Bloomberg Audio Studios, Podcasts, radio News.
This is Bloomberg business Week Daily reporting from the magazine that helps global leaders stay ahead with insight on the people, companies, and trends shaping today's complex economy, plus global business, finance and tech news as it happens. The Bloomberg Business Weekdaily Podcast with Carol Masser and Tim Steneveek on Bloomberg Radio.
Hi, everyone, Welcome to the Bloomberg Business Week Weekend Podcast. Another busy week that started with the unrest in California over the Trump Administration's immigration policies and closed with progress on trade between the United States and two of its major trading partners.
We also spend time at Cisco Live twenty twenty five in San Diego. We're going to share some of our favorite conversations from the conference.
There's a lot of good ones. We're going to start this hour with trade. US Treasury Secretary Scott Bessend, Commerce Secretary Howard Lutnik, and US Trade Representative Jamison Greer were in London early this past week negotiating with officials from China in the hopes of making a major trade breakthrough.
The talks lasted around twenty hours, taking up all day Monday and most of the day on Tuesday before wrapping up late that night.
Yeah, it felt like.
They just kept pushing on and pushing on and pushing on. We finally did hear from Secretary Lutnik and Trade Representative Greer, who finally stepped outside close to midnight in London Tuesday and talked to reporters about the high stakes negotiations and how the group finally agreed on a plan to move forward. Here's the US Commerce Secretary.
First, we had to get sort of the negativity out, and now we can go forward to try to do positive trade, growing trade, and beneficial to both China and to the United States.
From me, that's Commerce Secretary Howard Lutnick, just moments after leading trade negotiations in London with Chinese officials.
The framework the two sides agreed upon is based around a consensus from meetings in Geneva held last month that first lowered terriffs between the two countries. At the core of the talks highly coveted rare earth minerals.
As negotiations were going on, we wanted to touch base with an expert on these materials and why they're at the center of talks between the US and China, we caught up with Graceland Baskrin, director of the Critical Mineral Security Program at the Center for Strategic and International Studies. She helped explain why rare earths are so important.
First and foremost, rare earths are in everything, and we've seen that in the last couple of weeks. As companies have not been able to secure the licenses they need, we have seen disruptions to manufacturing.
We've seen it most visibly from the automotive sector.
For example, Ford actually paused the manufacturing of its Explorer in Chicago because it.
Didn't have these rare earths.
But we have to remember they're in defense technologies, they're in cancer treatment, they're in MRI scanners, they're in just about everything electronic. So what we have is both this huge reliant and there's no significant processing capabilities that exist operational today, which means that we are wholly dependent on China for something that we need.
A lot of.
Grazlyn, Carol and I spoke to r J. Scaring, the CEO of Rivian, the ev maker, and we asked them specifically about rare earths, and partly what he said and he's been clear that he is working to lobby the Trump administration. His folks are communicating with them about the importance of rarest But he also told us, hey, necessities essentially the mother of invention, and I'm confident that we're going to be able to do some of this stuff, at least in the near future without rare earths. Do
you think that's the case? Are we going to be able from a technological standpoint not to have to rely on rare earths for some of these things.
We are going to reduce our reliance. I mean, never underestimate the power of American innovation. We created a vaccine for a virus that we didn't know anything about a year and a half prior to that. So we will absolutely see I think a reduction in our reliance on them number one, in terms of how much we need them. We're starting to create engines without them. However, that is a process. The second thing we're going to start to
see is more capabilities come online outside of China. We soft, for example, that Linus is now starting to produce small scale at its facility in Malaysia.
The US is kind of trying to warp speed our own facilities in California and Texas. So innovation is going to be a huge.
Part of reducing our reliance and increasing our capabilities.
All right, but how quickly does all that happen? Because I think the other thing we followed up with him our desk Garrent, is that none of this is going to happen overnight, right, So how long does it take to kind of make the switch? And then in the meantime, how much are we still reliant on China or how long will we probably be relying on China?
I mean, we're going to be relying on China at some level in my opinion. I know opinions vary on this for probably a few years. In terms of being able.
To meet all of our capabilities, we have to remember this isn't just a problem of capital. In December twenty twenty three, China actually banned the export of processing technologies parts of this supply chain. We actually realized we didn't even have the technical know how in terms of how to do those midstream. So one is building the technical know how and that takes a bit of time, it's expensive.
So realistically, even if we start to be able to get off take by the end of the year through domestic capabilities and potentially through non Chinese foreign sources before we are comfortably reliant on non Chinese capabilities for probably a few years.
Gracelyn, how much of this is an environmental story? To what extent is the processing of rare or something that's done in China because other countries didn't want to deal with the environmental consequences of this.
The environment is no doubt one of the challenge challenges that has limited other countries from developing the capabilities.
You know, we use the term nimb not in my backyard because sometimes.
We don't want these things in our backyard and we're very happy for them to be done far away.
So that was certainly one component. The second thing I want to add to the environment component.
Is generally the United States has deprioritized mining over the better part of several decades. Between the nineteen fifties and the nineteen nineties, the US was actually the top rare earth producer in the world. We were the top uranian producer in the world. We had a Bureau of Mind that was responsible in government for coordinating all of our mining activities.
We closed that in nineteen ninety six.
We seeded a lot of our advantages to China and Russia during this time, and that's how they really created a dominant position where the US once was. So the environmental challenges certainly slow down permitting and some of these other dimensions of mining, but we also have to remember there was a broader government deprioritization of the agenda, which is how we found ourselves where we are today.
Well, that's what I was going to ask you. Was it because the government wasn't prioritizing Was it because companies, you know, really concerned about the bottom line and profitability and making money that they said China can do it more cheaply and more easily for us? Is it because we didn't have the labor pool? Why was it that the US went from being a leader to basically, you know, now it's China as the leader when it comes to rare earth minerals.
So if this was a multiple choice exam, all of the above, I hope would be one option.
But there's a few things right, So when.
The government deprioritizes something, it creates a negative signal as a result of government deprioritizing it, and a lot of this moving offshore to China, Russia, Australia.
Over time, then mine engineering.
Programs have less, you know, less need to produce mine engineers and other skilled labor for mining operations, so.
We have a huge workforce shortage here in the United States.
Along with that, again, it is much cheaper to do it in China, where there are low environmental standards, where labor is much cheaper. If you go to a mine in Montana, you're not going to have a worker making under one hundred and twenty thousand dollars.
We don't have to pay.
A worker in China that same amount, So basically the overall production.
Costs fell to That was Graceland Baskren, director of the Critical Mineral Security Program at the Center for Strategic and International Studies.
We should point out after our conversation Bloomberg News reporting that the Trump administration is developing a plan to use Cold War era powers to prioritize and fund rare earth projects it deems critical to national security, and that is according to people familiar with the matter. You can check out the entire story find it on the Bloomberg and
at Bloomberg dot com. So while trade talks dominated the overall news cycle this past week, one top business story for US was about one of the world's largest marketcap companies, and yes, it's definitely a household name.
Apple hosted its annual Worldwide Developers Conference in Cooper Tino, also known as WWDC. The company unveiled its biggest redesign of its operating system in years. It's called Liquid Glass.
Not a big surprise since Bloomberg's Mark Gerrman had reported a lot of the details in the weeks leading up to the conference and got to say, investor is not too impressed as well, as they sent shares of Apple lower during the event keynote.
For more on Apple's new look and what we learned at WWDC, we spoke with Bloomberg Technology co host ed Ludlow, who was at the conference in Cooper Tino.
You know, you have to remember this is Apple WWDC. It is a developers conference. A lot of what is announces targeted at a group of people who are critically important in contributing to Apple's ecosystem. But it was almost kind of like a sell the news kind of event, both because a lot of what we heard Mark German had already reported. We all know that, and also because you know, expectations that Apple at some point is going to give us a lot more on AI. Those expectations weren't met clearly.
Hey, let's stick on the liquid glass portion of this. My question then, in my question now after it being unveiled, is is there a learning curve here for consumers? We've seen mistakes that Apple's made in the past with redesigns Apple Maps more than a decade ago certainly comes to mind there. But but is this going to be a learning curve? Like how different is this going to be as an experien I mean, they're over a billion iOS users out there. They're going to have to get used to those.
Yes, but a billion iOS users. But you have to think about it in the aggregate of people that use a broader Apple ecosystem, so a lot of liquid glasses. Its piration comes from the work that app we're done with vision Pro, you know, literally the UI and UX of that experience using vision Pro, the translucency of it.
And so by making it uniform, you know, having the same look across iPhone, mac iPad os and vision Pro and watch, you know, it just makes you more familiar and I guess open to using the full hardware suite of products, let alone the software suite of products. But again, like, this is a stock that's been driven by trade headline,
trade risk, tariff risk. But it's a company who, like will tell you timetime again that they're just focused on iteration design, you know, executing on making things beautiful every year. And the problem is that last year twenty twenty four, and indeed the year prior when they announced the vision pro So this event the reason I'm here, all of us gathered at WWDC, there was much more of a clear,
tangible news catalyst around it. This kind of goes back to what WWC has always been, which has been for the people, the community that build for Apple, for the App store and for all of those devices.
Okay, and that's I guess part of the value of going to WWDC, right ed, is kind of talking to some of these developers, feeling the mood about what they are saying. So what are you hearing from them? These are the kind of insiders, right So I'm just curious.
So yeah, we actually did get some AI news, and that piece of news was that Apple is going to open up its on device LMS, Apple's own work in proprietary models to the developer community. The reason that that's a substantive piece of news is that's exactly how it works with open ai and Microsoft and Google and the Gemini teams. A lot of the success they've had in basically commercializing that technology is to say to developers, have at it. Here's the underlying model. Use it to make
whatever you want to do better. And that makes the iOS as an example, more appealing. So that was a big step forward today.
But to the.
Outside of the investor community, all of us that are consumers that have iPhones, there wasn't something there for us in that respect. And what was so interesting is that sort of dropping the stock at the start of the keynote.
It actually.
Coincided with Craig Federigi, that the Services software chief, saying we will have a lot more to talk about with Siri later, you know, because this had nothing to do with that the way that one as a consumer interacts with their phone right now.
Later, how much later, I mean, our investors and members of the developer community still waiting for a SERI that is as smart as some of these virtual assistance that we've seen from LMS at this point.
Yeah, it's interesting, like you know, I'm here at WWDC, but I still was following what German was putting in the blog in real time, and he made the point at the end that Craig was kind of at pain to point out, this is what is new with Apple Intelligence. Here are all of the things that Apple Intelligence can do, rather than you're going to have to wait till later in the year to get more functionality out of it.
On the other hand, like one of the things we got today was live translation in the I message and cool context but also through AirPods.
Right, But I even get.
Like a sense of deja vous when I was at metaconnect at last fall. You know, Mark Zuckerberg was there on stage with his Rayvan Metas demoing exactly the same thing. You know, the ability to live translate in a nation like America where many people are bilingual. It's awesome when you consider the number of people that use an iPhone and are on the iOS system or macros or iPad os. But it ain't that new, and all the other technology companies are doing it as well, and that's kind of
the frustration. It's like, give me the all singing, all dancing version of Apple Intelligence, which we interact with through Siri principally, and that's the bit you're going to wait for. And I didn't answer your question to him because they didn't say later in the year.
No, that's it, that's okay, that's okay. I mean, look, it's a question that you know, nobody really has an answer for. Maybe Apple doesn't even know at this point.
Well wait, it's kind of like, hey, Siri, like, what are you going to get better?
Okay, Okay, let's not make everyone's phone freak out. Carol Masser, Hey, ed before we let you go, You've been to a lot of these events, and as you mentioned, this is one for the developer community and Apple's app store and what you can do on an Apple device would be
nothing without the developer community. Have you had a chance to hear from any developers, to speak to anybody who's attending this and just get an understanding for what the vibes are in twenty twenty five when it comes to the developer community.
So literally.
We had a guy called Paul Hudson on the show earlier. He is the creator of hacking for Swift, Swift being the platform that many used to code for OS application and OS software and kind of what we've gone over is exactly why he was excited that. You know, when you build for one of the Apple operating systems, you usually use an Apple tool to do so, and he's
you know, it was pretty jazz and excited. A lot of people look at Craig federigas like the man, the rock star, the silver Fox of software, and on stage, you know, they got what they were asking for to some sense that like having on device LLM access is huge for a lot of developers. It just makes their life in doing something easier. But the root of your question is the most important bit. We have the App
store today for a reason. It is attractive to consumers, and there are lots of apps there that are made by third party developers. You know, I'm trying to think of a great example that Apple didn't make and someone else did, but you get the idea. And that's why all of this kind of update today is tracked because it adds value to the broader services and ecosystem offering that Apple has.
All Right, basic metric, you got Apple shares down nineteen percent year to date, Look at something like meta it's actually up nineteen percent year to date. Yeah, what is it that the investment community really wants to see from Apple to give it some love.
So I'm actually just pulling chat GPT up on my phone because I was talking to chat GBT about this earlier. Fewer than sixty percent of analysts rate Apple, are by the lowest among its big tech peers. Apple's AI efforts are seen as safe by analysts. Reminder, it trades at twenty seven times earnings, which is slightly off its historic high. The point is this, so far this year it hasn't traded.
I know, it's just so much fun to like use this as a tool, you know, in putting data from the Bloomberg But like, so far this year, that nineteen percent, it has nothing to do really with the AI stories. You guys know, it's tariff for risk, the relationship with the administration and the presidents of the President of the
United States. But the risk going into today and what was almost confirmed in the early minutes of that keynote was well, what if AI does become the key risk of the stock because investors lose patients and they say, we want to see something more tangible. It's a little unfair because I don't think we went into today, thinking we would get anything like that. It was really clear what we were going to get. It was for the developer community, and we did get it.
And we do have some extra time with you, which were grateful for and you're you know, you work, keep in the seats warm for you here in the San Francisco studio. Liquid glass. You know one thing that Apple is trying to figure out, and I think it's fair to say every consumer tech company is trying to figure that out. Is the form factor for the must have device of the future.
You've talked a.
Lot about where humane did not work. We've talked to you a lot about what we could see from Sam Altman and Johnny Ives. When it comes to what you saw with liquid glass, does it give any indication as to how Apple is thinking about this?
It does.
Look.
Look, what Mark Gumman's reported is that the seventeen generation of iPhone will have glass curve size with no sort of screen border, right. And liquid glass is a concept, the kind of translucency of it, the way that you whatever is in the almost the foreground of your your smartphone screen or iPad screen is not interrupted by what's popping up. We think that there will be some hardware toiration on that to come. But you do raise a really good point, which is why have liquid glass uniform
across all devices? Because increasingly it's not the be all and end all. The thing that's in my right hand a smartphone. It is that I take messages on my Apple Watch. Other smart watches are available, I take, you know, different platforms that I rely on OS for on my MacBook. Other laptops are available, et cetera. You know, it's just a way of making it more uniforms. So I think
that there's definitely acknowledgement there. Whether the iPhone seventeen has the features that Mark reported, we shall wait and see, but there was certainly a nod to that during the keynote, even if they didn't explicitly sail by the way, the next version of the iPhone it's definitely going to leverage this liquid glass concept.
Great stuff, ed Thank you so much.
We know a busy day, early day, Edla though, of course, co host of Bloomberg Technology out there at Apple headquarters in Kupertino, California. Of course, at the Worldwide Developers Conference.
You are listening to the Bloomberg Business Week Daily podcast. Catch us live weekday afternoons from two to five pm Eastern. Listen on Applecarplay and Android Auto with the Bloomberg Business app, or watch us live on YouTube.
President Trump's immigration crackdown was definitely top of mind this past week. The administration's pushed to ramp up arrests and deportations led to protests bringing up across LA and to other cities.
The big push for more arrests means there's a growing need for detention centers, and in several towns across the country, private facilities are picking up the slack.
It's interesting to consider that while some communities, including LA, are demonstrating how much they dislike the immigration policies, politicians and leaders and communities that have these private detention centers, some of them feel differently.
Bloomberg News Investigations reporter Rachel Adams heard was behind a big take on the budding deportation economy. She recently joined Carol and Me to discuss what she learned about these so called ice towns. Welcome to a Mana's ice towns. These are communities convinced that their financial survival depends on locking people up. It's not a new phenomenon. In this iteration of the classic prison town, though, Carol, many of the people behind bars haven't been convicted of crimes.
Let's get to this Bloomberg BusinessWeek story. It is, I guess you could say, a slice of life in America, but not necessarily a good one. But let's get into it with Bloomberg News Investigations reporter Rachel Adams heard she joins us from our Houston bureau. Rachel, good to have you here. What is an ice town?
So, an ice town is a town where the local politicians, local community is convinced that they need these immigration detention facilities that operate under these lucrative contracts with ICE in order to be financially sound. A lot of these towns they've had a private jail or private prison detention facility for decades, and with Obama era sentencing reform, the federal prison population dropped off, and so to keep these facilities profitable,
they signed contracts with ICE. And now we're seeing just a huge influx in the number of people detained at those facilities under the Trump administration's, you know, mass deportation agenda.
So, Rachel, take us too one of these ice towns. Torrence County, New Mexico.
Yeah, so we went to Torrance County, New Mexico, and at the end of March, that's when the County Commission was voting to extend this contract. And it's a really rural community. It's one of those areas that used to be kind of this bustling hub thanks to the New Mexico Central Railroad. You had people from all over who would come through towns like Estancia where this detention facility is located. But then inter State forty is built and
commerce really just drops off. And so during the Reagan administration, when you're seeing these policies like the War on Drugs that's massively expanding detention in America, there are these private prison companies that want to get a piece of that business, and New Mexico was a great place for that for them. Because land was cheap, they could build these really big detention facilities by the standards of the time, and so
that's what happened in Astancia. Corcivic then Corrections Corporation of America, built a facility outside of town at the time. It's since been annexed into the town, and things were pretty good for a while. It employed people. Business was good for Corcivic, but then around twenty sixteen you had the federal prisoner population drop off and that's when the facility ended up closing and Corsific said it was no longer profitable.
And then ICE came in to basically save the day for Corcivic and offered this contract that guaranteed revenue every single month whether or not there were any ICE detainees in those beds, and that's the contract that allows the facility to operate today.
So just to set the record straight, I mean, these are facilities that have been around for a long time, different administrations, correct, Correct, Okay, And the point is too though, that the conditions aren't great and they're needed in order to house kind of the roundup that is happening under this administration.
Correct.
So, during the last ten years or I guess five years since this contract has been in place in Torrents County, a lot of the people that it was detaining on behalf of ICE.
They were recent border crossers.
These were people who were coming up the border from Mexico a lot of times into Texas. They were seeking asylum, and while they were seeking asylum, they had you know, turned themselves over in a lot of cases, and now they were in ice detention, being held at the Torrens County Detention Facility and Astancia. After Trump took office, you see a really dramatic fall off in the number of people crossing the border. And so all of a sudden
you have room in the Torrens County Detention Facility. And so because you have this room, you're seeing ice flying people from all over the country, a lot of them coming from New Jersey, from Florida, and they're flown into the Albuquerque Airport and bust to the Torrents County Detention Facility and that is where they're being held for months
in a lot of cases. And this facility has had problems for the last several years, and it's really starting to kind of come to a head with the increase in the number of people that is holding under the Trump administration.
Okay, well, you describe the facilities at least one of them.
What was it like?
What is it like?
I mean, what we've heard from people who are currently being detained at the facility is that it's cold, the food is undercooked, inedible at times. We've heard complaints about
access to medical care. There are a lot of people who say that they have tried to ask to be deported because they do not want to spend another day in iceed attention, but that it's incredibly difficult to get the attention of any ICE officers because it's all you know, prison staff, not necessarily anyone who can actually talk to them about their case. But the most dramatic issue that we've heard about recently had to do.
With the water supply.
This is the desert, it's incredibly dry, and the town of Astancia was having an issue with one of its main water pumps, and so that led to water being intermittently shut off to the facility. We heard that there were toilets that were overflowing, and that there was feces on the floor, that the smell was overwhelming. We heard people say that they had been limited at one point to two water bottles a day, that they were thirsty.
At one point, the facility brought in trash cans full of waters to try to flush some of those toilets to deal with the odor and the hygiene issues, and someone we talked to said that he had seen people who were thirsty drink out of those trash cans because they were so desperate for water. So it sounds like this is improving over the last few days. We did hear that ice was on the ground recently, but certainly it sounds like it has been a rough couple of weeks inside the detention facility.
It's interesting because I'm thinking Corcific, right, which is the company that's been around and is running these facilities, right, I guess it's close to four deck for four decades.
It looks like I just.
Remember that period and like investing like private prisons, everybody thought it was like this great investment play, but you got to remember it's people, it's humans. There are conditions. I am curious, like are these people being held for due processing? Like who determines the fate of these individuals is ultimately the federal government.
I mean a lot of it is held up in immigration courts, and that's an incredibly slow moving process. So I'm sure you all know. One of the issues with these multi state transfers. I mean, people are thousands of miles away from their families, but they're also far away from their lawyers. If they had lawyers in Florida, for example, because their case was there and now all of a sudden,
it's in New Mexico. They have to try to find a lawyer that can represent them in a different court, and that isn't always easy to do, especially because these tend to be more rural areas where pro bono lawyers are harder to come by, located farther away from the facilities themselves. So just the act of moving someone to a facility so far away, it does impact that person's right to do process.
There's some context here and sort of like the internal workings of how these communities feel about these that I think is really interesting. In your county commissioners and how county commissioners look at this, you highlight Linda Yamario, who of the Torrance County Commission, who actually did go inside one of these facilities, and then you and the team
wrote that she voted to extend the ice contract. Talk to us a little bit, and it's not really town gown relations, but it's the way that the community feels about this and the way that local elected officials feel about these.
Yeah, it's really striking to sit in on these County commission meetings because you know the contract is up for an extension. Every so often, every several months, but no matter what, at every meeting which is held, you know, bi weekly, there are a bunch of lawyers and advocates and even people who are in detention who record testimonies that they play to the commissioners during public comment, urging them to shut down Torrance County Detention Facility for ice detention.
And you have these three commissioners who sit at the front of the room and they're listening to this and some of the testimony is in Spanish, and Commissioner how Tomeo speak Spanish, and so you can kind of watch her face as she's listening to people describe the conditions inside the facility, and what she told people at that meeting in particular was, look, I hear all these concerns, but I don't know what to trust. And so she ended up voting to extend it. Sorry I lost my earpiece.
But it's unclear whether she'll do that in October, because she did go again and say that that things seem to be going well, but she does still have concerns.
You know.
I think it's interesting and what I find fascinating is that this is a story that's got so many different angles to it.
When you think.
About what these detention centers do for a local economy, there's that, but there's also then, as we've been talking about the conditions, and you do wonder about the oversight, especially in an administration that has been cutting back on some of the federal spending in different areas. So, you know, Rachel, when you look at this story and this reporting, I mean, the cities or the local municipalities in many ways want them right because it provides jobs, it provides some economic momentum.
At the same time, there are concerns about kind of who's watching out for what's going. What does Corcivic say about all of this?
Yeah, Corcivic talks about how many people they employ locally. They talk about the salaries and the benefits that they provide. They you know, say that they I think they said that they hire or have more than three hundred jobs related to these facilities in New Mexico. So, I mean, they definitely see themselves as a major local economic benefit to the area. And that's you know what we heard
from the Estancia mayor as well. I mean, he talks a lot about how much of the grocery seats tax which is New Mexico's version of a sales tax that course of it contributes to. So it's interesting to hear kind of the I mean, I think Commissioner how Amos, Yeah put it with ysinkly when she had like on the one hound, they're a major employer.
Yeah, such a great story, so much in there, and as we said, it's a slice of life in this country. Bloomberg News Investigations Report Rachel Adams heard Rachel, thank you.
This is Bloomberg.
This is the Bloomberg Business Week Daily Podcast. Listen live each weekday starting at two pm Eastern on Apple car Play and the Android Auto with the Bloomberg Business App. You can also listen live on Amazon Alexa from our flagship New York station, Just Say Alexa played Bloomberg eleven thirty.
Let's Turn to Healthcare. US Human and Health Services Secretary Robert F. Kennedy Junior removed every member of the CDC's Vaccine Advisory Panel in an opinion piece published in The Wall Street Journal. He said he made the move because members of the committee were receiving funding directly from pharmaceutical companies, a claim that so far has not been confirmed.
With uncertainty at the federal level. We wanted to get an overview of the industry with an executive who thinks about the space a lot. We turned at Wendy Barnes. She's president and CEO of GoodRx. We talked to her about the headwinds facing her industry and it really ranges everything from tariffs to the impact of new regulations.
A lot to talk about. She kicked it off highlighting a new company initiative aimed at supporting independent pharmacies in the US. It's called Community Link.
You know, I would say the company's been on a multi year journey to best partner with retail pharmacies, with community pharmacies really being at the top of that list, and this is really the culmination of that work. And so what you referenced, the good RX Community Link is a portal by which community pharmacies sometimes referred to as independent pharmacies, So think of those as really your non
chain pharmacies. It can be as small as a single pharmacy, or some community independent pharmacy owners have fifty seventy five or one hundred locations, but largely they're run independent of a broader management philosophy and this is our effort to contract directly with these pharmacies in a cost plus reimbursement mechanism manner.
Look, it's not.
Any new news to you or probably to your listeners that pharmacies continue to be under reimbursement pressure, and we believe that they are key partners in our broader mission to make medications more affordable and accessible for every American. And so this effort is really one in which we're trying to bolster reimbursement for those independent pharmacies such that
they can work with us directly. And in addition to that, we're giving them access to ninety plus brand deals that we have secured with pharmaceutical manufacturers over the last to twenty four months, and those continue to grow such that the reimbursement on those same drugs is favorable to those pharmacies.
It's really no secret that pharmacies have long struggled to have a favorable margin profile on many brands that they fill, and this too gives access to those programs for those independent pharmacies.
Well, Wendy anecdotally speaking, as the large pharmacies and the chains have come under pressure, whether it's overstore closures or different things happening just in the space we've all reported on what's happened in cities and the way that some of these pharmacies have closed down in certain cities. Are
you seeing more people go to these independent pharmacies? I mean, anecdotally speaking, I can say that certainly the way my family has changed its behavior just thanks to availability and what pharmacies tend to actually have what we're looking for. But how have you seen that affect the overall landscape.
I don't know that we've seen a meaningful ship to independent from chain. I think there's still a pretty good mix of grosser versus independent versus chain, and candidly, you know, mail order slash digital pharmacy.
I think at the end of the day, what we see is.
In the seventy plus odd thousand pharmacy options that we have as consumers in the US that as a consumer, you just really want to get your drugs on your own terms, whether it's mail to your home with your preferred community pharmacist or at a chain or grosser. And we pride ourselves on working really with all pharmacy such that you can get your drugs when and where you know you desire to do so.
Do you see it all that the independent channel is shrinking. We've certainly seen it with some of the big pharmacy chains, right, We've talked about it, You've seen the headlines. But I'm just curious when it comes to independence, Wendy, do you see any shrinkage in terms of the number of outlets that are out?
There are no, you know, I.
Will say in the numbers that we've looked at over the last couple of years, there are still a good number opening any month. I will say in our book, with the pharmacies that we work with, I haven't seen significant trinkage. But I think there are certainly other data sources that would suggest that there are a number of independent and or community pharmacies that are closing in or selling their business to other owners.
But in our particular book, it's held pretty study.
How have independence traditionally reacted to the program at GoodRx, because it's our understanding they've been a little skeptical about it. How do you make sure that they're embracing it rather than treating it with skepticism.
Sure, well, I think you know, Look, it's early days, to be clear, and we're going to be on a journey here. Communication is going to be the biggest part of this program, in addition to delivering what we said we intend to deliver, which is fair economics over the course of these contractual agreements, as pharmacies agree to contract with this directly, and that again can all be conveyed
through that portal that we stood up this morning. So for me, it's really going to be about how we deliver on the reimbursement mechanism through this partnership, and.
I think, you know, the ongoing communication.
Will support how the pharmacy community embraces this over time. You know, we've certainly been setting out communications early on.
We've been in dialogue with.
Many independent pharmacy owners to speak through the program. But I think at the end of the day, it's really going to be the outcome of the program and the economics that we deliver that we'll prove it out over time.
Well, that's why I want to kind of just dig a little bit deeper, Wendy. You know, for this to work for you guys, this new role out, this Community links program, the economics have to work right for the independent pharmacy. So I'm just curious what you can tell us how the economics are different for the pharmacy that directly contacts with GoodRx versus using maybe one of the pharmacy benefit managers the PBMs and their networks to dispense
a prescription. What can you tell us more specifically about the economic advantage for independent pharmacies to use your route versus going to the PBMs.
Are happy to do so.
So it's really the agreement directly with US is rooted on a NADAK plus reimbursement. So NADAK is one of multiple benchmarks that is effectively a cost mechanism that stands for national average drug acquisition cost, and we are negotiating with independent pharmacies to do this with that predicated as the baseline plus an amount that keeps them profitable on filling these claims. So that's really the precursor, if you will, and or the meat of how this agreement will work,
which again provides favorable economics to these pharmacies. The distinction, Carol, as you called out to perhaps doing it through a larger payer, when you think about how it works through let's just say a larger PBM, there are multiple patient pay slash cash networks through which an independent pharmacy can participate, and when they're in that type of an algorithm, for lack of better description, typically it's going to search for
not necessarily favorable margin for the pharmacy, and so there would be times in their previous arrangement whereby they may not always have favorable economics to fill those prescriptions. And by contracting directly, we are putting terms and conditions in the agreement that allow them to have certainty around not only the reimbursement, but also access to drugs that they're
not fulfilling today in patient pay programs. Again, that would be the list of brands that I mentioned that I believe are ninety plus at this point that they're going to have certainty around how they're reimbursed in this direct engagement through us that they're not getting through their arrangements today.
Thinking about healthcare in the United States, it is inextricably bound to politics, and no question, if what the President calls the Big Beautiful Bill were to pass, we could see, at least according to the CBO estimates of a dramatic increase in the number of uninsured almost eight million. What does that mean for your business and how much your business comes from Medicaid and uninsured people today.
So interestingly, of the users who access our pricing on any given day, roughly ninety percent actually have insurance today. So I think there's a bit of a misperception that the overwhelming number of consumers who use us in fact don't have insurance, which is not the case.
Having said that, who does.
Come and look for pricing through our different platforms is your consumer who's motivated to check for competitive pricing, whether they have insurance or not, and we continue to be the number one platform to look for the best price deal that you can find on any particular medication.
So as it pertains to.
What may or may not pass in the bill, whether it's inclusive of Medicaid cuts which continue to be advanced, and or other cuts that would put pressure on your typical how hold income, what we do know is that those should produce tailwinds for us as a business for consumers whether they're insured or not, but specifically as it relates to Medicaid cuts and the number of individuals that may fall out of coverage, we do believe that that presents an opportunity for us to continue to support any
American that really is struggling to purchase their medications.
Wendy just got thirty seconds. You know, the word that we all use to describe the environment right now is uncertain. There's a lot coming out, certainly leaders and CEOs, just in your industry too. You've got write aid doing a bankruptcy. You know, I think about the macro within your business and just macro overall. How do you look at the outlook?
Just quickly, you know the outlook?
You're right, turbulent is a fair description. Having said that, businesses that stay the course and have a mission that makes sense for the consumers in which we serve will continue to thrive our goal of courses getting medication affordably and effectively into the hands of every American you need it, and we believe will be.
In a position to take advantage.
Wendy Barnes, President CEO of GoodRx.
You're listening to the Bloomberg Business Weekdaily Podcast. Catch us live weekday afternoons from two to five pm Eastern Listen on Applecarplay and Android Auto with the Bloomberg Business app, or watch us live on YouTube.
Plenty ahead in our second hour of the weekend edition of Bloomberg Business Week, including a deep dive into Cisco Live twenty twenty five, held this past week in San Diego. Tim and I were there.
We were broadcasting live with some of Cisco's top experts and leaders as we unpacked what's next for the company and how it's navigating the ever evolving world of tech.
Got to say, it sounds like Cisco is definitely going into a big new chapter for the company that has been around for a long time. We want to share with you a few of our common conversations from the event, and there's no better place to start with than AI,
which really dominated a lot of our conversations. CISCA recently announced that it's updating it's networking and security products to make artificial intelligence networks work better, and we got into that in more with Cisco's Vice president of Product, AI, Software and Platform DJ Sampath.
DJ here, I mean, you look around at any of the trade floors and like, this is the massive, massive, and dominant theme here. I want to talk a little bit about the news that we got earlier today that Cisco's updating its networking security products to make AI network speedier and more secure. It's part of this broader push to capitalize on the AI spending boom. Talk to us a little bit about what prompted this, and we're going to speak about it throughout the program because this is
some big news to emerge. But what you're hearing from customers on the customer front, where we are in the AI moment, and really what prompted a move such as this.
Yeah, Tim Carrol, thanks so much for having me. You know, there's a huge tectonic change that's happening right now.
If you think about the moment where chat gpt became and I came into our zeitgeist in November twenty twenty two, we've now started understanding that AIS here, it's not a buzzword. We go to chat gpt for a lot of questions and existential ones like, you know, should I take this job, should I not take this job?
A lot of interesting things.
That you start to see that are being broken down very systematically by AI as a TOT partner.
Right.
But what we're seeing now is that evolution from AI being an application or a chatbot to an agent agentic AI. Right, So you've got agents that are autonomously going on and performing.
Tasks for you. And as you start to think about that.
What's going to happen as your workforce, which has now gotten you know, one hundred percent of humans, is going to be augmented with you know, AI applications, AI agents, robots and humanoids.
You know, in fact, you know you're going to start to see that happen.
And these these robots and humanoids are going to have AI models sitting inside of them.
So what we're.
Seeing is that you're going to start to see a lot of change in the way these networks work and how employees are going to work inside of these organizations. So so we're we're just tremendously excited about that, and we're getting ready at Cisco.
We know it's coming.
We've we've helped usher in the era of the Internet, and we're now ready to usher in the era of AI AGENTIC AI more more specific.
What's surprising you about kind of the progression? You're right, it's like two years ago.
Right, it was like bam.
But I think for us, like it was just over the last few months that we have increasingly been using like chat, GBT, right more aggressive.
Way use a lot.
There you go.
Yeah, here's what I'll tell you. Right in the you know, in the trenches right now. The speed with which AI has been moving has been absolutely mind boggling, you know, the progress that we're starting to see. I'll give you an example. You mentioned Claude. Claude is an amazing, you know, piece of uh A software from a model perspective that it loves you to write code really well. Claude is one of the top ones that they used to write
new software engineering. A lot of people thought that, you know, AI is gonna, you know, sort of come for our jobs, and you know, software engineers are probably not the ones that it's going to come for. It's going to take a long time to get there. But guess what, you know, it's here and it's now, but it's not taking away the jobs. In fact, every single developer is becoming a lot more productive using AI. So essentially what's going to happen is it's not like you're gonna have to worry
about jobs. I think, you know, people that use AI are gonna you know, take the jobs of people that don't use AI. So you're going to just start to see that AI becomes extremely helpful from a productivity perspective.
You know, I want to go back to agentic ai, and it's I know, it's something that you think a lot about. We hear it a lot on from Salesforce for example. Right now, I still think though, when it comes to the general consumer, agentic ai is like for the early adopters right now, it's for the people who do want to pay two hundred dollars a month, for example, for a version of chat GPT that will make reservations for you at a restaurant or help you with travel plans.
What does it look like beyond early adoption.
Yeah, I'll talk about the you know, one of the key things that you'll you'll notice is that the rate of which things move, Like, you know, if you go back to just using the chatbot, the very first time you use chat gpd'd had a knowledge cut off, right, you can only get information from like twenty twenty one.
And then they opened it up to being able to do live web searches, right that is actually you know, teaching a model how to use the tool, and then you saw that it went from just being able to do a search to like, hey, you can now start to connect.
These models to your emails.
You can now start connecting these models to your salesforce and instance, and then slowly it's expanding. It's now connecting into your Google Docs and your Microsoft Office, you.
Know, environments very quickly.
Or what you're seeing is that you know, we're in this in this exponent of like the speed with which things are moving, and we're just at that beginning part of that exponent So you're right. Sometimes it feels like some of these things are very early from a consumer perspective, but if you started thinking about what's happening inside of the enterprises, the enterprises are starting to adopt AI a whole lot faster because they need to maintain the competitive edge.
But a key thing for them to be able to adopt AI is security.
Is safety and security You've got to make sure.
Well, that's what I was going to say, because in order for you really to kind of max at the potential of it, you've got to open it up right to kind of all your different systems.
So how do you make sure that that is really really secure?
And that's top of mind when it comes to a gentic AI because you are now saying, hey, you're going to allow these agents to execute tasks on your.
Behalf all the way to completion.
So which means you've got to make sure you know you're thinking about safety and security. And now I'll talk about what we launched earlier this year at Cisco. We had a product called Cisco AI Defense. Knowing that this is going to be top of mind for a lot of enterprisis, we build a technology that does you know, provide you with the visibility of all.
Of these agents, gives you.
The ability to validate these models, to make sure these models are not vulmotable.
Because here's what happens.
Right if you go back to the age of the Internet, when the Internet came out, the attacker started doing a whole lot of interesting things. They would do the dial up service attacks and DDAs and so on and so forth. We're starting to see similar types of things happen with AI applications. So we've got to make sure that we're protecting against attacks, new types of attacks like prompt injection attacks and so on and so forth.
What okay, So what about the type of attack where this is a little scary, but I think this is what's happening, where we are mimicked, our voices are mimicked, and perhaps a loved one is called and AI is able to pretend that we're in a situation that requires the help from a loved one. I mean this could happen, This could happen.
No, absolutely, like you know it's happening now.
Just to be very clear, because spam in phishing attacks with AI can become really really normally, like you're going to see a lot more of these attacks. That cost to do those attacks are going to drop dramatically, which means you've got to use AI to be able to you know, defend as well. It's not just about using AI for attacking, you use AI for defending. So there
are two parts to this. You want to make sure you're securing the use of AI, and you also want to make sure that you're using AI to make security better.
And at Cisco we're doing both simultaneously.
So what's the thing that keeps you up at night?
Like, I just do think about some of this is happening, Like we've moved so quickly. It seems like all of a sudden from generative to agentic AI, and I think people are just catching up to be like what is this all about? So what is it that kind of worries you as things move so fast?
I think.
I think one of the one of the biggest things is, you know, there's always this feeling like we're not moving fast enough, you know.
Because do you feel the pressure from customers.
I think at this point in time, everybody's feeling the pressure when you talk to anybody that is in AI that the rate at which AI is moving, everybody's just trying to make sure that they're they're they're keeping up with the advances that's happening across the entire ecosystem.
Hey, DJ, I'm curious about the fragmentation in the security market right now, and I'm wondering how you see Cisco potentially playing a role when it comes to consolidating all the features and the functions that are out there.
You're you know, this is a top of mind issue for a lot of cisows, you know, chief information security officers that are looking to consolidate a lot of the different point products that are there inside of their ecosystem. Cisco, as you know, you know, with the acquisition of spun because one of the largest security companies on the planet at this point, and we've taken a very intentional approach to consolidating a lot of these things into a single pane of glass.
What we launched earlier this year was something called.
Security Cloud Control, and this is a consolidation effort. You know, you don't have to worry about each and every single point products instead of your ecosystem, but you bring it all together and you use AI assistance to be able to manage all of these products. So, you know, we're entering into that AI and agentic era. You've got to have those capabilities so that the defenders have the advantage over the attackers that are cloning your whits and doing
those fishing attacks. You really need to give the defenders an advantage, and that's really what we're doing at siscoll.
You know. One of the things, you know, when we look at kind of the universe, your competitors as well, and when investors think of networks security, they often think of palolp On Networks and some other players that are out there, especially when it comes to the firewall space. I'm just curious, are there products dj or software that can kind of change market perception, investor perceptions and market share dynamics.
Absolutely one of the key things about you know, the way you're starting to see these attacks evolved, especially in the age of you know, AI and agentic era, you got to realize that one of the core enforcement points for stopping all of these bad actors and bad attacks is in the network.
Now, when you take that security and the profits of security that.
We have and refuse it into the fabric of the network, that's when you're really moving the needl And so when you think about Cisco, you see that we bring both of these.
You know, we've been doing networking for the past forty.
Years long time.
We've been doing security for a pretty long time too. But when you start to combine those two things in a single place, man, like we've got we're miles ahead of everybody else that's out.
There thirty seconds.
What's surprising you that you're hearing at this event, like we're just blown away and the conversation just get about thirty seconds, Like what are you hearing as kind of a dynamic narrative.
One of the core things that I'm seeing that's that's really surprising is the number of customers that are implementing AI infrastructure projects inside of their own enterprise environments. They realize that AI cannot be somebody else's AI. AI has to be their own and they want to run this inside of their own data centers, and that to me is very welcome, and I'm very happily surprised by that because we're ready.
To help them.
So that means a lot of on site on premise, right, It's exactly right.
It's not all up in the cloud.
Not all up in the cloud, because they have lots and lots of this data that's proprietary that's sitting inside of their own environments that they're not quite ready to push it into the pod.
Which I think is surprising. But like here, it's happening. DJ, thank you so much. Great way to kick off our coverage here.
Thank you, Carol. Thanks dam.
DJ Sampath, Senior vice president of Cisco's AI Software Platform Group.
This is the Bloomberg Business Week Daily podcast. Listen live each weekday starting at two pm Eastern on Apple car Play and the Android Auto with the Bloomberg Business app. You can also listen live on Amazon Alexa from our flagship New York station, just Say Alexa played Bloomberg eleven thirty.
We are across the country. We're on the West Coast, staying here. Some great conversations here at Cisco Live twenty twenty five. There's so much happening around Tim and me, and we've got a great guest to kind of get into some of the enterprise connectivity issues collaboration, because that's what he oversees.
Honor Rock Deingra is Senior vice president and General Manager of the Enterprise Connectivity and Collaboration Group at Cisco. He's also founding executive sponsor of Cisco's Responsible AI Initiative, which means we're going to talk a little AI as well. I want to talk you. You ever see this diverse portfolio
of collaboration tools that are all under Cisco's umbrella. When you think about what collaboration in the workplace means right now with a distributed workforce, with some people working from home, some people at the office, all at the same time, what's the one solution that can bring them all together?
Thank you for having me Tim on the program.
So when you think about collaboration, it's more than just the tools you used to meet with people or to call people.
You know, obviously those are the essential tools.
Where it starts with foundational technology, which is the network you want to make sure that everyone can connect seamlessly and securely to the network first, no.
Matter where they're working from.
And you know, our definitely of the workplace is quite expansive, and we talk about a regular office, we talk about a retail store, a factory floor, or your home or your car.
Wherever you work from is a workplace.
And so it starts with foundational elements of connectivity and then once you're connected, then it's all about delivering the best possible experience. And so one of the things that Cisco has been innovating in is this concept of distance zero.
And the idea is very simple.
When you're meeting with other people who are not in the same room with you, it should feel like there's no distance between you. And so we've been working on technology, whether that is video conferencing technology, whether that is services that we deliver to amazing experiences that bring people closer to.
What's the biggest hiccup that prevents people from actually experiencing distance zero? Is it connectivity issues?
So I think it's a mix of things, right, So, first, this connectivity has to be rock solid, and you know, when you're thinking about the diversity of spaces that we're talking about, sometimes you don't have a solid Wi Fi in a conference like this, it can be very tricky Hotel.
As an example, like getting gay dropped on right the computer with spine the phone was and it was like it was actually really really hard to work.
Yeah, we take connectivity for granted, but it's actually still very complex, and especially for people who build and manage these networks, the complexity is only growing with emergence of AI agents and digital.
Workers in the workplace, It's only going to get more complicated. So we're ready focused on how do we deliver amazing.
Experiences for people who are connecting, but how do you make it easier for it to manage it all under ug.
Our networks generally ready for AI or would you say no, I.
Would say they're not ready right now. And here's why I say that. Look, everyone can feel that this is transformittive technology and we're at the beginning of a major technology shift. When you look back at other technology disruptions, you know, the beginning of the Internet, mobile devices, cloud computing, it causus to take a step back and reimagine how network should be built and managed. And AI is much
bigger than that. When you think about the proliferation of AI agents in the workforce, digital work that are doing human like work. You know, they're communicating and doing work like us, but at machine speed and scale. So the networks have to really deal with an influx of much bigger traffic patterns and security soface cities that are exploding there.
Well, one thing I want to ask you. I know we're like we bought, although we're like dying to ask questions. I am curious.
About because we keep thinking about the Capex spand right and all that is AI, But I'm also thinking about are they not going to do the spend because they're worried about the power being there for all the AI? And I use the term AI loosely. Right, there's a lot going on, but is the power not necessarily there and the systems there? We keep doing some reporting on that front that's going to slow some of that spend.
So I think, first of all, I think this is going to be a central technology. This is almost like if you ask someone to do you want connection to the Internet, of course the answer is yes.
So this is not going to be optional.
Organizations have to figure out how to enable their workforce with AI tooling, But how do you do that in a scalable fashion? That's really the question. And I think there are two parts to this power equation. So one is for training models. These large data centers that are being built out around the world to create these models,
they are very power hungry. But when you think about deploying your applications in your office environment, you have to make sure that even there you're building scalable and devices that are power efficient, and that's what we're focused on with the number of new product announcements that we made here. These network devices, whether there's the switches or routers or wireless access points, they're built to be very efficient from a follow consumption point of view.
Because you over see this portfolio of tools that are meant for helping collaboration across different areas, you must have some really cool statistics when it comes to who's working from home, who's returning to the office. We've been talking a lot about this, even five years in past COVID. Even in London, for example, many banks are facing a shortage of desks because of a real estate squeeze, but they're trying to get people to come back into the office,
so there's this tension happening there. What are the patterns that you're seeing emerge in different parts of the world right now when it comes to return to the office, Yeah.
So I think most companies have settled into a rhythm.
Most companies have people coming back into the office at least part of the week. Now, it varies quite a bit across the globe, and it also depends on the organizational culture. But what we're seeing is most companies that are back in the office is two to three days a week, at least in the United States, there is generally a culture of remote work and hybrid work, much more than what we see in the rest of the world. But even here, most companies have settled.
Into a rhythm.
Now, what's essential is even if people are in the office that's three days a week, there are other people who are remote, and you might be working outside of for the organization with partners and customers. So there's always going to be people who are remote who are not in the same room as yourself. And this is why at Cisco we are very focused on how to build the right technology so it feels like there's no distance between you.
So mobility within the enterprise, is it going to impact more the people who are working from home already or is it going to be more influential on the people back at the office.
I think it's both, right. I think it is really a hierarchy of needs for employees.
Right.
It starts with if I'm in the office or in my home office, do I have the tools, the basic tools to get my job done? So that's number one, right, And that's where connectivity and you know, meeting tools and things like that come up. But after that, it's how do I collaborate with my team? How do I get more creative? How do I actually get creative work done when we may or may not be in the same room.
And if I get to the office, I want to make sure it's worth my commute to the office that I can find a desk to sit with my team, I can find a confrom room when I needed.
The enquirement has to be very supportive. If I'm going to make the strip through the office.
I think about kind of the connectivity, connectivity that we have as we go from office to office. Right, whether it's printers, I mean, something basic, but we can it works, but exactly we pop into a system or and.
Everything I try to do to make it not working.
Able to remote access like it is pretty amazing.
It's a reason why they have us practice that stuff serious requirement. Hey, we don't have a ton of time with you, but I want to talk a little bit about AI in a couple of different context One about AI embedded in WebEx and call center products and sort of the progress that you're making there if you're able to monetize it. And then I want to get to save AI in just a second. But first talk with WebEx and the collaboration for it.
Right, So WebEx really has kind of two parts of the portfolio. One is really focused on meetings and callings and things like that, and that's where AI is really an assistant for you. It can generate meeting summarys, it can take notes for you, it can do things for you in the contact center space, in the customer experience space. We have now new AI based agents. These agents are completely autonomous. They can talk to a human as a human.
It's pretty crazy.
So they are voice.
Enabled, but if you want to chat with it, of course, you know, chatbots have been around for a long time.
What's different about these agents now is they can.
Actually take actions, so they'll understand your intent and you know, if you say, look, hey, I want to reschedule my doctor's appointment, They'll go ahead and do that for you, send your notification as a confirmation. And we're seeing a tremendous demand for this technology and this is starting to make a lot of difference in the customer service ex fegion.
So they're not going to get added TODE when I call my doctor's office and.
Like they have.
It's like the Seinfeld where you know, like your folder has been flagged already, so you get directly.
I definitely have been flagged in a good way and in a bad way.
Public I want to ask you though, during COVID you mentioned WebEx. I mean, why didn't you guys take better advantage or more advantage during COVID. Do you just talk to us a little bit about during the lockdowns be you know, just like Zoom did.
Yeah.
So, I think our focus has always been helping large enterprises connect their employees and we were very successful in that space and we continue to do that. CISCOULD never really took this COVID as an opportunity to go into consumers. We did make WebEx available for free to lots of governments. Charitable organization educational institute. So we did that, but our focus was very much around being an enterprise product.
Now, when you think.
About where we've come along with the event of AI with autonomous agents as part of the WebEx and all of the stuff that I was talking about earlier about distance zero and conference room, I believe that the product has come up very long way and it's amazing, and so you actually have to see to believe it. A lot of people don't haven't experienced it in a while, so I would encourage anyone to take a look and see how far we've come.
I mentioned that you're founding executive sponsor Cisco's Responsible AI Initiative, Right, what is responsible AI and what's irresponsible AI? And I want you to answer the irresponsible part first.
Yeah, So when we started this journey several years ago, generative AI is obviously the you know, the trend right now, but we started this even before people are talking about GENERTI VII. The idea really was how do you make sure that bias doesn't creep into AI systems, that if AI is starting to make autonomous decisions that are consequential, that those.
Are done in a responsible way. And how do you put guardrails around those systems.
When we are building technology, how do we make sure this technology is used more for good than for the various reasons. So that's really the genesis of that program, and then we came up with a few principles and then that translated into development guidelines for our peak.
Can you, I mean, when you look across the landscape, can give us examples of AI that's not responsible, something that can concerns you out there?
Well, I'll tell you, Like these systems that are starting to show up in workplaces right now, they are not very deterministic. You know, they're powered by language models that can hallucinate, so they can say bad things, they can make things up.
And so you're talking about deployment of potentially tools that everybody knows about across enterprises and not living up to the expectations that the customers have.
I'll give you yes, and I'll give you a couple of examples. So imagine if there was an automated system that was approving your mortgage application and that system just decided to decline your application for reasons that are not even explainable because this algorithm is a black boss.
Well, that is not very good. So you want to make sure that there's a way to protect against that.
Or imagine hate speech creeping into this chatbot that you're interfacing with. There's in a customer service scenario, how do you protect against those things? You can't do that if you haven't systematically thought about being responsible with these systems, and that is where that program comes in Handy.
I don't know how you feel about this, but it seems like people accept answers from lms in a way they didn't accept answers from Google. From Google in the early part of search. Do you feel that way?
I think yes, And I think there's also emerging research that people are nicer to AI than they are two people, which is interesting.
I think they like feel like they're like this relationship with this thing.
Something'll be pushing back on chat gip too.
Yeah, I'm often cross referencing, like going somewhere else and like checking out that.
This is why it is really important that we build trustworthy, worthy systems, and then we go out of our way to make sure that you can trust the response that are getting out of these systems.
This is why these programs are critical.
Anor like I think about right, So we Google and we do trust our responses and there are reviews, Like there's different ways of kind of checking on it.
What will be the gut check on llms? Will there be rating systems? Will there be like so, how does that progress?
Right?
So, one of the things that we are doing here at Cisco earlier in the year, we announced a product called AI Defense, and the idea is very simple. First, it starts with observing how people are using these applications, and a second, you can start to put some guardrails so you can start to sanitize the inputs that people are getting to these large legit models. You can also do the same thing in the other way around. So for example, you wanted to protect people from hate speech,
you can put those type of guardrails. You wanted to make sure that these models are not hallucinating and not giving you wrong answers, you can start to put guards around that. And so you will see a lot of these tools emerged and the models themselves are getting better with these type of productions.
But I think there's always going to be a.
Need to have tools that you can control as an organization that not only comply with the law of the land, but also with your have values, with your policies, and so it's important to have those types of tools.
Yeah, it's interesting. So I spent a lot of time with chatchipt, a lot of time with anthropics Claude. And you know one thing that I've noticed about Claude, and this is Sonnet for it says, well playing with it's like pull this up and says Claude can make mistakes, Please double check responses, whereas there's no, there's not necessarily that same disc that's after every answer with claud You don't have that same answer with chat GPT, with Quad, they won't generate images for you.
Do we have it with even like Google.
With Google's AI feature.
Like normal search?
Well normally, but I feel like that's just in our DNA as people who search for things like I feel like I feel like Internet culture, we are taught to be skeptical of what you find on the Internet. Yeah, we're not necessarily taught to be skeptical of the answers that these lms are giving us. Is that wrong?
Well, I think there are different companies who are taking different approaches. So do you just pointed out with what you saw on Claude? So I think there is a need to drive more of illness, that you should not blindly trust what you get out of these systems, and you should always have sort of a skeptical mindset for for these type of things. But companies who are putting these disclaimers, I think they are doing the right thing by making the making the user of challenges.
We're speaking with anurag Denngra, senior vice president and the general manager of the Enterprise Connectivity and Collaboration Group at Cisco.
So I'm just thinking, you know, our audience is a smart one. It's an investor audience. I mean when you think about you know, when they're looking at companies who are embracing AI and building out their infrastructure, what do they want to see from their companies to make sure that they are doing that well. So, like, what's your thinking about what businesses should be doing to create safe, secure AI infrastructure?
Right, So the first thing I'll say is you want to make sure that your infrastructure is ready to deploy AI at scale, because this is being the different Deploying AI could be the difference between you being a leader in your space and being irrelevant, and so it is very important to start with what do I need in terms.
Of network connectivity.
How do I have scalable networks that are not just ready for the traffic that you're running on them today, but the use case that AI is going to unlock tomorrow.
How do you secure it?
This is why all the announcements that we made here at Cisco I are really about AI, the secure networking products, scalable devices, security at the core of the network.
So that's number one.
Second, with applications like AI defens that I was talking about earlier, Yeah, the start to put guard rails around safe.
And secure use of AI.
And then if you are in the business of building applications, building products, then I would encourage every company to think about how you're gonna protect your systems. How are you gonna make sure that your systems are used for good. This is where programs like the Responsible I program come in interplay.
I feel like, especially with publicly health companies, will be like a whole new section, like when they either report or something right like that we will be dealing with to get an idea of how safe and secure their systems are.
Yeah, I think in some ways it's gonna be like security. Yeah, and you know, As these these challenges become more more well known, As companies become more aware, they're start to demand systems that can provide the visibility on what's happening with the use of these sools.
It's definitely a new world order, but really fascinating. An Eric, Thank you so much, covered so much ground. Thank you, Thank you, good luck for having good luck at the good luck at the Steve Rakdingras, Senior Vice President, General Manager, Enterprise Connectivity and Collaboration.
Here at Cisco Live twenty twenty five, you.
Are listening to the Bloomberg Business Weekdaily podcast. Catch us Live weekday afternoons from two to five pm Eastern. Listen on Applecarplay and Android Auto with the Bloomberg Business app, or watch us Live on YouTube.
We're highlighting some of our favorite conversations from Cisco Live twenty twenty five.
Now.
As President chief Product Officer at Cisco G two, Patel is responsible for charting the path forward for the company through all the changes and the growth of artificial intelligence. It's a lot.
We spoke to him at the conference about everything going on at Cisco, including why the event itself is so important.
Look, these are our customers.
There are partners and their companies who are considering buying Cisco technology, and one of the things that we want to make sure that we're very clear with is what are the core problems we're sol and what are the key innovations that we're making. And we want to make sure that we can provide that to them and have them talk to the technical experts so that they can actually learn about the products and you know, you know,
kind of pressure testimony they're here. We're in a really interesting place right now because we're in the next era of AI. So we've past two and a half years since chat GPT came about. We had this kind of seismic shift that happened in the industry, and we have these chatbots that could intelligently answer questions that we.
Had, and we're now going to this next.
Era where there's going to be agents that can conduct tasks and jobs fully autonomously on our behalf that initially they'll start working by themselves for a couple hours, and then over time they might be a couple couple of days and then a couple of months and a couple of quarters, and the possibility of the kinds of problems we can solve are so amazing, and we want to make sure that we provide the infrastructure to make that happen.
So are you talking about with the agentic solutions that you're talking about are you talking about in France specifically.
Correct.
So we've been working on, you know, working with people on making sure that we can provide the infrastructure for training for the past couple of years.
And that's where a lot of the that's where a lot of the spendspen in, a lot of the spend, a lot of the rewarding by shareholders of certain companies, no question.
Has been to date.
But now what you're starting to see is the demand signal for inferencing is going up quite a bit. And if you think about what's happening and inferencing, when you just had an interactive mode where you would ask a chat, bought a question, well what happened is you would have spikes of peaks that you would get on compute usage
and on data center usage. When you have an agent and that agent is fully autonomously and proactively working on something and they're just kind of cranking away and there's not just one agent, there's tons of agents working together. What you see is a sustained demand for inferencing capacity, and it's persistent inferencing that's needed. So your infrastructures kind of assumptions, the amount of netw bandwidth you need, the amount of you know, kind of safety and security that
needs to be in play is very, very different. And so right now we're working with a lot of the governments who want to make sure that they have sovereign data centers all throughout the world.
We are also working.
With That's a big demand push right now, isn't it huge?
Huge menbuss I was in the Middle East, Yeah, just a couple of weeks ago, and we announced partnerships with Humane and Saudi Arabia. We announced partnerships at G forty two at Stargate UAEE, And the reason for this is because there's an insatiable demand right now. If you think about what's the constraint in AI today, it's infrastructure and it's safety and security where people have to trust these systems otherwise they're not going to use them.
And so those are the two big constraints.
And of course the third one is a skills gap where you have to keep training people up. But the first two are pretty material and imagine if we had a world where infrastructure was not a constraint. Imagine all the problems we could solve that we're currently not able to solve because of the constraint of infrastructure.
But some of those constraints are serious to you too, And I think about you as you guys are ramping up and building out and providing what customers need, whether it's enough power to do it all, how are you guys kind of balancing that you're doing what you need to be doing, but making sure everything else is there to support it.
I mean, if you think about the large constraints, they compute their power because GPUs are power hungry, and then there's network bandwidth, and then there's safety and security, like those are the four kind of high level buckets. If you think about power, what we are trying to do is make sure that we can have the lowest amount of latency, the highest performance, and the most power efficiency because how long it takes for a packet to get
to the GPU to get processed right really matters. And if you have idle time on a GPU, that's actually not a good thing. You want to make sure that you're keeping the GPU utilization high and so we need to make sure that we have power efficiency. Every killer water power we save is a killer water power you can give to the GPU, and then we meet need to make sure that the backups keep flowing.
So capability and efficiency, you're constantly like balancing.
Constantly balancing it, right, Yeah, absolutely.
Because you've got to make sure it's efficient, otherwise it's not going to be financially going to make sense.
That's exactly right, because if you if you don't have the right level of power efficiency. But for firstly, all of the data centers that are getting built right now, you we are running out of data centers to build in certain.
Geographies because you don't have enough powers.
You actually tend to go to where the power is, and data centers are getting built where the power is.
So it's a very important commodity.
Is that true globally?
Yeah, it's absolutely.
The data set like running out of places to run data centers because of lack of power.
Yeah, you just need you need to be not too far from the power so that you're you know, the the entropy is less and how that works.
You're good, Carol. Enthusiasm between the two of you, that holistic questions that I want to make.
We said we can go as long with you two as you will give us.
Yes, that is as long as you want.
That is the answer. Can you talk a little bit about the relationship that you have with in video, Yes, because I think a lot of in best of trying to understand from a customer perspective, from just a relationship perspective, what can you tell us?
So we have enjoyed a very meaningful partnership with them for the past six years. You know, we six years ago we decided that we're going to use the nvidio chip sets and our video conferencing devices, and that made those devices very you know, kind of AI ready, and now what you see is a lot of kind of sophisticated AI capabilities are uploaded on the device itself, so that even if the device integrates for a third party like Microsoft Teams, they can benefit.
From the innovations that we've made. So that's been a great partnership.
What we did in the past few months, though, is we have doubled down on the partnership with data centers, and so we had announced that we have a AI factory and secure AI factory, which is basically what Nvidia has as a reference architecture, which is a suggested model mode of how companies should think about building out their infrastructure.
And for the very first time they had a non Nvidia.
Silicon provider that was actually part of that reference architecture, who was Cisco, And we were the first one of those that they actually incorporated into their architecture. Why is that important because customers look to nvidio to say, hey, what does your reference architecture look like? And now Cisco is part of that reference architecture. So that was the
first thing. But what we've also done and what we announced today was our safety and security capabilities are pretty sophisticated, and we want to make sure that we actually tie that with Nvidia's you know, kind of efforts around if you're building an open source model and an application with an Nemo framework, we can now validate that model with Cisco. And so we aren't just helping with Nvidia and building a secure AI factory with the infrastructure and a full stack, but we're also helping.
But how do you keep it secure and how do you keep it safe?
And so because one of the challenges that you have is if people don't trust their AI systems, they're not going to use them, and so we want to make sure that people are trusting them.
We're talking with G. Two Patel, President and chief Product officer at Cisco.
We're here at Cisco Live twenty twenty five. What's the productive conversation around So much focus is on, especially when it comes tot Nvidia in their demand for tips, the hyper scalers, how much they are doing to get an idea of where we are in the AI race. But I'm also you know, you talk to you folks, and it's a lot of on premise build out.
So what's the Yeah, are both okay?
So what's the mix in terms of how we see this continue to grow?
So I think the way that you will see this progress forward is there's three or four different models that will be out there. Firstly, one of the very counterintuitive things is hyperscalers are going to grow very rapidly, Private data centers are going to grow very rapidly, and neo clouds, which are actually AI specialists, are going to grow very opidly, and.
So our service providers.
So when there's you know, kind of a rising tide, all boats rise, and that's what's happening right now.
Now. What is happening that's unique with.
You know, data sovereignty and with you know kind of nationalistic behaviors that are kind of progressing from from a geopolitical standpoint, is there's more and more of an acceleration of demand of private data centers.
Yeah, we talked about this on prem on prem.
And specifically for infancing, I think for training.
That's interesting.
You know people might go to the cloud.
Brother, do you see that as being part of a government regulation, a sovereignty issue or is it more an individual on an individual company basis.
It's actually a combination of issues. It's you know, there's a sovereignty issue. There's also a data proximity issue. There might also be some companies who have enough scale where the scale itself might give them an advantage that they're like, hey, I don't need to go pay someone something. I might want to have some some of the workloads on my own premise. And so it's a combination of multiple issues. But what you are seeing is there's this kind of
upsurge of demand and acceleration. The good news for us is we help the largest of hyperscalers with intra Cluster communication. You know, on on the networking side, we are also working with neocloud providers and then and then we've had relationships with service providers for the past couple of decades, and now what you're starting to see is a resurgence of demand on the enterprise that we know really really well.
We've actually been working with enterprise data centers for the past forty years and now we are actually starting to see a massive resurgence because of AI inferencing demand and.
The agentic workflows.
When every company is starting to rethink their workflows, whether it be a BB a healthcare company, or a bank or a financial services or insurance company, every workflow is getting reimagined. Infrastructure is going to be built out everywhere, and we want to make sure that we are the common networking substrate and we are the common security and safety kind of provider across the board.
So G two.
In an environment where we constantly are talking about an uncertain environment because there's stuff coming at us even today as we kind of continue to watch the negotiations between US and Chinese officials over in London, how would you describe demand the CAPEX spend that you are seeing, whether it's the hyperscalers, whether it's those folks building their enterprise systems, like the different buckets that you play into.
The demand signal is stronger than an ever.
I've never seen anything, you know, stronger on the demand signal side, because you've you've got trillions of dollars that are going to be flowing in for these data center buildout capacities.
So the demand signal is very strong.
It's actually also happening globally, so it's not just happening in the US. You're starting to see demand in the Middle East, You're starting to see it in Europe, You're going to start to see it in Asia. And one of the things that's important for America is we want to make sure that American technologies are being utilized in all these data center buildouts because that's good for America
and that's good for the world. Frankly, and so you know, Cisco and Nvidia and open AI and others and all of our partners, I think a MD it's extremely important that we all work together to make sure that these these kind of data center buildouts happen with our technology.
You know, you're you're shocking me with the just the striking magnitude of these investments. It's what is what is not being Where's money not being spent as a result, as budgets get reallocated toward infrastructure, what's not getting money.
I think what you'll find is the.
In the short term, you actually have to take certain amount of dollars from existing projects and move them towards AI. And almost every company is doing that this and you know, I'm going to take twenty percent out for my existing projects move.
Them towards AI.
Over time, what you find is AI will get so much productivity that those twenty percent of the projects will get rehydrated. Because if you think about like, for example, we're one of the first partners with OpenAI on their codex project, which is the autonomous software Engineer. It's no longer where you just have an autocomplete where you can have an individual developer that gets productive.
You can actually now tell an agent or.
A group of agents, go solve a really hard problem for me and come back to me in a couple of weeks, and we'll be able to actually have a world that we are living in in relatively short you know, kind of time horizons where it'll be able to solve that problem and come back to you, and you'll still have the human in the loop that will be able to observe and that will be able to do things. But imagine what that does for the constraint we have
twenty seven thousand engineers. I've never felt more constrained in my life about the ideas I have compared to the resources I have to prosecute those ideas. And we will actually get to a point where we can unlock those ideas in a very different way.
What I was just gonna say. We could go on forever with G two, but apparently his team is telling us that you have to go speak there's even speaking gate.
Keep going.
Oh, he doesn't want to do it. He wants to hang out.
With us, and you guys an't too fun.
Hey, we're gonna get in trouble if we if we do that, well.
I mean I'm gonna go go.
I mean, I mean we're just kind of blown away every day, like what we are learning about this and how it's impacting our world. I mean, I don't know six months from now, a year from like, how should we be thinking about how dramatically this will impact everything?
I think, firstly, your imagination will be challenged quite a bit because humans have a very hard time thinking exponentially for a sustained period of time.
And this curve is exponential.
And when it's exponential, it's not happening in a single dimension, it's happening across multiple dimensions all at once. And so if you think about it six months or twelve months or eighteen months from now, no longer are you just going to be thinking about this as a more efficient way for getting an answer. But you will have companions and sidekicks and a staff of people that are digital workers that can get a job done that you would
have otherwise not had a time to do. And one important area to talk about is a lot of people worry about ISAI I going to take my job? I actually worry more about people that use AI really well.
There's a higher risk of them taking your job than AI taking your job, noted, you know.
And the other piece that I worry I also think about is if you actually have there's no one that I know that comes home every day and says, you know what, I've gotten everything done on my checklist, right, And so the reality is is when you there's eighty percent of the work that we do probably doesn't get checked off because we don't have time to do it.
Now AI agents will start doing that. And then the final category that's the most exciting is it's AI will be able to do things that we were not able to do because it'll have original insights that don't exist in the human corpus of knowledge.
And that's the area of AI, but I think is the most under hyped. G Two, we got to leave it there.
Your crew is about to come on set, and where.
You're to go and just where you know.
I have a companion inside can but I'm ready for my AI companion.
Inside than product officer as let's go here at Cisco Live twenty twenty five.
This is the Bloomberg Business Week daily podcast, available on Apple, Spotify, and anywhere else you get your podcasts. 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 terminal
