Bloomberg Audio Studios, podcasts, radio news. Bloomberg Tech is live from the heart of Silicon Valley with ed Larbow in San Francisco.
This is Bloomberg Tech coming up.
Microns sows after its AI fueled forecast shatters all street projections, plus Qualcom jumps after posting a strong annual sales forecast. We are joined by Qualcom CEO Cristio armon live and Apple raises prices on some hardware and extreme measure in direct response to the memory and storage crunch. And it is memory that is still at the heart of what's happening on Bloomberg Tech.
A super bullish sales outlook.
From Micron, which is currently on track for its biggest jump in a month, but at the open on track for its biggest jumpsince twenty eleven. We will get to the details. There are two other big stories, one directly linked. Apple is raising prices on some hardware citing the memory and storage crunch and storage prices a Qualcom, fresh off its invested day, has a lot of news and extended conversation with the CEO. It is still memory and it
is Micron, and here's what you need to know. Micron didn't just beat expectations, reset them higher, forecasting as much as fifty one billion dollars in fourth quarter revenue, well abuff all Street estimates. More importantly, Management says AI driven memory shortages could last beyond twenty twenty seven, keeping memory prices higher for the foreseeable future, even as Micron boosts
it's manufacturing capacity. And that helps explain our other top story, Apple raising prices on some of their hardware offerings as the cost of key components memory stays high. Bloombers Mark German has the story, and you write that this is an extreme set of measures that Apple is taking. Where are the price hikes? What is Apple saying about the reasons why? And this is all about memory at the end of the day.
Yeah, these price hikes are largely unprecedented in Apple's modern history. There hasn't been a time where they've hyped the prices across several different products, across much of its product lineup. This is the iPad, this is the Mac this is the Home pod, this is the Apple TV, this is the Vision Pro. It's essentially everything outside of some accessories and of course the iPhone, the most important product. The AirPods and the Apple Watch. So this is a very
big deal. I think that the price increases are pretty significant, but they're not necessarily going to be a big deterrent to customers, particularly on the Mac side. The iPad makes me a little uneasy given all the competition in the tablet market and upgrade cycles, and then the Apple TV and home pod prices are pretty significant, and given that there's so many cheaper alternatives from Amazon, Google and elsewhere, that could be a little dangerous for Apple in that market.
Mark what struck me reading your report is how detailed a pool was in its explanation. Maybe you think that they weren't as detailed as perhaps they often are, But they're trying to tie this to what's happening in AI data centers, and they're trying to explain that for the longest time possible they held off on doing this.
They have no choice. Now.
Yeah, let's remember other companies in the consumer electronics space. Apple competitors hiked prices significantly in the prior months. Apple, you know it's a good thing. This time is fairly late to these price increases. About Apple's Apple, They're the most popular consumer electronics company. Everyone's buying these products, and so this is the biggest of all big deals when
it comes to price increases, that being Apple. The most interesting thing to me in Apple's statement wasn't that they're blaming AI data centers and memory.
We know that.
It's that they seriously hinted that there will be more price increases to other products. They said that these are the price increases that are happening today, leaving the door open. I would say to iPhone and Apple watch price increases in September when announced the new models, Boobo's.
Mark German, who leads our coverage of consumer technology, thank you very much. Yesterday, Microns stunned with quarterly estimates that absolutely shattered will streets expectations, signaling that the AI infrastructure boom remains strong again. Micron shares are on track for their best day in one month, but at the open, on track for their best day since twenty eleven. Supply is tight, it is going to stay tight. There is no line of sight to supply improving, and memory prices
will remain high. That's the story. Chief market strategist Carosh Life of BEMO Wealth Management is with us for more and you have a very interesting, almost historical look at what will happen next, and that is that historically, for decades, tech had been a pretty reliable deflationary force. Mark just outlined, actually more price hikes could come on that higher end Apple hardware. What happens as a result.
Well, I think Marcus are really trying to struggle with the fact that it looks like we've got an economic regime change, if you will, underway, and it starts as Mark talked about, it starts with the data centers and the build there. But it also you have to think through what life as a semiconductor company has been like because they go decades with very big pea contrough and so they want some reassurance from these companies before they
start expanding capacity. They need reassurance that that demand is going to be there. They get some of the reassurance they're signing longer term contracts. There's really slowing expanding capacity, but in the process of really definitely targeting their production much the same way that or that many of the semiconductor companies did during the pandemic, and they diverted it away from auto companies and put it towards the PC
manufacturers and the AI creators. If you will, So they're being very strategic in how they're doing, and the industry overall is trying to watch its margins. I think what investors have to deal with is, you know, there was this presumption early on that higher prices and more over here just came because someone was shaking a money tree over there. But it's a pretty zero sum game in terms of if you're a company, you've got you're paying
in one category and giving up in another. And so Apple sending that message that they've held on for a very long time, not past pricing increases along. That's actually pretty indicative across the economy where a lot of companies have done that. They've held it as long as they can, and so it leaves us the consumer and us as investors to be very discerning about where we're going to see it pull through, where we're willing to pay up.
I think Apple is betting that people are willing to pay up for those those specific products.
What Micron said last night was remarkable. The number is fifty one billion dollars up to in the fiscal fourth quarter when the streets saw forty three.
So there's just a.
Simple element of outperformance relative to expectations, but the admission that they have no line of sights to the supply situation improving and therefore prices remaining high. How does the market model for that with everyone else in the equation, the capex deployers, the hyperscalers.
I think it's it's difficult to model because basically what we're doing is as an economy, we're trying to rewire the whole economy very rapidly and make up for decades of under investment and infrastructure because we haven't invested in the energy grid. We haven't needed to because the demand wasn't up, and now we're asking demand to go up many multiples of where it has been historically, and so markets have to adjust to the fact that you'll probably
have higher endemic core inflation, if you will. But you're seeing from the fixed income markets that they're actually pretty tolerant of that because of the fact that it's coming from steady demand. It's not coming from workforce demanding higher wages, but it's coming from growth we're putting in the economy, and economics one on one would tell you that in the long run, and we invest in that in the capital infrastructure, it does tee us up for an era
of more productive growth. But there's this interim period that it's going to take us a while. We're probably talking later in the decade or early next decade when we finally see the big deflationary impacts of this come through. Because in the meantime, we all want access to it, we need data centers to get access to it, and there's that bottleneck in the middle that's going to need sorting.
Carol, just really quick before we let you go, Micron. The funny thing about it in other memory names is that they're trading it like a pretty reasonable multiple.
Well, they're super reasonable, and like I've heard this morning, like you've seen even within VIDIA two. Every time they report, people look at the stock price, you know, simplistically on the front. They look at the stock price and forget to put the stock price relative to the earnings. And the valuations have been coming down on these names actually for quite some time because the earnings are so unbelievably strong.
It's a stock that literally is priced oney one hundred and ninety dollars a share. But yeah, I looked at twelve month fort EPs. Carrisflyfe for BIMO Wealth Management's great to have you back on the show, Thank you very much. Coming up, we're going to stick with Semiconductors. A conversation with Qualcom CEO Christiano Amon on data centers, deeper push into data centers, M and A. Qualcom's been busy and they gave some pretty interesting numbers for the future. That
conversation's next. This is Bloomberg Tech. Okay shares of quodcorm We're up around four percent. They had opened up more than eleven percent. The company unveiled ambitious long term growth targets at its investor day in New York. Quilecom is projecting more than fifteen billion dollars in annual AI data center revenue by fiscal twenty twenty nine, and forty billion dollars in annual sales from businesses outside of smartphones, double
the target that it is set two years ago. Joining us now is caulled Com CEO christianom On and Bloomberg's the clothes anchor remained Boss Stick and Christian. It's great to have you back on Bloomberg Tech. And here with remain and I if we could, can we just start on actually the near term because the five billion dollar data center number for fiscal twenty seven is really interesting to a lot of people. I think they really want to understand the composition of it right. Data center can
mean a lot of things. Where is Quark I'm going to move most quickly.
Yeah, very good. Happy to talk to both of you guys. Look this exciting time, especially because we kind of upgraded our non handset revenue as we've been diversifying the company from twenty two to forty and enter the data center market. And five billion in fiscal twenty seven. It's a high confidence number that we have based on customer engagements we
have right now. Most of their revenue is from a custom ASA engagement that expanded since the acquisition of Alpha Wave and started to see in their revenue some of this innovative technology we came up for celebrators. It doesn't require HBM memory is our high bandwidth compute and as we keep executing on those customers, we're excited to see the revenue going to about fifteen billion in fiscal twenty nine.
I am curious Christiana when it comes to those longer term targets and the bridge between the short term. So what you have put out there for twenty twenty seven. Is there any sense I know you haven't named the customers that get you over that bridge here, but is that going to fill the whole of that five billion based on the one billion that we already know is out there?
Look the five billion, as I said it, I'll give you the composition. I think what I can say right now, I think, as it expects, a lot of those contracts are governed by non disclosure. We have one large hyperscalar customer in the United States, one large hyperscalar customer in China, and then we have some other engagements that we had on a data center that is all into the five billion dollars. So the five billion dollar is it's a very high confident forecast to which we have the customers,
the capacity, the memory allocation and everything. As we think about the engagement we have today, and also we talk about a contract that we have with Meta also for two generations of the CPU, especially as we position ourselves for the GENTECHCPU opportunities, that those engagements are into the fifteen billion dollar number for Qualcom as an eventrate, those are great numbers, but really when you look at the scale,
what's happening in the data center. You imagine, as we execute and we show our differentiation, they could be upside. But we're just you know, being focused on the engagements we have today.
I do have to ask. So an investor seem pretty pleased is what they heard yesterday about your articulating this longer term data center strategy. But I'm curious. I mean, you famously exited the data center business back in twenty seventeen, twenty eighteen, and I know there were a lot of direct a lot of reasons for that that went beyond your operational control. But there was a lot of competition then.
There's probably even more competition today. Why do you think you can get it right now compared to what it was eight years ago.
Yeah, look the circumstances. I think when we started on the CPU, actually, I know people talk about data center as a general term, but we actually we start a long time with an ARM compatible eyes at CPU. Very early, there was a big software barrier. This is before AI, and you know at the time I did not you know, the market wasn't ready for it only until later. So it was kind of the probably the right decision for
qualk at a time. This is a different conversation. It's a different data center right now, this is about AI. But it's a valid question, right so I will give you the same answer that I gave when we entered the automotive business, when we entered the industrial business, and everything that we have been doing to diversify the company. Those are fast moving markets now changing with technology. So what we really focus is not what the data center is doing today, but what the data center is going
to be doing tomorrow. The data center is now moving to agentic. That's why you see CPU companies on the rise, you see demand on the rise. The data center is being disaggregated. Is not one single GPU that does everything. You have different harder and we're coming out with a comprehensive portfolio. We have. We have a CPU, we have a new memory solution that is not a memory from the memory provider, has actually been designed by Qualcom that
doesn't require HBM. We have a different accelerator for the desegregated computes, and we have our semiconductor scale which is good for customisings. So that's what we're doing. And you know, if technology matters and you have fast innovations cycles, there's room for.
Qualcom Bloomberg Tech Live on Bloomberg Television and Bloomberg Radio all around the world, and we're speaking to Qualcom's Cristiano Amon and a big part of Qualcomm's story is custom silicon now A six. You mentioned it a moment ago, Cristiano. Bloomberg reported on May twenty sixth that byte Dance was one of those HYPERSCALA companies, and at the time Qualcom
didn't comment how is that going to work? You know, there is a distinction between an A six program and export of accelerator, which is your own name on it, within the current confines of the US export controls that are in place.
Well, was still not commenting on it, but I'll tell you that like everybody else that has a significant portion of their data center revenues, also with customers in China, there is a there's a very clear set of regulations and specifications that you have to follow and you should expect to Qualcom it's complying with those and I think there's different types of products that are distant for China and those are the ones that we are shipping to our channel customers.
Thank you.
The data center business is multifaceted, so ASEX your own accelerators and now CPU. But the market's in a place where in video or AMD on an annual break basis bring a new generation of technology.
Right.
And so the question that a lot of people pose for you, Christiano is by the time the accelerators come to market and the CPUs in twenty twenty eight and customers deploy them, will they not be obsolete? Is it not too late to kind of get into those markets when you look at the peers and how they deploy.
Absolutely not. Look, I get this question all the time, and I think if that was true, Qualcom will not diversify from phones and enter the automotive industry, into the industrial, into the robotics, into the broadband. And I actually believe it's exactly what I've said before. I'm going to give you. I'm going to give you a simple answer. There are four vectors of differentiation accelerator. First of all, we know
that power matter. If you look at the demand that exists in the data center right now, is a demand that going like this and the energy available is like this, So that creates It's perfect for Qualcom because what we do is very power efficient. We design our technology always thinking there's going to be a battery on the other side, and we're going to have efficient and power consumption. That's
one vector of differentiation. The other one is what I said, we design for the future which does not have the memory constraint that you have today. We don't have to move a lot of data between the accelerator to memory, and that reduces thermal, reduces power consumption, and re increases the memory bandwidth. And we don't have to use expensive HBM. We have the ability to do this at a much
lower cost. Number three is we are coming in with very high density of transistors with our accelerator, and.
We're not a small company.
I think we we ship forty billion components a year. We do more than already new chips per year on leading notes, and I think we have the ability to execute. And if the data center is changing because of AGENTIC, and the data center is changing because of this aggregation, there's an entry point for Qualcom.
I am curious, Krishano, particularly with Meta and Dragonfly, how confident are you that Meta will still be there in twenty twenty eight as a partner, and I asked that only not so much from the Qualcom side. But Meta has sort of moved around a little bit with some of its strategies and how it's approaching AI. Is the agreement that you have with them? Is that basically firm look two parts. Yes, we have a contract with Meta. That's why we talk about them as a customer for
two generations and it's very significant for our CPU. We have a very unique CPU if you look at some of the metrics we're provided, and I think we have a track record that have a good CPU. The second part of the question, I will never bet against Meta. You know they have they have incredible first party applications, they have a lot of data, they have a lot of data from consumers.
They're well positioned. We are on the early innings of AI. Anybody that believes right now that whoever the players are on an I AI, this is what is already defined. This is a this is a decade long and I think if what we learned something about the Internet is your technologe is going to move very fast and the players are going to change. I will not bet against Meta.
Christiana, can I just jump in on Meta really quickly? What I find really interesting about it is the relationship was already there right snap Dragon in quest. I spent a lot of time down in Palo Alto on the Irian project kind of looking at what's under the hood. How much did that existing relationship on the consumer electronic side get the win on the data center side.
It's one hundred percent. Look at the end of the day, this is not this is not one time relationship. We have built over a long period of time, a strategic relationship and what happened Also have those relationship in place and we have more products, you have trust, you'll have the capability of execution. And by the way, this is also true what happened on customising. We bought Alpha Wave. They have some engagements, but the other part of Qualcom.
Qualcom has brought a relationship with some of those companies and I think that just scale and that's exactly what you're seeing.
Well, can I ask you a little bit about that? You mentioned Alpha Wave. You've obviously you have Dentana, you have the Modular deal. You made a lot of acquisitions in a relatively short period of time. I guess my question is twofold a should we expect more anytime soon? And based on the acquisitions that you've done, have you actually already sort of integrated that into some sort of sellable product that's already sort of out there as.
One, yeah, one hundred percent. Look if you look at our track record, right so, especially as I think about the diversification of the company. First was automotive. We bought a River and that is now the stack that you see on BMW for a highway autopilot. We bought Nuvia. You see now our CPU showing up not only in PC's and phones and automotive, but now in the data center. We bought Alpha Wave, and you see us now into the connectivity business and the custom AASAIC business, and the
Module acquisition is incredibly exciting. I actually don't think the industry investors really understand what that acquisition is. I think my prediction is people are going to read the three or four or five times and eventually they're going to get it, which is an industry friendly alternative that we're going to make it available upon close to everybody. So all of those things.
That Module acquisition you're Chris Latry is sticking around absolutely.
That's why you can see, if anything, if anything, what you should see about this acquisition. By being an all stock deal, it means that the founders are now fully vested on the future of Qualcom. And I think there's another thing that investors did not understand. They asked me the question why did this size acquisition do in all stock is because they're all now committed to the future of Qualcom when we are connected together, so all those
things as part of us doing that. And to your question is there are going to be some other acquisitions, but as we have been very disciplined as we're building industrial some of the acquisitions we did are Thewena with thirty three million. Developers would focus on industrial edging posts. Now you see us doing data center. Future acquisitions are focused on data center and robotics, which is another area of investment for Qualcom.
Christiano.
Smartphone and AIPC is still there for you, right and you'll be aware of the world that we woke up to this morning in memory with Apple raising crisis as well as your outlook for the PC and smartphone markets changed or shifted it all for this year next year.
That's a great question. I think you have to look at those markets very short term and you need to look at this market. So what's happening in the midterm for the very short term. I think the whole world is dealing with this emory situation right now, which is not great. And I think I respect the companies kind of make choice of the different industries. But then what happened is you don't have to demand the problem and
mobile NPC you have the memory availability and prices. I think what you saw the message from Apple, So we even expect when we think about, for example, the Android market in China, I think when you think about year over year, when you look at twenty seven, expect it to be flat to probably a down bias. And that has all to do not with demand but with the memory situation.
But the other parts more interesting.
There's a lot of clarity right now because of agents, how this market is going to change with AI very very clear, and I think we see a lot of those phones now becoming the customers of the phone as the human and the agents, and that is changing the architecture of the devices.
CEO Christiano Ammon and Bloomberg's remain vostic. Fantastic conversation. Thank you both very much. Indeed, welcome back to Bloomberg Tech. I'm going to take a look at today's big number, which actually may be our smallest big number. Yes, sub one nanometer. IBM has unveiled what it calls the world's first sub one nanometer chip technology, squeezing nearly one hundred billion transistors onto a chip about the size of a fingernail. For scale or reference, a human hair is about eighty
thousand to one hundred thousand animeters wide. We're talking sub one nanometer. Let's get to Micron. Micron's earnings and the memory environment are our top story right now. Micron's up twelve percent. It's on track for its biggest jump in a month. Earlier in the session have been much higher. Suppli is tight, prices will remain high. Bloomberg Intelligence writing, these earnings reflect AI related demand, but don't quite capture
the structural changes to the industry. Let's get out to Bloomberg Intelligence analyst Jake Silverman the headline there or the title the memory up cycle. For a few days now, you and I have been talking about is this a massive departure from the memory market's historical norm. You seem to be writing now that it is.
Yeah, it definitely is too many extent, to quite an extent, because in the past, what we've really seen is you see contracts on a monthly or quarterly basis, and so the fact that we're now seeing contracts extending out to twenty thirty in many cases, it just shows that there's a structural difference in the way that customers are thinking
about procuring their memory. And a lot of this is just driven by the substantial price increases that we've seen, you know, with memory probably increasing two hundred to three hundred percent this year for dram and Land. You know, the hyperscalers want to make sure that not only do they have the supply that they need, but they want to be able to make sure that they have a good idea of the range of pricing that they can expect as they're building out these AI data centers.
Jake, I'm about to have a really important conversation on software, so we should probably remind the audience, like what's happening in the background, and that is massive demand for memory chips in AI data centers, either for training or for running AI workloads.
Just explain the bigger picture.
Yeah, so when you're thinking about building out an AI data center, you need all of these GPUs. I think in video really clearly the dominant player on training. They have two hundred and eighty eight gigabytes of HBM per GPU in their NVL seventy two racks and whatever other
rack configuration that they have. But yeah, actually, interestingly enough, you've seen Google with their TPU going up from one hundred and ninety two gigabytes to two hundred and sixteen two hundred eighty eight for different skews, Trainium clearly increasing the amount of HBM. So it's not just in video, it's the entire ecosystem just requires more HBM. And then we're not even getting into the system memory that you need.
The LP DDR, the DDR that.
Supports all sorts of different functions for AI demand and that extends beyond training into inferencing, and we haven't even touched on ergentic AI. And so if you start to see a proliferation or an inflection of a lot of these different AI tools, it's going to require a memory as one of the foundational background enabling technologies.
Bloomberg Intelligence is Jake Silverman with the React. Thank you very much. Indeed, let's talk software. But it's annual config conference in San Francisco. Collaborative design firm FIGMA just unveiled a platform overhaul for the AI era, transforming its workspace into what it calls an intelligent canvas for full stack digital creation. It allows teams to build animations, integrate AI agents,
and write native code all in one place. Delight to say that joining us is Figma co founder and CEO Dylan Field back on the show.
It's good to see you. Good to see you too, Thanks for having me.
The idea is that you want your audience, which is developers and also customers, just to be able to do more in one place.
Designers, yeah, product designers and designers of all types as well as those they work with.
What's the breakthrough here? What have you done from the technology side that's allowed for that?
Absolutely, So, rather than having only vectors and images on the canvas, we've also added in all sorts of other things. For example, now you can have code layers where you're able to actually have a representation of code that's running
and interactive on the canvas alongside your designer presentations. Plus we've made it so that you can generate with the Figma Agent plugins that let your team do reusable workflows, run skills, and we're very excited about the generative effects and shaders that we've added shaders let you do advanced graphics work on the canvas. And finally we unveiled Figma Motion Figmotion and I'm really excited about and our community is honestly just over the moon.
Is it something you asketful? Sorry? Was is it something you asked for?
By the designers, they basically say like, we want this, can you building.
Since we've started? Okay, they've asked what's the impetus for it? Yeah?
So I think that if you zoom out right now, we're in a place where AI is trained on indistbution data and the center of the distribution is your average.
That's what's it's trained on.
And when you're prompting and you're getting outputs, it's so often that you're getting this average output and people can smell that they are almost it's just like AI writing. You ever get sort of triggered and a little bit like is that AI when you're reading something.
I think it's attributable too, yes, because they're not all equal, or just.
If you think something's been AI written in the first place, and the same thing is true for design. I think that as people are trying to stand out with their businesses be bold, have a point of view that will be rewarded, and only humans can really raise the ceiling here.
They can use AI to be more effective, but humans bring the point of view and they actually are able to use design as a differentiator to make it so that their product, their software, their business, their brand, their marketing, they're advertising stands out breaks to the noiseness increasingly wild attention economy we find ourselves in, which will only be more flooded by AIS.
That's quite big picture of stuff.
I mean, what I wanted to ask you is that the technology underpinning all of this. If you remember back to Open AI dead Day twenty twenty five, you and I had a conversation and your stock was going up a lot because you've got a name check on stage.
You telling me that and you know by that.
But in this case, let's take the coding layer for example, is it something you've built yourself or your model agnostic or coding platform agnostic.
Now we get to the net results.
We've always taken a modular approach on what models we can use, and so we can swap models in out whenever we like, on behalf of the user or at the user's request, and what we do is basically, you can take your design and just move it over to a code layer and it'll automatically convert it.
You can then make that interactive.
You can also in the future imagine a world where you're prototyping or even going to production with code Enfigma and by being part of that token flow, we're very excited about what that can mean for our users and their abilities to really impact the platform because as code is commoditizing, what we're seeing is that the value is moving up the stack to design. Designers are the ones that we think in the future will be building all the software so.
It can equal revenues.
As you know, a lot of your users watch the show and I always try to get their input what they'd ask you.
I have one audience question and I'm just going to read it to you.
Is there any plan to make Figma a platform for brand integrate using MCP? And so this person says, for instance, I may use a different application to build my.
Website, etc.
But my AI agent needs to access what he calls his ground truth, his own IP, his own branding, logos, artwork, et cetera.
Not only that, but also we want to do even more when it comes to your digital asset production workflows. So as you're figuring out ways to create systems and workflows to do advertising, to do media generation, one of our bets is Weave and Phigma. Weave is a platform that is able to compose models and workflows so that you can still have that human touch, that craft and really shape the output of models like Clay, and through
that you can get to extraordinary results. You can then run those workflows through an API and pull them into whatever system you like. But we're very excited to be that system of record, not only for the brand assets and design system, but also for these workflows as well.
Big the co founder and CEO Dylan Field back on Bloomberg Tech, Thank you very much.
Cue.
There's a new threat to tech stocks, and it's not what you'd expect. Public anger towards AI has Wall Street strategist warning it could spurt anti AI policies, creating another challenge for markets already grappling with a shaky rally in tech stocks, marked both by both rapid growth but also sharp volatility.
Boy did we learn that lesson this morning.
Bloomberg's Matthew Griffin has more really interesting read. We've done a little bit more on the show in the past six to twelve months about the kind of pr crisis that AI is having with Americans. At least now this is translating to the desks of Wall Street strategists who are saying something could happen as a result, just explain the report.
Well ed, thanks for having me on. And like you said, strategists are starting to tell their clients pay attention to this polling data, pay attention to what's happening at the local level, because again, Americans, they are very unhappy about the idea that AI might take their jobs. They're very unhappy at the prospect that data centers might be pushing
up electricity prices. And what strategists think is there is a chance that's going at some point to translate into policy that slows the rate of the AI boom, slows profit growth for the hyperscalers, for the chip makers, you name it. It's things like a pause on data center construction. And you look at a market where the number one question is what's the return on this investment? That could be a pretty big risk eventually.
So that's why I want it's going next is this hardware related, like there is a ban on building data centers, or is the concern that there is some kind of heavy regulatory intervention on the software side.
Well, if you look at sort of the worst case scenario that Wall Street's talking about, there is an example of each One would be a national moratorium or pause on data center construction, and the other would be some kind of tax targeting big tech companies or targeting AI producers. Now, we have seen progressive politicians members of Congress pushing for these policies already. We aren't there yet, but we are seeing them at the local and at the state levels.
There are a bunch of cities and counties that have paused data center construction. There's actually a moratorium sitting on the governor's desk here in New York awaiting her signature or her veto. And we've seen governors in Ohio and Illinois pausing tax breaks for data centers. So we're starting to see this kind of thing percolating.
Biomas Matthew Griffin with a really well read story on the Bloomberg terminal and dot com. Thank you so much. Let's stay in the realm of the AI build out. According to a new report, AI revenue may have reached a tipping point. Analysis from exponential View shows global AI sales excluding China reached twenty five billion dollars in the first quarter of this year, preciation from data center and chips investments. Well, that's expected to be just twenty one
billion dollars. So is the AI build out finally paying off? Azeemas are founder of Exponential View and an aufer on that report, as well as a Bloomberg contributor joins us. Now we've talked about this for so long, the machine where you know the numbers that are going in capital expenditures, it's just been very hard to find the numbers that are coming out the other side.
You seem to have found them.
Well, thank you for having me on your show, Ed, and I thought it was really important that we do find them. The supply side is well understood. How many picks, shovels, chips, data centers are being built and how much does that cost. Piecing together what's happening on the demand side is much harder.
The biggest AI companies are privately held, they don't need to disclose, and the hyperscalers, the big three AI is only a small subsegment of a subsegment of their businesses, and so they don't really disclose, but there are enough clues.
And for the last six months, my team and I have been looking at every public disclosure that we can up and down the supply chain, from the hyperscalers, from well sourced leaks, to reconstruct the AI economy and to reconstruct specifically, what is the top revenue coming in from customers, enterprises and consumers without any double counting, And that's where what we publish today, and that's the twenty five billion dollars in Q one twenty twenty six that you referred to,
and I think it's really critical this lifts the fog of war.
You can also look forward, right, so you're exactly right. I mentioned a moment in time and that was that period gone. But if you looked I guess on an annualized revenue run rate basis, you can say per the report that actually there is continued growth accelerating growth in those revenues too.
Absolutely.
I mean we view the if you annualize the current month, we're at one hundred and seventy billion dollars of annuallyzed revenues without any double counting that represents a two hundred percent growth rate on the previous point, this time twelve months ago. But the current growth rate is still at
around two hundred percent. And what's really remarkable, ed is that you would expect that growth rate to start to slow down, and in fact, we were caught out by this because we had in our modeling from the start of this year predicted it would slow down, and what turned out to happen was that Anthropic went on a real tear and kept the growth rate.
High, which you're showing the chart that goes back to what I said at the top, the point being quarterly revenue pace and current number is now exceeding the depreciation on the hardware side. I'm assuming it is em that you're a great student of history, just as I am.
And what's so interesting is like you can go back and say, well, what happened at the advent of the smartphone, or even of cloud computing the Internet if you want, where does what's happening AI on those real revenue stack up against those prior moments in technology history.
Those moments are epic. They were some of the fastest growing business waves we've seen in the United States worldwide. Ever, in history, AI real revenues is growing roughly three times faster than the Internet grew, roughly three times faster than mobile grew. So this is a really fast in demand product that enterprises across America and consumers around the world are really buying into.
At the moment, a z in real quick to end.
What happens next, Well, what happens next is we see more of main Street trying to get this right.
We see more competition. I think people are concerned with the price of some of the proprietary models. We'll see large enterprises rotate towards open source, which provides better value for money. But I do think that as enterprises get better at doing this, as main Street understands how to use AI, they will double down there.
Zeemers are founder of Exponential View of Bloomberg Contributes. It has been really great to have you on Bloomberg Tech such command of the data.
I really appreciate it. Thank you very much.
Now coming up on the show, Google's AI efforts have taken a hit with the departure of even more top talent. It's been a bit of a wave of late. We'll have the details next This is Bloomberg Tech.
It's time now for talking tech. I'man Hira honand first Up. The AI Boom is making its way onto NBA floors. Cloud infrastructure provider Iron is taking over the Golden State Warriors Jersey patch starting next season, betting the partnership will raise its profile in the Bay areas booming AI ecosystem.
The deal also makes Iron the team's official AI cloud partner. Plus, Anthropic is accusing Ali Baba of accusing nearly twenty five thousand fake accounts to illicitly access its claud ai model, and a letter to US senators and White House officials, Anthropic called it the biggest ever yet by a Chinese company to piggyback on the work of leading USAI labs. And Amazon is doubling down on India's AI boom, announcing another thirteen billion dollars to expand its cloud and AI
infrastructure in the country, including in Mumbai. The announcement comes as CEO Andy Jasse visits India, where he's meeting with Prime Minister in rand Remodi. With a new investment, Amazon's total spending in India from twenty ten through twenty thirty is expected to top eighty eight billion.
Dollars ed thank you, I Hira, two top AI researchers leaving Google Foranthropic. These are just the latest departures in recent days rattling Google's AI ambitions. Bloomberg's Google reported g you Love broke the story and joins us with the details, and the story was so impactful it was shared across industry. Let's start with those kind of two key figures are they and why it's being seen as such a blow to Google's AI efforts?
Thank you, ed so.
The two senior researchers who are moving from Google to Anthropic are Jonas Adler and Alexander Pritzel, and both were viewed as key contributors to Gemini, which is Google's flagship AI model.
And also highly regarded by their peers.
Absolutely, yes, yes, their peers were sad to see them go. Jonas Adler was playing a key role in AI coding and Alexander Pritzel was involved on the pre training side, which is where models are kind of trained on large quantities of data. And I think, especially following the news last week of Nome Shazir moving to Open AI and John Jumper, a Nobel laureate heading to anthropic. It has sparked concerns about who else.
May be on the way out.
We've been talking about that on the show for that extended period of time, right, and it's played out in the stock you know, Alphabet, the parent of Google's shares have been under pressure.
What's the bigger picture?
Like this is a mass wave of departures and so Gemini is in trouble or is it not as severe as that?
To be clear, those two departures, you know, the are the only you know names that we've reported on beyond John jenberin Norm Shazier. We're talking about a small number of people. However, the AI talent pool in terms of the number of people who really can move the needle on these models is relatively small, and so individual moves, you know, can matter. Noodle stresses, of course that his bench of talent.
Is very deep.
Yes, thank you, and you know engage with us on that story. It was a very detailed report. Is there anything else in there that we haven't gotzi that you think you know, the world needs to know about.
We got a little bit more color about what happened around the time of Nome Shazier's departure. Computing power is a precious resource in Goodle. They have a lot of computing power, but they also have a lot of mouths to feed with cloud clients research their popular products. And so we learned that shortly before Nom Shazir announced his departure, the computing power for one of his projects was shifted
to a larger pre training team in London. It seems it was an attempt to boost collaboration, but we're still piecing each other. But compute can be very political within Goodle.
Bloomberg Studio Love who broke a big story. I highly recommend you go and read it. So that does it for this edition of Bloomberg Tech. This is what markets look like. Those are the stories. And as that one hundred modestly higher micron the most severely bullish outlook for memory in a supply constrained environment and price is high, and then Apple raising prices on some hardware because of that memory situation. Also recap the qualcom CEO conversation on
the pod. You know where to find it. This is Bloomberg Tech