Bloomberg Audio Studios, Podcasts, radio news. Bloomberg Tech is live from coast to coast with Caroline Hide in New York and ever Low in sentences.
Go this is Bloomberg Tech coming up.
Bloomberg News pulls back the curtains on Apple's revamped Siri design ahead of its WWDC debut.
Plus Snowflake jumps the most It's twenty twenty after the software maker gave a stronger outlook and sign a six billion dollar deal with Amazon.
We'll hear from the CEO.
And Meta introduces paid chatbot subscriptions to help offset its AI infrastructure costs.
First we look at infrastructure and how actually an infrastructure deal with Amazon is helping summon the narrative around Snowflake. Extraordinary move, the biggest jump in the stock since twenty twenty. We're looking at thirty four percent gain after a thirty four percent in product revenue for this company. It was a beat, it was a raise, and there's real acceleration in people using their AI coding tool. In particular, twenty two billion dollars added in a market cap so much to digest today.
Ed a lot of news headlines this morning.
One coming from the information that Microsoft next week is going to introduce a coding model, a model focused on coding, a marketplace we've covered so much of late the stock off session highs that we saw gains three and a half almost four percent, So the market taking it seriously.
If we hear more, will give you more.
Well, and now we can give you more on what to expect with Apple's Siri overhaul, because it's going to take center stage in the company's next major software update, and Bloomberg News has details on what to expect ahead of the WWDC debut.
Now, these Bloomberg.
Created illustrations, they offer a look at the revamped Sery interface, including a new chatbot style app that the images that you're currently looking at they're based on information viewed by Bloomberg and people with knowledge of the company's plans.
Well those plans.
It's Bloomberg Consumer Tech and Apple Managing editor Mark German.
So the look the feel is really building.
On some of the updates we've had in the past and improving them.
Well, I'll just say this, this is an incredibly exciting moment for Apple.
I think consumers should be pumped.
What Apple is doing here is they've seen everything that Open Ai and Google and Anthropic have done. They believe that AI has a place at the center of its products, and they're finally going to implement that.
So this is a really big deal for consumers.
People have been clamoring for a version of Siri that works properly for the better part of fifteen years, and my strong belief as we are finally going to get this fall So I see this as a major development and accomplishment for Apple in its quest to try to bring AI to the masses, taking a slightly different tack than their rivals in the space.
There's two parts to this, as you report it. There's the standalone Siri app, let's say it's a kin to a chat GPT app. And then there's how you interact with Siri on the screen of your phone. You know, however, you open up the phone, we're going to go through some of the images that you included in the story, But could you just explain those two parts that we're detailing, and as we cycle through the images, we're showing them on.
The screen right now.
Again, these are Bloomberg generated images based on our reporting, both discussions with sources and documentation that we viewed but those two new features.
Please, So to launch Sirie today, you say, obviously the SII wake word or Hey Siri if you have an older.
Device, or you can hold down the power button.
That continues, and there's a new animation that pops out.
Of the dynamic island.
Obviously they added that with the fourteen proback in twenty twenty two, so this is with modern iPhone hardware in mind. The second thing you can do was swipe down from the top center of the iPhone. So how you open notifications today will be how you open a new serie interface called search or Ask, and that essentially takes you into a type to serie interface basically a system wide AI agent. You can tell it to get things done on your behalf. You can do searches on your device
as well as the open Web. So there's a perplexity competitor from Apple, built by Apple, developed by Apple, designed by Apple in there as well. And then there's the serie app. You know, chatbots have taken the world by storm. Chat GPT has nearly a billion users. People are using Gemini, people are using claud. This is clearly something that people want. So this is a product an app in line with what you're getting from Google Gemini.
Obviously, as we reported last year, the.
Underlying models for a lot of these new technologies in Siri are powered by Gemini and running on Google's cloud infrastructure.
How worried should like a chat che should open aib right now? You reported that maybe they even considering legal action. Is this going to be a real standalone competitor.
Well, we've seen this time and time again where Apple releases a standalone app of its own that's built into the operating system. I think chat GPT has the strongest brand. Serie doesn't have a very strong brand. I still think they should, you know, rebrand the whole effort. That's neither here nor there, at least for now, but definitely having a chatbop built into iOS, mac os, iPad os north
of two billion devices that is threatening. I would say to Gemini to chat GPT and to Claude, especially if over time Apple is able to make it really competitive to what you're seeing to chat GPT.
So we'll have to see how this plays out.
Over time, but at least, you know, in the short term, a lot of people are going to be introduced to the concept of a chat or a conversational AI interface who haven't used it before, despite the popularity of church.
Gipt, Bluebo's Mark Gumman, Thank you very much. Let's get to another story. Shares a Salesforce up about a percentage point, kind of a muted move, and you're gonna understand why. In a few minutes time, the company gave a revenue outlook for the current period that just fell short of an this estimates Bloemberg's Brady Ford covering Salesforce with us. Now we're kind of waiting for a different story in a minute's time, which might give us some context on
that move. But what was the story of Salesforce? Forget the outlook the numbers. What were they saying?
Application companies are trying to reinvent themselves as AI companies, and so Salesforce came out and said, we have an AI product that's kind of ramping and revenue.
We have all these.
Positive traction points, but our core products, for you know, sales and service are slowing down, and so it's how quickly can they reinvent themselves At this point, not very I mean their app a percentage point today, but if that chart zooms out, it has not been a rosy picture for Salesforce, and the last year or two it's down.
Thirty percent year to date.
Brody, I loved your content on LinkedIn, just showing that, like basically all earnings caols are now turning into podcasts, so they're.
Trying to reframe and rebrand the way in which they present these results.
But like many off is saying, this is still a record quarter, So how do we get more confident that.
There is going to be that.
Second half inflection on organic growth?
Yeah, it's just that reacceleration story, right, I mean, their biggest products keep slowing down, They've done acquisitions, they've added new products, and so there's kind of a lot of rosy things to look at. But until that kind of core sales, cloud service, cloud that really built their house, until that speeds back up due to AI, they're going to stay in that penalty box.
Should we just very quickly talk about the actual story today? Teams throw it up on the screen. Snowflake is going absolutely parabolic.
Yep.
Why, it's a tale of two cities, right, I Mean snow makes data infrastructure software, So a lot of companies have their most important data sitting in Snowflake and they have to use it if they want to do all these cool AI features on top of it. And we're seeing in the numbers that demand for Snowflake's products is kind of going gangbusters. So they had a great day.
They did, and Brodiefd, you're gonna stick with us to talk about a little bit more Snowflake so you can see that move means it's added twenty two billion dollars to its market capitalization. I mean after beating and expectations. They've also locked in that massive six billion dollar infrastructure deal with Amazon. We could talk about that with Brody in a minute, but first of all, just take from the Snowflake CEO himself, Shieto Ramaswami.
We spoke to Melia.
First of all, they're.
A longtime partner. They are the biggest cloud service provider that be run on top of and on on top of Azura as well as GCP. The really important thing with Amazon is how we go to our customers.
Together.
Bothrians are extraordinary value that we deliver for our customers, and with deals like this we get massive economies of scale that let us pass on some of these savings back to our customers. We ounced a huge change in how we price AI that makes AI a lot less expensive for our customers. It's aided by deals like this because of this ability to bulk purchase confidently, which, as I said, in turn, we give to our customers create amazing products on top.
This deal makes us much more effective.
Together, Amazon is interested in solving customer problems, and having a data platform is a key part of solving customer problems, us being able to go to market together. We work at every level of the hierarchy. Matt and I are garment, the CEO and I are in constant touch, but so on our teams. It's that ability to collaborate at a deep level to solve complex problems, for example, like ed data migration that makes us pretty unique.
It is truly a better gather story.
What timing Caroline's conversation with the CEO of snowfake and audio issues there that's tech happens.
Is this stock up thirty five.
Percent because of a compute deal with AWS or is it up for a different reason, Like it's really difficult in this moment to see the impact of that relationship in how the market's cheering the name today.
It seems that most of the rally is due to their own products and the fact that they have a pretty strong outlook. But the Amazon deal is interesting because a big part of the software story right now is when you're spending all of this money on LMS on AI features, that does hurt your margins, and so anything that companies are able to do to get economies of scale and drive costs down, which appears to be what's happening, that's also good news.
I mean, ed will know this more than anyone.
The graviton offering that comes from AWS and the idea. In your story, you make clear that maybe they're pushing towards that because of.
These efficiency gains.
And we're going to hear from Shreidar a little bit later in the show about that brody, but well, broady, how are we seeing the adoption of coding tools Cortex in particular seven thousand more than subscriptions?
Is that a lot? When you're thinking about the competitors out there.
It's interesting because a lot of application platforms, they've had their own tools like a coding assistant or other productivity tools right and there in the platform. In a lot of cases, we hadn't seen great uptake for it. And so Snowflakes saying, Hey, actually the coding tool that we're putting on here is being used. It is driving revenue. That's kind of new. We haven't seen that from a
ton of other companies. And so I think that's a pretty significant positive point that's being reacted to here.
And you just think about where Shreidar comes from.
Like his whole business AI business was bought by Snowflake neither before, So no wonder he's managing to integrate at Bloomberg's brody Ford.
Great Nawk, thanks for joining now.
Coming up, Navy is going to be joining us making a major push towards electric marine travel, deploying one hundred electric vessels are across the Maldives. How nice because speaking with the CEO next, that's Bloomberg Tech.
US based maritime technology company Navier is deploying one hundred electric vessels across the Maldives to build an inter island transportation network linking airports, resorts, local communities. Roll up marks a major milestone for electrified marine mobility. Joining us now as Navy CEO some pretty badasharia back on Bloomberg Tech, and with respect, last time you're on the program. Nava
was in a very different place just getting started. Fast forward, you're going to put one hundred of your vessels in a really interesting market, and you're really ramping up commercially with one hundred million dollar deal.
Let's start with Moldives. Piece.
When we lost FOT, you weren't looking at this kind of luxury into island market.
Now you are.
Why Actually we always saw the potential for you know, transportation with the N.
Thirty for islands.
But what is interesting here is that Maldives is this place where there's a natural need for the technology, and that aligns very well with our vision. You know, it has a vision for twenty thirty net zero and there's over a thousand islands today, there is over you know,
two thousand, eight hundred gas guzzling boats. The fit is so on spot, and you know, we are very grateful for the partners we found who have like the you know JIH who have the similar vision of like developing the country's infrastructure.
Let's talk about the N thirty Pioneer edition. What does it show in terms of performance? What does it do that's unlike anything that's on the market.
Yeah, absolutely, you know, at a high level, our goal at NAVIA is to build a standardized foundational layer I would say the best possible platform on the water with you know, dual use case, whether that's transportation or defense.
And we are very focused on.
Making this reliable, long range and applicable for different kind of sea states. So this is not just the deployment of the boat, but this is also the deployment of a network.
So usually when you go somewhere.
You know you have one of boats that takes you to an island, but in this case, it's more like what you see in the land.
What do you see in the air? Right?
You have United Airlines, you have Black Lane, you have four seasons on the land right, But when it comes to the water, there is no standardization.
And what is different is that you know.
You are in this beautiful resource right, which are so into sustainability, and then you get on the water and there's.
A huge disconnect.
And that's where we really come in. You know, that experience part of.
It some pretty How much pressure are you under to deliver for JIH and the Moldives project, like a hundred of these are they built? How quickly do you build them? Where do they get built? How do they get to the Moltives give us a sense of how real this is.
Yes, absolutely, so, just to give a background, you know, jih is led by Mama Dali Jihana. He's one of the most influential business leaders of Maldives who have literally known as the man who built Maldives, driving force behind most of the prestigious hotels and resorts like the World War four seasons, and they have also a presence in the GCC. So I'm really working very closely with them the first year, this very year, we're starting with five vessels.
The first one is going to get there you know, end of summer, and then we are going to test out this fleet and at the same time we will be working very closely with you know jih to plan out the infrastructure, the routes and then the you know, deployment is phased over the next three years. So there is a bit of like the infrastructure planning, route planning and then the software lure of it.
Right, So we want to make this.
A very seamless experience where Malis really become almost like the you know, it becomes a playbook of how to replicate this in you know, many other places.
Well, how do you replicate that into a defense narrative not just luxury briefly.
Right, because you know, if you go back to it, the company is really focused on what we called building the generalized marine vessel platform and the that is the standardized core.
Right.
If you forget the if you strip it to the physics of it. The role of a vessel is to carry per unit payload pernit mile reliably efficiently and go the longest distance at speed. What do you put on top of that, Like people often get caught into this, Oh, is it is it a luxury boat?
It looks too pretty. No, forget it, it's just your physics, you know, it's just.
The physics of the vessel, right, And our goal is to get as many vessels out there as possible for us to win as a generational maritime company. What you have seen today in asymmetric warfare, Right, we have to move away from exotic vessel building to standardize you know, scalabild systems, and that you can only do when you're going to dual use and commercial use cases forces you to like ruthlessly cut down cost.
So that's a big part of it. When you have dual use platforms.
You streamline everything the building, maintenance, supply chain and so on.
So some realty fascinating, some reti battetaria. Thank you for joining us and Ivia and discussing the mold. These move meanwhile, from electric vessels to electric vehicles. Weimo set to deploy new autonomous vehicles purpose built forever taxi use without human supervision.
Now topped the OHI.
The cars will be made available to select riders in San Francisco, Los Angeles and Phoenix. You were one of those select individuals. Talk us through the ride you took and am I saying it right?
Yeah, o hi ohi as in the city in SoCal but also like oh hi. But for Weimo, like this is very serious, right, this is the next phase of them scaling. So the Bloomberg Tech audience has probably seen one of the white Weimo Jaguar Eye paces. This is a vehicle that Weimo developed with Zeka, an arm of China's Gli, and you know it's a completely different design,
more like a shuttle. But the point being that they final assemble these in Mesa, Arizona, and at scale of like you're talking tens of thousands per annum.
So in the first instance. Yeah, free rides.
Get user feedback, but if Weimo's going to break into the mainstream, this is what they see as being their mass volume transporter.
Fascinating and particularly the China Angler as well.
Ed particularly the China Angle.
But I think that's the story here that because it's final assembly in Mesa and that the skateboard arrives without any of the self driving.
Tech, they get going here, you are there.
I am a lot of fun read the Bloomberg story will have more on the socials later now coming up, Andthropics explosive growth has turned into a highly sought after employer.
We've got more in that next. This is boombog Tech.
Landing a job at Anthropic is so fiercely competitive that applicants are spending get this, six hundred dollars on.
Average on private interview coaching.
Candidates say the startups intense culture screen feels that's like an interview and more like therapy on the company faces some incredible pressure to survive economically while keeping its values. People like really cammoring to be part of that story. Bloomo, Joe Constance and place to say it is with us you do this fascinating deep dive what it's like to go through interview rounds and anthropic What is it like?
Well, I mean, I mean, for for the first things, it's competitive. There are so many applicants. Now, there's so many people who are you know, just would be thrilled to join. Even the most seasoned engineers, the most high level executives, recruiters tell me, are willing to take the call from the recruiter and the interview process.
While a lot of it is.
Pretty standard, the culture interview, from what I hear from candidates and recruiters is a little unusual and more of you know, a lot of companies, the culture fit interview is kind of a vibe check, just to make sure you know you're not.
Odd.
But then this interview is a little bit more.
They want you to be old in many ways, I mean differently, you'll push back against thoughts at all.
Different They have a very defined sense of their own culture and so they are looking for particular people to fit that environment.
Hey, Joe, a lot of people I know work at them.
FROPIK are pretty odd and if you're watching the show today, you know where to find me. So if you're a candidate, because I imagine Actually a lot of the blom Beg Tech audience are aspirational. They want to go and work at anthropic What does forty six hundred dollars actually get you? Like, what are you paying for? And is it working?
Sure?
You know, part of this story was interesting to discover a little bit more about this, this cottage industry of interview prep companies, these career coaches that are really selling you know, their services to help people prepare for these what feels like and you know, oftentimes very high stakes sorts of you know, rounds and rounds of these interviews and skills assessments, and so in some cases it's just some resources to prep candidates on you know, what types
of questions they can expect. For other coaches they're offering mock interviews.
And what time can you expect because it is different, you are going to be sort of pushed more than you might be elsewhere because they're really so based on the mission right briefly.
Right, So, I mean that's that's the thing that I've heard from candidates that sometimes they're not quite expecting as much intro respection, you know, these types of questions that are really pushing folks to reflect on, you know, past experiences, decisions they've made, have they felt about those decisions, which is a bit unusual for people who are used to just talking about, you know, this project they did at work and how it went and my great.
Is failing being how how I'm too orientated?
Yes, exactly, the most constants with pretty much the most read story on Bloomberg today, Thanks very much. Coming up, Meta is selling consumer subscriptions to its METAAI chatbot for the first time. Got that story next halfway through the program. Heck of a lot more to come. This is Bloomberg Tech.
Welcome back to Bloomberg Tech. Let us take a look at today's big number, twenty two billion dollars and counting is how much the market cap for Snowflake has grown from just one day's market gain. The stock we know is surging today. So after the software maker gain a stronger than expected annual outlook, we in fact managed to speak with a Snowflake CEO, should A Ramaswami on how AI is really helping to boost their bottom line.
Well, we had a landmark quarter caroline strongest sequential dollar growth in company's history. Product revenue op to one point three three four billion dollars of thirty four percent net revenue retention rate, a key metric we watch up to one hundred and twenty six percent. But I think the bigger news really was this is the quarter where we clearly
showed that AI is compounding snowflakes advantage in data. We did this ole fashion way by creating amazing products like Snowflake Intelligence, which is our work agent, which doubled it that option with respect to our accounts, and our coding agent, which excoed our coco which is used by more than seven thousand our counts. And this is what gives us confidence in the business. That's why we raise that your items from twenty seven to thirty one percent started performance.
But I think is much more of what does this mean for our future that you're very happy with.
Another big story in the world of tech is Meta. This is a two day chart. Yesterday the stock up almost four percent. We're basically flat today. Meta is selling consumer subscriptions to its Meta AI chatbot for the first time. Two tiers basic tier seven dollars n nine cents a month that's cool, and then one that's like kind of higher tier Meta one plus like I guess those of you that are more in it, we write a Bloomberg News that this is about offsetting the AI infrastructure costs,
but other people have slightly different take. I want to get to sweader Conjuria Wolf Research managing director joins us. Now, because you put research out right when the when the news comes, and you're basically saying our thesis, which we'd already outlined, is playing out you multiple new revenue streams. This is a potential bigger total addressable market for you, rather than an action to offset that high capex or high infrastructure spending.
Yeah, that's right, and thanks for having me ed so at a high level. One of the deeper dives that we did prior to the news coming out was part of the reason why Meta is even underperforming Google and Amazon is that Meta is spending like a hyperscaler without any clear line of sight into demand that Google and Amazon have in their hyperscale business. And so where is Meta spending all this money to really justify this type of spend and cap x and when will we see
that revenue? So that's the fundamental question, and so in that when we dug deeper, it could be subscription or it could be a gentic commerce or it could be business AI. And now we're starting to see with this product release of subscription across consumer businesses and Meta AI that this could be the beginning of.
It about how this is going to scale.
First quarter, it was about one point three billion dollars that was in the non advertising revenue.
So tiny, how much could that rise?
Yeah, so in the non advertising revenue right now, a lot of it could be business AI that they are actually monetizing in WhatsApp through their businesses in WhatsApp. Now an add on is going to be subscription. So a clearest comp that we have today is Snapchat, and Snapchat arguably has surprisingly done a great job in converting its da user, daily active users, and monthly active users to
subscriber base. If Meta can do something similar to that, and I'm not saying it's going to be exactly the same, but say low to mid single legit percentage of their daus actually convert to subscription, well, that in itself implies about a one to three percent percentage point uplift to their revenue. In other words, approximately anywhere from five to fIF teen billion dollars of incremental subscription revenue from consumers in the next three to five years.
Look at it differently.
In a world where you're paying monthly for Chat, GPT, Clawed, Gemini, etc. Is Meta AI worth paying eight dollars or twenty dollars a month four.
That is going to be a key question for them. So there is a consumer subscription piece, there is this Meta AI type subscription piece which competes directly with Gemini to your point, and then there is a business like if you are a creator, then you can subscribe as well. I see great value in the consumer subscription and the creator subscription because it allows them to create more content.
But in the question that you are asking, where if I have a Gemini subscription or a GPT, do I need Meta AI, jury is still out on that you would need it. If you're running out of capacity and you're paying twenty bucks, you don't want to pay additional higher tier to a one hundred or two hundred dollars tier, and you want just a little bit more of access capacity, maybe you bade bucks do Meta Maybe in that scenario
does make sense. Or if it is highly personalized social sort of a use case that they give us which cannot be created by Claude because they don't have that information, or perhaps the GPD perhaps there is a use case in those instances. Yes, but I'm not fully sure. Jury is still out on that. On the meta AA subscription scaling.
Schredo Cajuria, great research, great analysis from Wolf Research.
We thank you, Mick.
NASA it selected Blue Origin, far Fly Aerospace and other private space firms to help build out its long term lunar ambitions. The agency says monthly missions could begin in twenty twenty seven, laying the groundwork for astronauts striveging live work on the lunar surface. We spoke with the NASA Administrator, Jaredisingman about the timeline for a moon.
Base starting in twenty twenty seven. You should see see a near monthly cadence of robotic landers on the Moon, several rovers. In fact, we initially we provided in a ward for the first two, you know, crude and autonomous capable rovers for the lunar surface. So when our astronauts arrive on Artemis four and twenty twenty eight, they're going to already have some infrastructure of the moon base waiting for them.
They're already going to have a rover waiting for them.
And then in that timeframe, it's not just intermittent anymore. It's not just those monthly visits. But when do you think people be working? Humans might even be living in some capacity on the moon.
There, So we are approaching the moon base in phases. So Phase one is a lot of littles. We are dusting off the playbook that worked very well for NASA in the nineteen sixties. We're getting back to an iterative approach. So you know, there was the Mercury program before there was Geminy, There was Geminy before Apollo, and an awful lot of Apollo missions before we went right to the moon landing on Apollo eleven. We are doing the same thing now. So Phase one we're calling it a science
of survival. We're not going to lock in what the mobility strategy should be, for logistics for astronauts, the power strategy, the surface comms, the orbital coms. Why would we try and nail and get all of that perfect today when we haven't been to the Moon in more than a half century. So Phase one will be a lot of landings again, that near monthly cadence to learn and inform Phase two, where perhaps now you're putting a lot more
tonnage on the lunar surface. You have a lot more direction as to the type of hardware and capabilities you want to lock in on, so you don't need to have maybe monthly landings when we get into phase two, but you have a lot more direction as to what should work for our intended objectives, which is to build
out that habital environment. And then phase two we're going to learn now having astronauts go from let's call it a period of maybe even days on the lunar surface in phase one, to potentially weeks in phase two, to where you might get by the time we move into Phase three, a similar astronaut rotation like you see on the International Space Station, where we could have crews potentially being on the lunar surface for months on end.
You don't have that marked on your calendar, administrator, when phase three might have a base that has humans actually living and working inside it.
Oh, we absolutely have time frames.
I mean we are looking at basically twenty twenty seven through twenty twenty nine for phase one. You have twenty twenty nine out into the early twenty thirties for Phase two. But again, this is all going to be informed on what we learn during those first landings in phase one.
That was NATA administrator Jared Isicman.
All right, coming up on the show, we're going to be joined by Eric Vistria Benchmark for his outlook the physical AI space. Yes, we're going to talk about hardware. This has Belen bog Tech, French AI startup Mistro AI is expanding into advanced manufacturing, striking deals with new customers Airbus and BMW as it looks to so called physical AI to fuel growth. CEO and co founder of the men, she spoke with blue Ink Tech Europe's Tom McKenzie.
For US, it's a massive market.
We see in particular the europe is our in core market and one of the Europeans the strength of Europe is to is in its high end manufacturing. So it's the manufacturing world. Is the thirty trillion market. If you think of a police that I can bring to it, if you only look at ten percent of that, you're looking at the three trillion market. And that's happening in the next five years.
Man Over in Asia, Mini Max is annualized revenue more than double these past two months to at least three hundred million dollars. As the Chinese AI startup prepares to roll out its next flagship model, Mean Max co founder and President's Yaean joined Blomberg Steven Engel on the sidelines of thebs Asian Investment Conference over in Hong Kong.
Agent and the models are really important for manimalization, but definitely the foundation model is a key. You will see a better performance. Models with specialization and differentiation will drive the token consumption, also drives the enterprise and the consumers retention and the consumption. The model performers product commantization will become the frive wear. So you will see we did
the end to end optimization with the whole group. There are lots of innovation, technical innovation inside, so you will see we can provide probably similar performers models with probably lower price even sometimes higher margins.
Yeah, so how do you change your revenue mix where most of your revenue is coming from your consumer facing products like the chatboxing and also high level which just last year's that's like the text to video generation. But next year, as you go into your model three, right from two point seven to three, how is your revenue mixed going to change?
Yeah?
So the number you mentioned is last year's number, but right now it's almost like fifty enterprise and fifty consumers, so the enterprise increase a lot. And also, yes, the model is a key, so we think the model is our product, no matter it's a B to B R digital C. It's all the Chinnel's for the commercialization. So we've spent most of our resources and the spendings of
the model layer. Yes, we are going to release I'm three very very soon in a few days, which probably I think, which is the first open source native multi model.
That was Min and Max co founder and president Yea Ian along with our own Stephen Engel. We're gonna stick with AI and we're going to discuss the outlook for physical AI with Eric Visher, a partner at Benchmark Today's VC Spotlight.
Been really looking forward to this one.
Eric.
Ten years ago, you're very welcome to be here. Ten years ago you led a series A in a little company called Cerebras. Fast forward ten years what an IPO, but it's indicative of where we're at right now in this demand for fast inference.
I just want to start with that case study.
You know, the timing of this IPO and where it fits in what's actually happening in physical AI right now.
Well, I think that it's very very clear that the demand for inference and AI is operal charts, and I don't think that's going to stop anytime soon. I think it was Alex Scattered at at Whale Rock who kind of recently said, if you think of the population of the world, there's like one percent of the world.
Is maybe AI power users today.
So if one percent of the world is AI power users today and we're completely compute constrained, for as far as I can see, what happens when three percent of the world, or four percent of the world, or five percent of the world becomes AI super users or power users. And so I think that we are going to be in this compute constrained world for quite some time. And I think that that's going to lead to a lot of success in all of the hardware layers. And it's
also going to lead to the bottleneck moving around. You know, some days, some months it's going to be memory, some months it's going to be data center and power. Some once it's going to be chips. And I think that bottleneck is going to keep.
Moving around some days it's all of them combined, Derek, So totally at this exact moment, where are the startups you're most interested in, or the ones you already sit on boards of starting.
To innovate at the edges?
When are we going to really see like movement of photonics or a different type of compute being used.
Well, I think, I think you know.
One of the benefits of being an early stage ventry capitalists is we have a very long time horizon, so we're not trying to figure out what's going to happen in eighteen months or twelve months or twenty four months. We're really looking and trying to have some idea of what might happen in five years or seven years or ten years. And as we look out, we are really excited. For example, we invested in star Cloud, which is a
space data center. We invested in Sunday Robotics, which is a domestic robot And when you kind of work at companies like that, they're very much on the frontier. It's going to take a bunch of time. There are very capital intensive projects, but they're amazing teams that are doing really cool development, pushing the edge, and it's going to require a lot of flexibility on their parts as the market evolves. There's obviously a lot of unknowns, but there's also tremendous possibilities.
I just want to go back to Cerebris for a minute, and to some this is now ancient history is academic. But two days before the IPO I broke a story that arm and soft Bank had basically gone to Andrew Feldman and said we'd buy you for a large number. And Andrew very quickly shut it down, and the rest is history, of course, because they went public. But it's somebody that joined the board in twenty sixteen and has had three if the five major investments go to IPO.
What would you have made of that outcome instead of going public.
Well, obviously you can't comment on that specific reporting, but we're a public company now.
I think that has opened up.
A lot of possibilities in terms of what we can do. We've raised a ton of capital to finance the business and allow us to take advantage of the tremendous demand. And I think we're really at the beginning. You know, if you feel like there is an end in sight, you might take a different, different tact. But as far as we can see the demand is tremendous.
You say, you're just backed a company and it's about orbital space centers, and that immediately makes us all think of SpaceX and how they're going to be sucking a lot of the oxygen out of the room when it comes to a public offering. How are your companies currently feeling the need for cash? How are you thinking about permanently fundraising or looking for exits right now?
Well, you know, it's really interesting.
Right now, the venture landscape is very much have and have not, which is, if you are oriented around AI and you're growing really quickly or have something that's very much on the frontier, there's almost there's almost like limitless cash available and funding available. And if you're not, even if it's a good business that would have people would have fallen on all over themselves. For you know, five
or six years ago, there's almost no funding available. So it's a very it's a very bimodal setup right now, which is which is challenging certainly, But each of these financing eras kind of they come and go, and you know, one of the really important things for any of these companies or entrepreneurs is just to keep on, keep on grinding and finding the way that there's building a company
that is impactful as a roller coaster. And it takes a long time, and even if you take the Cerever's journey, there's been lots and lots of ups and downs over the time. Obviously, taking an AI sending conductor company public in maybe twenty twenty six is about as good timing as you can get. But you know, part of that is the timing. Part of that's just luck. And then obviously it's all built on top of a decade of the team grinding.
At okasor another interesting one that we're looking at in your portfolio. We have to talk about that another time. Eric, it's great to have you on. Eric Vishrha of Benchmark. Now coming up TikTok, but it's moving away from music, scaling back tis with major music labels.
We'll dig into that next. This is Bloomberg Tech.
AI could provoke a fifteen percent displacement of knowledge workers. That's according to Muddy Waters Capital CEO Carson Block, who joined Bloomberg's has Linda Amen in an exclusive interview to discuss AI demand listen to this.
Our house view is that we're going to see fifteen percent displacement of knowledge workers. You know, we think it could be as soon as three years, is it four?
Is it five?
At some point and it's in the single digit number of years. This will this will be a factor or this this will occur in our view, and yes, there will be jobs that are created by AI, but we're talking about net losses because the technology is increasing in capability faster than we humans are able to adapt to it.
Let's talk about the displacement of music labels.
Maybe over at TikTok, because music has been called to TikTok's identity since its days. Is musically helping artists like Little nas X or Olivia Rodrigu just global stardom and waits Now TikTok there is skinning back ties with major music labels and focusing more directly on those artists.
It's according to sources, that's.
All discovered by Blue Megs Alex Levine along with Ashley Carmen.
What is happening with a company that identify with music?
I mean it's it's in the icon, it's in the branding. How are they moving away from labels?
So exactly?
Music has really been part of its DNA since the very beginning. It is the thing that made helped make TikTok this global cultural phenomenon and got more than half
of America using it. Though TikTok continues importantly to work with major music labels, including some of the world's biggest, it is deprioritizing those relationships in part by building out projects, prioritizing internal efforts to actually have products and services that compete directly with the labels and that allow the company to have sort of more direct relations with the artists rather than through the representatives.
Alex, TikTok changed music. Now labels worry it's leaving them behind. But present day, how does music work on TikTok. So it's Friday night, I'm kicking back on the couch. I go to YouTube on the TV and play concerts, music videos. I don't think present day like, you know what yep, music music video like, just explain what we're talking about mechanically.
So mechanically, when you open your app, you've got your four you feed. Every video that you see is going to have some sort of audio behind it, whether that's people speaking or whether that's music, and oftentimes those sounds are our songs that have gone viral, and sometimes it's
new songs from emerging artists. Sometimes it's made you know, it's it's global hits from artists like Paul McCartney, like Bruno Mars, and sometimes it's simply just you know, sort of repetitive meme type noises that you can find through
through various other means on the app. But I think that there's always sort of been this question, especially more recently, about whether blowing up on TikTok or going viral can actually mint a legitimate star and and have have them develop really an enduring career from that.
Bloomberg's Alex Savine with what's going on in music on TikTok, Thank you very much.
Character that does it for this edition of Bloomberg Tech. What an edition has been?
Yeah, a lot of market moves, a lot of great interviews, a lot of top stories, recap up on the podcast. You know exactly where to find it all the Bloomberg platforms and online, Apple, Spotify.
In iHeart, this is Bloomberg Tech