Bloomberg Audio Studios, podcasts, radio news. Bloomberg Tech is alive from coast to coast with Caroline Hyde in New York and Vla Low in San Francisco.
This is Bloomberg Tech coming up. Memory chip prices strike again, this time taking a toll on Cisco, even as AI demand surges will break it down.
Plus, Apple's revamp of its Siri voice assistant hits new snags that could delay the release of some features.
An Altruist launched AI tax planning tool, and within ours wealth management stocks tumbled. The CEO of Outist joins us to discuss.
First we check in on these markets that in many ways are still digesting AI disruption. We focus in on Cisco that I know you'll go to. We're off by one percent on the Nastak one hundred. That anxiety of costs, of memory costs really building into the benchmark. But more broadly, we're all eyes on costs for consumers. CPI print coming later in the market. Just a little bit nervous ahead
of that. Bitcoin down another day of five percentage point sixty seven thousand is where we trade for this particular risk asset ed What.
Are you looking at there's a lot of pieces of consecutive red on the screens we've been putting up. Cisco no exception, down eleven percent, on track for its biggest drop since May twenty twenty two, and prior to today's session it was up eleven percent and had momentum. Outlook is strong on sales for the current period. AI demands there, but there's margin compression memory chip prices. Let's get more with Bloomberg Intelligence senior hardware and network analysts Wou jin Ho.
It's actually probably a good time to remind us what Cisco is and what Cisco does. But the memory issue is factoring in to servers, and right now you are leading with your REACT piece on the Bloomberg terminal with the memory chip issue. What do we need to know.
So really quickly with Cisco, is is the leading global provider of a networking equipment for corporate as well as for data centers, and they've actually had a foray into cloud data centers to help power some of the AI workloads. Now, what happened last quarter? They got the AI orders, you know, momentum and drove a lot of the upside on sales, but the gross margin issue especially given from the dram, pricing is a lot bigger than we had thought.
Chuck Robins, though, was trying to say that, like, if anyone's going to navigate this, they're going to navigate it. They're already going to their suppliers. They're already thinking about increasing prices. How well can they handle it?
Yeah, So we did the math for the third quarter, there's about two hundred basis point product derosion on gross margin. But when we look at the fourth quarter, you know they're able to stabilize that. Now, if we think about twenty twenty seven, they actually have a fairly sizable software business, and if software really starts wrapping up, we could start seeing some gross mogstability in twenty twenty seven.
W Jino Blomberg Intelligence. There we have your p Cisco's results outlook pinched by memory price spike. Look, all of this is as Cisco's numbers. They highlight the heavy infrastructure build out, but have investors really questioning what happens next for the implications to software too, joining us now to
go across the whole gamut. As Tony one portfolio manager tro Price Science and Technology Fund in his latest thoughts, he says that the recent software sell off is more of a valuation reset and that companies that successfully embed AI into workflows could drive stronger productivity and margins over time. And Tony, what's so great about having your voice on this? Shows you time and time again have been like the proof point the ROI is going to be productivity gains.
Suddenly we've got the productivity gains. Saw a shining a light that we're actually worried it's going to really hit valuations. How do you weigh where we are in terms of the selloff and if it's overdone?
Yeah, well, I think it's the key question the market is wrestling with bright now. And so when I think about it, it's it's really just what is on the right side of change, I think, and what is the inevitability here? And I think what you're seeing with plod cowork is that these agents are becoming more autonomous and they can actually execute things on their own, do more than just like you have to prompt something like chat to be team, you know, for for your your own
personal chat bete. But so what it's it's exciting is that these enterprise agents, I think, are going to be more more of a force here and as we're starting to see that really pay off. And when we think about it, like I think the why the markets reacting to the way it is because if you think about it in the in the future, like, there are wide sweeping implications for what modes mean, like in terms of software as well as information services. So it's definitely a disruptive potential force here.
Disruptive to some of those in your portfolio. I'm looking you are exposed to salesforce, you are exposed to accenture, which people have been worrying about the consultants side of the equation. Even Shopify and delimit strong numbers got beaten up yesterday. How are you being discerning at this moment of where you actually buy into on a dip because you do think that they'll manage to ride the software disruption.
Yeah, so I think that you know there it's more important that you are the platform and not just a feature. Because you're a feature than AI agents can encode software so easily and coustiously that you essentially can be embedded in a bigger platform. And so I think that increasingly you want to make sure that you're the layer that all these agents need to coalesce around, and I think that there's going to be a lot of change here
in terms of and we honestly don't know. And so I think the market is in a sell first and ask questions later mode, and so it's important to kind of stay on the right side and manage these these changes in the market.
Tiny Nvidia CEO agents one has this line that software where is now worth paying for? The example he would give is Cursor. But I guess what we're trying to do with you is define what software is that is
worth paying for. You just brought up Salesforce, right, so in the most recent quarter they had like more than a billion dollars of arr directly from AI and they talk about all the almost nineteen thousand agent force deals that they have, yet they're still subject to this same AI is going to eat software's lunch chat that's going on right now. Why do people not believe the traction that they have as an example that it's worth paying for.
Yeah, So I think.
There's two things that I think about.
One is that there's just so much competition when you think about all these AI neted startups and they're pricing in a way that is really compelling and to drive ad option essentially, So you think about like image generation, right, like you pay twenty dollars at chat to bet you can generate unloaded images, whereas on Adobe or Firefly you
have to pay per image generation. But you know, you can think about you know, cloud cowords in a similar way in terms of what they're pricing versus the value that they're delivering is really compelling. And so I think, what you just have a lot of price inflation, and that's why I think a lot of these you know,
software companies are looking to drive adoption. You know, the economics aren't as good as the seat based model, and we're going from like a sea based model to outcomes and agents, which is a very different, different version.
Just for reference to audience. And as that one hundred near session low is now down one and a quarter percent. This story started in Europe in August. I was there in London. I remember the day all the European software that was sold off. Earlier today we spoke to the ceman CEO. Let's get his take.
We know that AI has the potential to disrupt some of the software businesses. I mean, if you talk about software, it runs on call centers and the like. I mean, that's fully on the block. But if you go on the other end on industrial software, simulation software, physics, space software, let's we don't see that at all. We rather see an enrichment of the software which allows more and more customers to use it.
Another software CEO whose fate is very closely tied to a relationship with Nvideo, and I think it's important to point that out. But his argument is that actually specialized software is getting better because of AI. You buy that.
I think so, And I think that most I think a lot of the economics that we've made and the specialization of the niches, so you know, domain specific knowledge will always be very important, and so you know, I think just general knowledge is probably going to be not take the economics, and so what really matters is like being able to put this into ROI now. At the same time, it doesn't mean that if you have a stronger model, that doesn't mean it doesn't create more competition
for existing comments. So I don't think the battle is one or loss for these companies, and I think it's just important that they stay you know, front footed and incorporating the latest technology trends into their.
Platform and your science Technology fund has all the trends. You're in hardware, you're in software, you're number one holding is alphabet at the moment, then in video. But going back full circle to Cisco who exposure there? What do you think about the memory price implications and how you can weather who are the winners and loses in that situation?
How long for?
Yeah, So I think that you know, memory is fascinating industry because one is secular trends and AI is now this new tam and then we've also gone through multiple years of under investment because memory has been a downturn.
So I think one it.
Is that they're secular, but there's also cyclcal elements a bit. So you know, in terms of our portfolio, we definitely own a lot of the bottlenecks where the economic value capture is in such as memory. At the same time, I think that you know, Cisco networking and this is also going to be part of in AI. I think that the implications of the memory a shortage is actually becoming more widespread and having implications that a.
Lot of times the market didn't think about.
So you know, it is fascinating, But I do think Cisco continue to be a really impeign company that's driving a lot on this AI adoption tourny one.
Duro price great heavy back on Bloomberg Tech. Really appreciate it. Now coming up, Apple's new Sery runs into more snags during testing. What's delaying the AI up grade? We have the Bloomberg recording next, This is Bloomberg Tech. Apple's revamp of its Siri voice assistant has hit new challenges that could delay the release of some features, according to sources, Bloombo's Mark German, as always, has the details. Let's get
into the details. What do we need to know? What you what are you hearing from inside the company about the latest snag with artificial intelligence powered Sery.
So Apple has been planning to launch the new Seri.
This is an AI revamp that initially introduced in June twenty twenty four at its Developers conference, and I'm going to tap into features like your personal data to fulfill queries onstream content and precise control of apps. It was supposed to launch next month in March. Is part of an update called iOS twenty six point four. This had been the plan for several months. It's actually been the plan for well over a year at this point. Now
that is being delayed and staggered potentially right now. The delay is moving many of the serie capabilities to iOS twenty six dot five that's scheduled for some time in May June, so about a couple month delay. But some of the features, including the personal context features, those are likely not to launch until September as part of iOS twenty seven, so a bit of a delay there, you know. Altogether, it's possible that these new features are not going to
launch until two years after they were initially introduced. The setback for Apple here is not necessarily in bone sales. People are still buying iPhones. You saw the eighty five billion dollars worth of sales, which could have even been more if not for some manufacturing challenges. But the challenge is where is OpenAI whereas anthropic? Where are Google going to be at the end of this year? Right, They're
likely to be well ahead where they are today. So that's really what we have to pay attention to.
Mett's talk Mark about how therefore the Google introgation is going, because in many ways it just fits into the overall Apple intelligence stop start and how they're going to be building out from within.
My belief is the only reason they're going to be able to get this done this year, which I still believe they're going to do despite these latest snags during testing, is because their models are being improved by the Google Gemini team as part of that collaboration between the research labs at both Apple and Google. So that is going
to be a major boon for them. I do expect this big serie chatbot revamp in iOS twenty seven coming out with the iPhones in the fall to be pretty nifty and pretty cool, a new way to control your phone with AI. They just need to pull it off, and I think they will. But the big question is how impressive is it going to be in September compared to whatever Open AI in Gemini has for their own customers and bropias with Quad in six months from now.
Bloomberg's Mark German, we appreciate you on all things latest with Apple. Meanwhile, let's turn our attention to social media now, because Instagram boss Adam Massari testified in a landmark case they're like in social media to digital casinos. Masari told the jury he doesn't consider problematic use of platforms like meta to be the same as quote critical addiction. Blom Meg's legal reporter Madlyn Mackelberg is with us. You were there in the hearing, NLA, and so the push on
Adam Massari anything different. We've heard on other occasions that he's testified.
So he covered a lot of similar ground to things that we've heard before in congressional testimony and other public remarks about this issue of social media addiction. Obviously, what's unique here is the context in which he was giving this testimony, which is this landmark case that's testing whether juries will respond to claims that these companies knowingly designed their platforms to addict young users. But we heard him
cover a lot of different ground. As you noted, he talked about this issue of clinical addiction versus problematic use, which is something that we hear a lot from social media companies, referring to people who use the platforms more than maybe they feel good about. We also heard him talk about some internal emails in the company when he was asked about them by the lawyer about decisions they were weighing about certain features on the platform related to photo filters on Instagram.
Mad this is a process, right, both sides present arguments of plaintiffs, what do we need to know about what happens next, who else might testify, and what is under consideration at the cort of the trial. That's right.
So we're in week one of what's expected to be an eight week trial. Adam Asseri was the first executive to be called to the stand, but he will certainly not be the last. As you said, this is the plaintiff's case. They kind of get the first chunk of time at the trial, and then the defendants in this case, Meta in Google, they're going to get a chance to respond.
We're also expecting to hear from Mark Zuckerberg of Meta, We're expecting to hear from YouTube CEO Neil Mohan at some point, and we're also we kind of got a little bit of a preview that we're going to be hearing from executives a little bit further down the food chain who maybe we're weighing in on some of these decisions.
But the heart of this case is about this question of addiction and whether or not these platforms knowingly designed these tools to hook young users resulting in mental health struggles. This case, of course, centers on one individual, a young woman from California and her specific allegations, but it represents thousands of other cases that have been filed. So there's a lot of attention on the outcome here because it's going to dictate how the rest of these cases are decided.
Bloomberg's Madeline Meckeberg, thank you very much. Now coming up, Altruist introduces a new AI way to plan your taxes. We speak with Jason wenk Voucheris's CEO. Next, this is Bloomberg Tech.
AI.
It's added again and it's been a wonkey, worrying few days for wealth managers and those who had shares in it considering the technology's impact. It's all started Tuesday, a startup called Altruist added a tax planning tool to its AI model Hazel. Within hours, wealth management stocks had tumbled. Charles schwab lpl Woymond James, as you see significantly, and that's as investors really.
Feared AI disruption.
The CEO of the company that kicked all of this off, Altruists, Jason Wank Jason, did this citate you?
I surprise?
Yeah. I think, to be incredibly honest, I don't think anybody could have possibly thought that you know, somewhere between you know, tens of billions and maybe one hundred billion of market cap would be wiped out in a week. AI is very powerful and certainly the market you know reaction tells me that there's a lot of people voting with their dollars in the potential disruption.
Jason, all of this is about enabling well wealth managers individual wealth managers. From your part, just remind us what altruist does and what this new development in terms of building in a tax tool does to help them.
Yeah, absolutely so.
I think the nail on the head, like the most important thing is that we're not trying to displace financial advisors or wealth managers.
It's not AI taking those jobs away.
It's AI to help empower them to do better work for their clients. And in many ways, when people think about wealth management, you know, there's there's deservedly a lot of critique. It could be things like why is it so hard to get access? Like the minimums are really high, Why are the costs very high? Why are the results sometimes inconsistent? And a lot of that is, you know, because of how much manual process and human labors involved.
You just have a lot of limiting factors.
AI very much is an incredible equalizer, you know, effectively with a tool like Hazel what used to take you know, teams that were you know, ten or more people working hundreds of hours and costing many tens of thousands of dollars, It can be done in like two to three minutes. And you know, users, the advisor users in our case are paying one hundred is dollars per month, right, so much more affordable, much more accessible.
And tax is just the start.
We launched tax first for what it's worth, because it's tax season in the US. We wanted that to be our first agent, you know, to kind of get out there in the hands wealth managers and advisors. But this is a pretty good harbinger of the things that will come. There'll be more agents doing more work that will continue to make advice better, more affordable, and more accessible.
Jason, you will have noticed that some of the CEOs of those wealth managers appeared on television in the days that followed the selloff. His child swab listened to what they had to say.
AI is a real accelerant. So it was puzzling to see our stock sell off related AI because we're benefiting it from it in multiple ways and bringing it to our clients.
It's allowing us to reach new clients.
And I think that's why you won't see the population of advisor's shrink, is that AI will be useful in reaching new clients that we couldn't before.
Has mister West or any other leader in that field phoned you since Tuesday and would you be open to working with these companies literally jointly. Yes.
So no, I haven't heard directly from mister Worcester, but we already actually work with any dependent financial advisor. One of the reasons why we stood up Hazel as a separate entity altruist of course as a direct competitor to Schwab, Fidelity, LPL and others. But we wanted Hazel to be something that any advisor could use to help any client, and we didn't want them to be forced into using a
single custodian. So there are already thousands of advisors who use Schwab, for example, as a custodian that are leveraging Hazel and doing so to drive better outcomes for their clients and a lot more operational efficiency for themselves. And we have i'd say, in the last forty eight hours. It's in the hundreds of very large, significant, you know, wealth management companies around the world reaching out to see
if there's a way that we can partner. So definitely looking forward to, you know, putting this tool in as many hands of as many capable advisors in the US as possible.
Let's talk about the tool and the technology. You know, what are the data sets that the underlying model was trained on. What is your coal competence that if there is concern from the legacy industry of wealth management, what is it they should be concerned about, And what is it that makes this model powerful?
Well, so kind there's two parts that I'll try to be succinct. But the first is I think what causes concern is not so much just the AI. It's the fact that we also have built a very modern infrastructure for our wealth business. So the things you can do on altruists you just can't do anywhere else in terms of the speed of opening accounts and funding accounts and managing of assets. If you take AI and you put it on top of bad, antiquated infrastructure, you're not going
to get a ton of value. It's you know, kind of akin to putting like a you know, self driving on a horse and buggy or something like that, Like you need to have like an entire vertically integrated ecosystem. I think, if anything, that's maybe where some of the market concern is is that what happens if something like you know is digital and it's twenty or thirty percent better, you add AI and multiple agents, it becomes two or
three hundred percent better. Now advisors could very much have a shift in where they decide to custoy their assets. That could be pretty monumentally difficult to overcome for some of the big incumbents.
As far as what makes the model unique.
The way that we built Hazel's we rely, like most AI companies, on having really clean data.
So data is a huge advantage.
I think, you know, mister Worcester was right in saying that data is very critical. In the case of the custodian. You have the ultimate system of record meeting, right. We have all of this data on clients transactions. That makes it a lot easier to give personalized analysis and feedback back to the users.
Jason Mike Vouchers, Sorry, we're out of time. Grateful for the conversation coming out. We look at way moo. We'll be right back. This is blue bag attack.
Welcome back to Bloomberg Tech. A lot of action in the markets today, let's just go through it. In terms of the NASDA benchmark currently off by one point eight percent, we're at session lows. This is as we worry about earnings and how they're implicated by prices. Prices and Memory Cisco in particular down significantly. I want to shine a night and off by eleven percent. So we've got that anxiety about whether AI is really helping these businesses or
whether pricing pressure is hurting them. Remember the macro pictures are about get CPI data, so pricing pressure for the consumer too. As we build up to that, the market's just a little bit jittery. Add Ian over in European trading check it out by twenty percent. This is the payments company Fintech, which is being hit by consumer.
Sentiment being on the downside.
Also a week of dollar hit them as well in terms of payments, revenue and their earnings outline, just really not living up to expectations. We move on and have a think about what's happening when you've got exposure to the private sector now many are saying, I'm talking BTIG in particular, saying that soft bank is like the way to play open Ai exposure as well, because they've already
put thirty six billion dollars in. They're looking to increase that maybe up to the tune of thirds thirty billionaire. They have an eleven percent stake in the business, and actually they managed to post a profit, the fourth quarterly profit in a row, not seen since twenty twenty one, as their investment stake in open Ai goes up and to the right, helping offset some other losses. But the stock traded in the US if you're looking at depository receipts currently up by thirty percent.
So we'll dig into that a little bit more.
End.
But it's interesting what's happening in terms of a soft bank story.
It's a little bit like shuffling the deck and like you know, you have to sell at certain holdings to fund investment deeper into the names that you're most interested in. Like it's not rocket science at the end of the day. Carry one quick question to ask is like how wide soft banks is exposed beyond one single name like open Ai, Given how ferocious the lab battle is right now.
Yeah, opening either in on Remember they've been going into chip design companies like I'm peer, they've been looking real exposure to ARM but they've been coming back on their T mobile exposure to be able to get more into these privately held businesses. Also been buying private equity companies right because they want the exposure to day to center.
Yeah, I'm glad the team got that up. I mean, you know, far right column says everything. Let's get to another story in the private markets that we've been trying to update you on regularly. Anthropic is close to locking in a deal to raise more than twenty billion dollars in a funding round co led by investors including Pieter Teal's founder's fund, The Shore and DRAGONEA that's, according to Bloomberg's sources, one of the largest startup funding rounds ever.
And who's been across at Bloomberg's VC reporter Natasha mascarinus important detail, Actually, Like some of those names are are interesting. You know, dragone has come up a lot recently. Weimo was something we reported a couple of weeks ago. What do we need to know? Like how close is this and who's leading? Yeah, listen.
On Monday, we were discussing where are the crossover funds in this round? Now we have some examples with d Shah, with DRAGONEIR. We also have a continuing list of traditional venture capitalists in the round. What I'm really paying attention to is the addition of Founder's Fund to the cap table. That's an early open AI investor and founder just Want is known for concentrated bets, So it is interesting to see an early open AI investor go ahead and be
a major investor in this anthropic ground. And Founder's Want is not alone. We're seeing Sequoia as well in the round, and the list grows.
It's interesting.
I know another VEC that makes concentrated bets and doesn't go in to other companies that seem to directly compete. So how much of a taboo is it that vcs back to horses, almost like having some sub option here.
Yeah, I mean it used to be a huge taboo because of concerns about information leakage or how do you really take that out of your head when you're in a partner meeting and discussing you know two different roadmaps that are you know, both in this case going towards a version of AGI. What I would say and what we are reporting is that certain investors get certain information rights. In the case of open AI, sources are telling us that if you have a certain amount of ownership in
the company, that comes with its own information. And so what we're really wondering is who's taking big checks in these rounds, because they are the ones that will have to answer the questions of how are we keeping the information separate?
Bloombergs and Sasha Mascarinas, thank you very much. Now, Waimo is aiming to have a big year, adding new US and international markets. The robotax firm also recently raised sixteen billion dollars from parent Alphabet and other outside investors. Co CEO Taqijra Maracana sat down with us and discuss the company's plans, including whether it's considering an IPO. Listen to this.
We are just laser focused on execution, you know, building weimo to be financially responsible, operationally excellent, and then make sure we maintain the safety culture. Like that's what we're really focused on. Having this vote of confidence, as you said,
not only from Alphabet, but from our three colleagues. From this round and from all of the new investors who decided to join our cap table and the existing ones who doubled down on their belief that this is the right opportunity to fund, and we just feel humbled, but also there's a lot to do, so we're really focused on making sure that we can scale, focusing on our two first international launches, you know, London and Tokyo, and scaling across the United States.
We do not yet have, although there has been progress towards just this week, a federal level framework set of rules. What do you think the direction of travel is with that? And you know, clearly if you had to deal with one set of rules and not a city by city, let alone state by state basis or case, you'd be making a bit more progress. I suppose.
Yeah, we think it's really important that there is a federal av standard. We've been advocating for sort of a safety case based approach because the technologies are different and we think that the burdens should be on companies to demonstrate why you believe your technology is safe enough. We also think there should be transparency requirements. You know, people should have to demonstrate how many trips.
Are you providing?
You know, I don't right now.
The balance isn't quite there.
I mean, some states require a lot of reporting, some don't require as much reporting. I think the United States has an opportunity with this technology to lead globally, and I don't think you can lead globally if it's a framework that's governed by multiple jurisdictions across the states. And it's a way to slow down the adoption of this technology not only in the US, but in other markets.
New York City, New York City, not necessarily New York State. A lot of people want to know what's the roadblock there upon the expression and what's the timeline. You know, you work closely with the authorities, but that is a big potential market.
Yeah, that's a market where you know, we're just going to have to do the work and demonstrate our safety outcomes and earn the trust and shizzle away at it over time.
Do they have the rules for you to follow?
They do not have rules that allow the human operator to be removed from vehicle entirely.
And until that changes, and until that changes.
But you know, there is an interest in doing this in the state, even outside of the city, and that gives us an opportunity to grow more fans, and fans actually are calling for this in cities where our technology can't be deployed. We are seeing organic campaigns spring up saying I want WEIMO in my town.
San Francisco. The Bay Area is my home. London's where I grew up. And I was studying the map of the burros that you propose to launch in and you correct me if my math is wrong, but just based on those burroughs at launch, this seems to be the biggest citywide deployment from the start that you guys will have done. Is that correct?
It?
Possibly, it's correct.
I mean, we're in the phases of figuring out the specifics around the launch and figuring out the actual framework around the launch, and so I don't want to speak to definitive word about what we're going to do. But what you're speaking to is we're not gated by the technology, right, and we have the appetite to scale, and we want to partner and do so safely, and we want to
earn trust. So that's like there's a lot of levers there that we have to figure out how to strike the right balance and how to make sure that we're introducing it to the community both to meet the demand and a grow I.
Asked, because of those twenty cities to come, London is one, yes, and London is now outside of the European Union, but it's you know, it's kind of your europe launch. Yes, what was that experience like? Within London's regulatory framework and the UK's regulatory framework.
They've been extremely forward leaning and interested in seeing how this technology could actually improve safety.
On their roadways.
And that's where I think, you know, we find a sweet spot when people are less sort of complacent about the status quo.
Weimo co CEO to Keedra Mawakana ed. I mean, we've got to check out the full interview. It's on Bloomberg or as a special edition of Bloomberg Tech Podcast. But really your conversation is about this extrawnay expansion into London and just what pace of growth, what signals you discerning as to how big it can be there.
Now you and I have both called London home, right and we've taken the bus and the tube and gone on bikes, and so how do they get past what's a really established transit network and right now there are several dozen Waymo vehicles testing in London. They are doing that in more than thirty of the sixty London boroughs, you know, like sixty percent of the city's already covered. That's way beyond what, for example, Tesla's doing in Austin right even the Bay Area here in terms of their
scale on footprint. So there's a big expectation that if this labor government is doing okay and passes the rules, then they can do this for real this year, no safety driver or supervisor and have a commercial service at scale that they haven't done that quickly in the United States.
Look, this is about disruption of transport, and it looks as though transport and logistics is in the either storm in terms of disruption today as well. If you're looking at the markets now, you and I have just been pouring over some of these market moves extraordinary for c H. Robinson down by twenty one percent, but you're actually seeing the Dow Transports the index having its worst day since April of twenty twenty five, currently down by five percent.
Now traders are telling Bloomberg Kuneneagel, for example, which is another European logistics company, traders are citing an announcement coming from Algorithm Now.
Algorithm is a.
US traded company and they've announced that as a leading AI technology company. They've published a new white paper demonstrating its semi cab platform reduces empty freight miles by more than seventy percent across active customer networks. Is this the next aishooter drop ed?
It was a severe sharp drop and we got to keep looking into it. Coming up from the program Back to Ai similarly secures one hundred million dollars in new funding as it looks to scale predictive AI mimics human behavior. We have more on that next. This is Bloomberg Tech. AIS Startups similarly has raised one hundred million dollars in a new funding round led by Index Ventures. Similarly as an AI lab, the wantes to help companies predict human behavior,
in fact a chain of human behaviors for more. June Park similarly CEO and one of the founding group is with us in San Francisco. This is very interesting because the field has tried to develop a model that is a behavior model, and it has had mixed success some failure. Let's start by talking about the data sets that the model's trained on. How did you develop what you think is a model that does give you an insight or a prediction of how a human might behave? Absolutely, thank
you for having me. Glad to be here.
So similarly is a company that are spinning out of Stanford where we led frontier research on creating agents and simulations that leverages generate to BAI technology. So today we partner with millions of real people and combine our data collection strategy with our modeling strategy to create agents that can actually behave in a much more generalizable context. So data here we might leverage include interview data that's unstructured, which is the kind of data in the past was very hard to.
Leverage literally people telling you their life stories. That's exactly right.
So we actually create ai interviewer that can go talk to people one on one, voice to voice and actually get life stories and their preferences around different policies and so forth with their consent.
And to what end I mean to build simulations that are one hundred percent accurate to then populate with agents tune.
How do you see this work?
You've done proving more disruption to the market right now, right.
So today we work Fortune ten companies in retailer businesses, to personal finance, to CpG companies and even to polling companies.
And today our customers use our technology to do concept testing. So, for instance, CBS has been a close partner for the past five months where CBS has now created hundreds of thousands of similis or quote unquote agents that represent their real customers, so they can actually do simulated focus groups to understand better and there's in their pain points and their use cases, but also to do story layout designs
and so forth. Galop, another polling company that's highly well established and respected, is now working with similar to create digital panel of their agents or their simulated human panels. Now, one last example that's interesting is, of course many of our customers are Fortune five hundred companies that regularly have earnings call, so use case there might actually be creating simulation of their earnings call. So this is an ass
that we get frequently. So in the recent earnings call we simulated we actually can predict eight out of ten questions that are actually asked.
In this court.
Sorry, June, you're saying the analysts almost don't need to be asking that question. You can predict with eighty seventy what analysts are going to ask CEO CFOs on earning course.
That's exactly right.
In fact, I actually before our conversation today, I had a chance to simulate the questions that twelve you might ask. And these are some other questions that actually have come up in our assimilation.
Oh so you seem well prepared. Look, there's a degree of skepticism about Similarly, to be fair, the founding members four of you are are largely academics at Stamford. You founded the company officially a year ago, but you've only been training the model for seven months. You say you have five months of commercial deals. How have you done that so quickly? So similarly, it is a.
Real combination of amazing front your researchers that lasers the ceiling of what the models are capable of today, but also amazing product and engineers.
So here on the product.
Side, we have people like Laney Allen and Mikha Kapor who have been product and engineering leader in the space of companies such as Figma and Heavier, and that combined with the rest of the engineering team and the research talent our neighbors, us to create frontier technology that actually gets routed directly.
To product June Park, who has preempted all of our questions through his own products of Civilly. We so appreciate your time, Thank you very much. Indeed, let's just take a look elsewhere in AI model world, because Alphabet currently spiking higher, up almost a percentage point. This is they've updated a release to Gemini three Gemini three Deep Thinking Reasoning Model. It's an update and they say that the mode hasn't been updated to solve modern science research challenges in particular.
Ed.
Look, Alphabet has been on a tear because of the idea that it's generative. AI is managing to leap frog or at least push ahead some of the rivals.
Yeah, it has, and look at the spike in the session on those headlines are coming up Coinbase releases earning after the closing bell, We're going to preview what to expect. This is Blomberg Tech, Asian delivery firm GRAB missed analysts expectations on its full year forecast to sign that weaker consumer sentiment is weighing on the company. Earlier, we spoke with GRAB CFO Peter Owe. Listen to this.
It was an amazing year for us. We achieved some new mastones in the business. We have now fifty million monthly transacting users using our app constantly on a monthly basis.
We also showed great profitability.
In the business to over five hundred million dollars in the Justicey you be done, but also our first YEARNIT profit in the business. As we look at twenty twenty six, we're firing on aus cylinders and we are seeing good growth across all the platforms, across all the different products
that we have today. We are confident going into twenty twenty six in terms of delivering that new guide that we gave up for twenty twenty six, but also we also gave a three year guidance at the same time of twenty percent revenue growth from twenty twenty five through twenty twenty eight.
So really long term perspective here.
Just dialing in on the profitability side, because you're right many loving the fact that you surprised you delivered this full year profit. What now to really drive up that return on equity? Our Bloomberg Intelligence analyst talking about maybe it's the net profit.
To really be driven higher.
Needs even more leaning into the fintech side of the business or some cost cuts.
Where do you pull.
Yeah, there's really three big pillars when you look at the business today. One is continuing the growth momentum in our rights business as well as our deliveries business. And our deliveries also now we have other products and services up. One example at grocery delivery continues to be one of our fastest growing it's actually growing one point seven times faster.
Than our food business today.
We also have dining in which is another unique set of product features that brings people the restaurants.
Also, at the same time, financial service is.
Also condused to be our fastest growing segment about business today. The loan box size that we have today, we've clicked over a billion dollars in loan portfolio and we expect to double that is we finish the year twenty twenty six. So we've got all the different products working together to really drive the growth in the missus but also margin expansion at the same.
Time, delivery to fintech grabs CFO Peter Ue there talking of fintech. Coinbase was about to report earnings after the bell and this comes in the context of a deepening crypto route which has seen the value of bitcoin in particular, plunge look more than forty five percent from its peak, and with that, expectations of coinbases revenue are not looking pretty Bloomberg Senior crypto reporter Olga Caraff.
Joins us now.
And it's all about volumes, right, and they're going to pull back, so inherently revenues must have been hit.
Absolutely, So what happens when prices drop continuously? A lot of the retail traders stay on the sidelines, and of course crypto exchanges they make a huge portion of their revenues from transaction fees, so they see a huge drop on that.
What I understand or I get the volumes concerned, But I've always understood that like coinbase others that it competes with, they benefit from volatility, right because volatility you have to trade in one direction or another. Why is that not a good moment for them?
So you're absolutely right, when there is high volatility, they do benefit. But typically what happens is that you have say several days of very high volatility, such as sort of in October when bitcoin dropped a lot, or in
early February when again it dropped a lot. So you can have a lot of volatility, a lot of trading during those days, but then you can have a sort of slower activity, lower activity in the days kind of in between those very high volatility days, and that can impact exchanges like for instance, we've already seen Robinhood report.
That is a fourth quarter crypto revenue.
Was down thirty percent, So things are not looking good for crypto trading providers.
Briefly, Coinbase has tried to diversify ahead of this. They're looking whether or not it's predictions markets, whether it's equity trading alga.
Is that enough?
So Coinbase has been trying to diversify for many years and these are really good initiatives that could really help in the years ahead. But right now, for instance, they just launched prediction markets, so it's a very.
Very small revenue driver.
They do have stable coin revenue coming from their revenue share with Circle that is helping to stabilize revenues quite a bit, but a.
Lot of the the revenue portions are.
Still all Bloomberg's ald creed. Thank you very much. That does it carry for this edition of Bloomberg Tech