Health and Safety team. So we have a workforce today of over in laian employees globally across six continents. Prior to Amazon, yes, I was a Navy F eighteen Super Hornet fighter pilot. So if you've seen the movie Top Gun, Maverick the jet and the main part of that movie, that's that's what I had the
privilege of flying. Definitely a childhood dream.
Do you still fly now?
Not currently. I do have a commercial pilot's license, but Amazon can seemed busy on a day to day basis.
What do you find a lot of your colleagues from the Armed services are doing now who were pilots.
Then, Many fly for the airlines, Many went into business. I have several fellow aviators that are in business now as well, and many have joined companies like Amazon. In fact, Amazon's very veteran friendly. We've hired over forty thousand veterans over the last several years. So it tracks people that are accustomed to working with large teams of people and familiar with operational type of environments.
Well, I was going to ask you, Sarah, like, what about that experience that you just has, I don't know, helped you tremendously in the job that you do at Amazon, And you've done multiple jobs and made your way through, So give us an idea of like how that carried over.
Yeah, I actually get that question quite a bit, And I know it's great because it is something that people are curious about because it doesn't necessarily seem like a common path.
But I think about that Fighter jet like, yes, this coordination, but a lot is what you do on your own right.
It is a lot of decision making, for sure, but it also takes a team.
You know.
Jet pilot is merely an extension of the team that helps ensure that jet is up and ready to go, the folks on the flight deck of an aircraft carrier, the folks that ensure that aircraft carrier runs safely.
You know.
So it is definitely a team effort, and so taking a lot of that teamwork in that team environment into the operational workspace is one thing.
You know.
I think decision making a bias for action, being proactive about things, looking ahead. These are some of the traits I think that translate pretty well. And yeah, I think that it's a combination of thinking big and also diving deep. Over any given day, we are going to move on.
I promise, Yeah, it's okay.
We have plenty of times.
Well, we just your backgrounds are like pretty fascinating tense moments that you.
Had flying, yes, landing on a moving ship at night in bad weather, and you know, landing at one hundred and thirty five knots and going to zero in less than two seconds. And you know, we call it trapping aboard an aircraft carrier. That that's just a lot thought, right, yes, yes, a carrier, yes, And so it takes a lot of training and concentration and precision to do that. Well and do that safely, so SOPs to you.
They're probably like, yeah, I got this.
Well it does help to put things into context.
Yeah, exactly.
All right, a little more on flying.
I don't know, no, I don't know, no, no, no, sorry, were I can talk about flying all day?
So good?
Yeah, but you know it is. But it is actually interesting though, what what you've done at Amazon in terms of building out that logistics network. It like used to be a unique thing to sort of, you know, get to an airport and see a prime air yeahane, But what a little over a decade ago the team over at Amazon started building out its own fleet. I think you worked with Atlas Air Is that right?
That was one of our carriers.
Yes, And now what's the relationship now is it? Does Amazon own its own planes? Does it contract out its planes? How does that work?
Well, we have a combination of we own some aircraft and we lease out other aircraft that they're all operated on our behalf by air carriers. So the pilots aren't actually Amazon employees, but they fly our branded aircraft. But going back to your point about seeing a primary aircraft is kind of a unique thing. I think back to
when I joined Amazon. At twenty eleven, our operations basically ended at the dock door of our fulfillment centers, and so over time, especially in twenty thirteen, we recognized in order to keep up with our own growth, we really had to start build building out our mile last mile transportation network. So it wasn't long after twenty thirteen that you started seeing branded trailers out on our interstates. You started seeing branded final mile delivery vehicles out on our roads.
We built out sorts centers, delivery stations, and you started seeing branded aircraft out there at various airports throughout the country, including JFK here as well, not too far from US.
So does that just continue in terms of the build out, Like, I'm curious, what's the next iteration or next stage of that.
Well, we worked to also make for an efficient operation as well. So I think about something that we as a network have done differently this year as we had versus years past, to regionalize our network. And I'm talking holistically, not specific to air but if I think about if customers here in New York City and places in order, over seventy six percent of customer orders are fulfilled within region, which means within a couple hundred miles of each other.
And what that allows for is it means fewer touches, which means less on transportation costs. We work to keep those prizes low for our customers. And how we've done that is basically, we're just more sophisticated with the placement of our most common and most popular products. So we're a bit smarter in how we place that that product.
Right Like, if I'm ordering diapers on Amazon, which I do all the time, it turns out that they're being shipped oftentimes from Staten Island rather than from a few years ago, they would actually be showed.
They could have been Kentucky, they could have been sent from anywhere Pennsylvania. We're working hard to make sure that we're getting them closer to our customers.
So building out that kind of localized distribution network makes complete sense in terms of time, I'm assuming money saved and access, right Like, it just makes sense.
Yes, Like the shorter transportation distances, so that reduces the cost or reduces the number of touches that need to be placed on the package to process. And of course we still have some unique items, so not every item is located near every customer. So that's why we still have an air air network to ensure that two day delivery, especially if there's something very unique, maybe located in California, and we want to get it to you within two days here in New York.
I know we've ten, we're gonna have another chunk of time. Are people opening up and saying, yeah, I'm okay with a two day delivery. It's not like I have to have it this afternoon anymore. Are you seeing increasingly people do that? Like I know in my household, we now are maybe starting a bucket and maybe letting the week go or even a couple of weeks before we kind of hit the buy button and so on and so forth. It's not like everybody's like, I gotta have this now. Everybody ten multiple.
Order unless unless which happens, which.
You gotta do. Are you running around the corner and pick them by? But are you thirty seconds and then we'll come back and talk more. Are you increasingly seeing people comfortable like waiting?
It is such a personal preference, And that's why when a customer checks out, they have multiple options. You've probably noticed that one. Yeah, so if you would like it next day and that's an option, get it next day. Sometimes you'll see order within the next two hours and you'll get up by four pm this afternoon. Sometimes it might be a two day delivery, or for folks that do want to consolidate their orders, they can specify the day delivery and we work to consolidate into one shipment.
I want to talk a little bit about the next stage of innovation and how you're thinking about delivery, because you talked about just this incredible build up that's happened over the past twenty years at Amazon. When it comes to aviation, when it comes to last mile, when it comes to your own brand. I mean, we see we have Rivians all over our neighborhoods delivering Amazon packages in addition to the Dodge rams. But it seems like more Rivians. Now,
what's the next area of innovation? What makes it more efficient?
Oh? Well, actually, speaking of Rivian, we just announced last week we have over twenty thousand electric Ribvian vehicles on the road. So when we think about sustainability and the efforts that we're leaning into there, we're continued to invest and lean in on the electric vehicle technology. So if I think about innovation to answer your question specifically, I think about the innovation that we're doing on behalf of our customers and to also help our customers with their
shopping experience. So we have a new AI power tool called Rufus on our storefront and so people can use it to maybe get more information about a certain product, or maybe they need some help with some sizing, or maybe like me, I need some gift ideas for my nephew. Whatever the case may be, they can use that tool to help them out. And so that's really exciting to lean into that and there's more to come in that space. And we're also leaning in on AI within our operations
as well. So we have a tool, for example, that our employees can use called Dragonfly, and from the palm of our employees' hands, they can enter any kind of safety suggestion, operational suggestion and it goes directly to their manager for action. And that's just one of many tools in the toolkit that our employees can use.
Are they using it?
Oh?
Yes, And it's awesome and the amount of feedback that we get and the ideas that are generated. And the really cool thing is one suggestion from an employee may not just be for that particular site, but it could impact operations across the network. So we're seeing some really great ideas coming from our frontline workforce, which is which is awesome, and it's a great tool that is one of many, Like I mentioned the toolkit for us to maintain that direct dialogue with our teams.
Where else do you think of AI? Like, I do feel like we're scratching the surface and just beginning to play around with it, either on a corporate or personal level. Right, But it's it's just I think we're all trying to figure out how do you think about? Okay, this is how AI is going to change. I mean, you guys are such a massive world globally, Like, how do you think about all?
Right?
I see AI doing this for us at Amazon someday.
Well, we have a bit of a glimpse into the future of AI automation and robotics with the site that we just launched in Streepoint Shreeport, Louisiana. We call it Shive one for short, and I was just down there a couple of weeks ago, and it is the most automated site we have anywhere in the world. We've incorporated robotics AI automation. We as Amazon have been using AI and machine learning for several years actually, but it's now
really kind of coming all together. And so when I think about the site in shreport that we launched, we now have autonomous robotic drives that move carts full of packages from one point to another.
How's that different from a Kiva robot, like, because you guys have been doing those.
Yeah, the Kiva robots, the traditional robotic drives that we use, they run off of basically a grid. So think of they are basically are guided by technology that uses what's called fiducials. So think of it like a two D barcode on the floor and it's going from one point to another. Whereas Proteus is automous, it doesn't rely on a grid. It knows it needs to go from one station all the way to a doctor without any type of grid. But there's a lot of safety features involved,
so it goes yes. So when it senses a person, it'll make some noise. It has a looks like a smiley face, and so it can looks like it can kind of see the person and it smiles and helps ensure that you know, it stops and yields to people, and we actually use similar technology on our traditional robotics floors are texts that help to maintain the robotic drives. They actually have a vest that they put on when they go out onto the robotic floor and it stops
all robotic drives within six feet. So it's pretty cool watching that because it looks like the person has a superpower stopping those robots, which is which is a huge safety a great safety feature that is definitely you know, necessary to help ensure that we're keeping that area safe.
What about when it comes to robotics and automation in the picking process, where you still have people actually you know, going and fulfilling those orders putting them in boxes.
Right, So we've incorporated robotic arms into some of our processes as well. So if we think of some of the repetitive motion or working to engineer that out, because the repetitive motion can lead to ergonomic injuries, which is one of it is the most common injury type across the industry, So not just at Amazon, but think of warehousing and logistics in general. So if we can engineer
that out, we're doing that. However, the site and shreetport, we still employed twenty five hundred employees, but thirty percent of those jobs are more technical than the jobs that we've had traditionally, Meaning what so to think of You know, we need folks to maintain our robotics and automation, so it takes some technical certifications in order to do that, and our teams can gain those certifications through a program called Career Choice at Amazon and so at no cost
our employees. So it's a it's a great benefit that's available.
The question that has to be asked when if we're talking about AI and robotics kind of merging together. We were out on the West Coast at an investment event and I was involved in a panel that was about AI, and the question that everybody says was a, I going to take my job, So help me understand. There's got to be efficiencies when you're dealing with robotics and so on and so forth. But what is the thinking about will it take some jobs? Will it create some jobs? Give us your insight on that.
We call it a collaborative environment, and there's a balance for sure. But again, when I go back to when I started in twenty eleven, we didn't have robotic drives, or robots in general across our network, and we had fifty thousand employees total across the globe. Fast forward to twenty twenty four and we have over a million employees across the globe, and we've incorporated robotics in our sites across the US and in other countries including Europe. I
was just in Japan a few weeks ago. We used robotics in Japan were I look at this holiday season, we're hiring two hundred and fifty thousand employees.
So we have said the same as last year.
It's on park with what we hire to as we go into the holiday season, and we continue that throughout the year, and so there are plenty of jobs available to include more technical jobs for folks that are interested in that more technical part of the work experience.
Do you think given technological upgrades that we've seen in recent years, we've sort of peaked when it comes to humans in fulfillment center jobs. Like you know, if we think about it from an efficiency perspective, you don't need as many people now, and even next year you'll need fewer and fewer people. Not that they's taking their jobs by any means, but efficiency.
When I think of efficiencies. I actually think more of our logistics network as a whole. So we talked earlier about regionalization. That's really where we're seeing the efficiencies of removing additional touches. That doesn't mean removing people, it just removes some of the inefficiencies that we have in our network. And we're continued to still grow. We're launching buildings new operations just about every month, multiple per month in some cases.
In fact, this year, we just launched our first Amazon operation in Hawaii and it complements the air operation that we have there as well, So we have an Amazon delivery station outside of Honolu, which is pretty exciting.
Any update on drones, It's been more than ten years since Jeff Bezos was on sixty Minutes Sunday after Thanksgiving Monday before Cyber Monday. I remember covering that it was great marketing, by the way, but still I haven't never had anything delivered with a drone.
Yeah, we're continuing to invest in that program, and I am excited to say that we are actually operating drones for customer fulfillment outside of Phoenix, Tallis in Arizona.
To be is it scalable?
I think over time, it will be you know, there's obviously a lot of regulations.
So TAM's going to get his diapers in Brooklyn High a drone from Amazon.
It may take some time, but I'm optimistic.
Actually, gone are my daughter's diapers?
You were going to go there?
I take it back.
I take it back at your daughter's diapers now, But I do wonder, like Ken, is it scalable in a city that's so dense.
Yeah, we're continuing to iterate on the drone design itself. It's all design in house. We're working with the regulators, of course, because drone delivery in general is a relatively new space and so there's been a lot of innovation, invention and iteration over the last several years. But it's exciting and I'm excited about it.
Okay, Cyber Monday, we promised, we promised. How are things looking so far?
They're they're looking strong. We had a record breaking Black Friday. Of course, Cyber Monday isn't over yet. In fact, deals are still dropping. But I was just at a fulfillment center in New Jersey this morning and despite the amount of volume, they're processing hundreds of thousands packages at one site. It's very calm, home organized. We spend the entire year preparing for this this season, and it's people. It's really fun for the teams as well. I personally have experienced
our holiday season in fulfillment centers. So it's been a great day and overall a great weekend for the for the teams and more so for our customs.
When is it the most stressfulst two weeks before one week before Christmas? When does it just get? Like when I get those start getting all those emails, like if you want to buy by now or you want to buy now, Like, when does it get Is it the toughest point for you?
I want to I want to equate it with stressful, because it's always a controlled operation, in the safe operation. I think it's really more so if if folks are like me who have a tendency to procrastinate to the last minute, it's just hard to purchase maybe those those last minute gifts. You know, there does come a point where there's a cutoff period for the Christmas delivery, So it's really just keeping an eye on those those cut off periods.
I got to ask you, Sarah, I think Tim and I both think about this. You because your VP of Level Operations and you've talked about kind of just everything around the world, how would you describe the global economy right now?
Well, what I what I can speak to and what I can say is what we are seeing in terms of trends is that our customers, and I would use that broadly, are looking to save money, they're looking for a bast selection, and they're looking for fast deliveries. And so we see that across the US, across the globe, and because we focus on our customer, that's what we continue to do ensure that the best experience for our Amazon customers.
Do you also handle grocery delivery and whole foods and stuff?
There's grocery whole food. It's separate, but we have visibility.
On it because that's a completely different animal.
It is, and that's growing space for sure.
Interesting. Yeah, very cool.
It's your favorite plane to fly in nowadays? Is it just like.
A Well, I'll always buias to the S eighteen super Hornet, So if I have the chance to fly out again, that would be great. But you know, in general, the thing about aviation is first and foremost is safety. And I think that's one thing that one of many things that the aviation industry does in general is do safety
really well. And we've taken a lot of those practices from aviation and applied them to safety at Amazon, which have proven over the last especially the last year and a half to be very succesful.
Think about right the standards you have to have for flying people.
Absolutely around the world.
Absolutely, Sarahoz, thank you so much.
Thank you for having her.
What a pleasure, VPA of Global Operations at Amazon, joining us right here in our interactive brokers studio.
This is Business Week.
This is the Bloomberg Business Week podcastle Apple, Spotify, and anywhere else you can get your podcast. Listen live weekday afternoons from two to five pm 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 Journal alone.