Tech Sell Off Continues as Investors Question AI ROI

24:04So we need to make sure there's an ROI component to it. But coming out of it, be able to see what comes out and fail fast if it seems like a given process doesn't lend itself appropriately. So I go to the sensitive question of those employees that are embracing these coding agents. Yeah. Do you hire fewer of them? How are they feeling about being augmented or replaced? Well, I would say it actually allows us to get to a backlog we could never get to. There's always more work than you can get to. There's always things that you could modernize. There's always more things to do. Our first reaction from our teammates has been this frees them to do more innovative work rather than the mundane work that was consuming their time and resources.

24:40And I think we're going to see that pattern play itself out over time. Certainly. Only what, a few weeks into the leading job. It's wonderful to speak with you, Hari. Hari Gopal Krishnan, he's the Chief Technology and Information Officer. over at Bank of America.

24:59Time now for Talking Tech. And first up, Xiaomi says it's looking to sell its first EVs in Europe by 2027, taking on the likes of Tesla and BYD. Now, the company announced its strategy after posting a 31 % rise in quarterly revenue, boosted by the launch of its second EV. Now, according to the Xiaomi president, the company is doing research and preparation, but has no specific product plan yet. Plus, shares of Baidu, well, they're lower today. The company posting second quarter revenue that slips slightly as it faces an economic downturn in China. Now, AI rivals and difficulties in net growth areas in particular.

25:32China's internet search is looking to Gen.AI to drive future growth, but is struggling as open source models like DeepSeek gain traction. And startup GameScience has unveiled its working on a successor to one of last year's best-selling RPG titles. Now, at Gamescom conference, it was over in Germany, the company showed off a teaser trailer to Black Myth, Sean Cuey, a sequel to 2024's Black Myth, Wukong. Now, according to game science backer Tencent, the project is, quote, little more than an empty folder at its current stage. Coming up, Meta making those restructuring moves of its AI group. They'll have the details next.

26:12But first, let's check in on these markets. We are down, and we're down significantly. The Nasdaq 100 really accelerating yesterday's losses. We bounce off our lows, but still we're seeing some of the worst sell-offs that we've seen in a few months. We're also seeing names like Palantir really eroding some of that valuation that we've seen, having its worst couple of days in three years. Other names such as Apple and contributing to the points perspective. This is Bloomberg Tech.

26:42Welcome back to Bloomberg Tech. Let's take a look at these markets. We're in sell-off mode. We are bouncing off our lows. nevertheless we're up by 2.4 % in the last two days. We're a sell-off since about April, in fact. We have some deep anxiety surrounding valuations, particularly in the AI trade, whether it's macro concerns, as we anticipate Jay Powell and an awful lot of Fed talk, what rates mean, but also, more broadly, what these valuations have got to extreme levels really mean when you've got, perhaps, enterprise not fully adopting in the productivity means that some anticipated, the MIT research report causing some gyrations.

27:16Let's move on to the individual movers that we keep an eye on. Look, I'm off by overall in the last few days, well, 5.5 % for Palantir, but worst couple of days sell-off in a few years. Intel is off by more than 6%, wiping out a lot of yesterday's gains after we were enthusiasm around SoftBank investing in this company and indeed talk of the US government taking at least a 10 % stake. But we pull back on Intel. As people really lock in profits on Meta as well, we're off by 1.4%. It's down for three straight trading days, having lost 5.6 % in the course of those three days. Let's just stick with Meta more broadly.

27:50And the company announcing yesterday yet another restructuring of its AI group pursuing superintelligence. Now, it will be splitting its AI team into four distinct teams. A move the company says will allow it to better capitalize on its recently acquired talent. Bloomberg's tech reporter, Riley Griffin, is here to break it all down. You're in San Francisco really digesting what seems to be a focus on LLMs, still on the frontier research. But then there's the application of this research and the infrastructure, of course. No doubt. Yesterday we were able to obtain a memo that was shared by Alexander Wang, the new chief of MetaSuperintelligence Labs.

28:28And he's really laying out the new structure for this organization after weeks and weeks of an incredible effort to hire talent. As you might remember, Caroline, this is hundreds of millions of dollars, sometimes nearing a billion dollar in total compensation for individual AI researchers. And now they are finding their home within Meta, within these four teams. Yeah, and the home. How are you seeing the, perhaps what some have reported behind the scenes, of some jostling or discomfort with what had been suddenly incredibly expensive talent coming in and a reorientation of where the focus is? Does this help better align that talent and actually push forward?

29:08And ultimately, is it open large language models, for example? Are they going to delve into closed ones, potentially? That is a really burning question for all of us. What we do know is that one team will continue to be dedicated to large language models. The LLAMA product that underpins Meta's AI efforts broadly. And the one team that I really am keeping an eye on is the infrastructure team. Mark Zuckerberg has pledged to spend hundreds of billions of dollars specifically on AI infrastructure. And now we are seeing talent from within the company go and lead that specific effort. We expect more capital to be deployed there this year and then certainly next.

29:49Some great reporting. Riley Griffin, we so appreciate it. Let's get some of the investor and analysis take now. Rohit Kulkarni is with us, Managing Director and Senior Analyst covering internet and capital markets research at Roth Capital Partners. And that commitment to infrastructure that Riley articulates there, that money, is it proving out? Hey, thanks for having me. I think so far, what we have seen looking backwards last 12, perhaps 15 months, Meta is probably the best poster child in providing kind of quantifiable returns on investment in AI infrastructure. By that, what I mean is the company has demonstrated accelerating revenues.

30:29The company has demonstrated upside to industry growth and a lot of other bells and whistles around AI investment. So looking backwards, yes, definitely. AI is the poster child of showing return on investment. Looking forward, that's the question mark in the stock today. Okay, so they've been able to make it work for advertising in the business model. But now for superintelligence and the race to be won, what business model does that align itself with for meta? How much are we going to see generative AI affect the way in which we interact and communicate? I believe the problems that Meta is trying to solve using AI and using superintelligence are just compounding with the data that they're collecting and the possible business models that they are exploring.

31:18Remember, they have invested a lot of money in the augmented reality, virtual reality glasses. They have been working on that over the last four years. There has been a reset in that. And now probably we are not now looking at 2027 for the actual launch. So there is a timetable to build AI within those glasses. I believe that's going to take up a lot of money. And second, integrating the three apps that they have. And those are the three largest mobile apps in the world with billions of people using on a daily basis. To make those apps super intelligent and to make them extremely sticky is probably a monumental investment opportunity for them.

31:56So these two are orthogonal investment opportunities, but I think both are probably going to be something that Zuckerberg is going to be fully invested in. At$1.87 trillion, is that a full valuation for this full investment theme that you have, Rohit? I believe the layers unpacking in the future, it's just going to be something that we are going to be excited about. I feel we have still not yet scratched the surface of the potential of Gen.AI on a very large multidimensional platform like Meta. And I feel Zuckerberg has the founder whip, if you feel. He's investing behind this as if it's an existential once-in-a-lifetime opportunity, even bigger than when iPhone was launched.

32:48So I feel the excitement and the kind of attitude that they can disrupt the apple cart while redoing something new is something that investors are willing to be patient around on Metastalk. And so are we. You're patient. You've still got a buy rating. Compare and contrast to some of the other key companies that you do analyze. When I'm thinking that you're looking at Alphabet and Amazon, are they as bigger winners? Have they been as bigger winners in the capex they've already spent? and do they have as good a roadmap as you currently articulate for Meta? So far, I think Meta is the clear winner.

33:26Amazon is so far probably a number three, and Google is a number two winner as far as demonstrating returns on Gen.AI investments. Having said that, I think the complex problems that each of these very large, massive companies are solving or trying to solve using AI are essentially compounding with every day. And that is leading to more investment and more investment needed to solve those problems. Think of the multidimensional e-commerce, cloud, advertising, prime video, same-day delivery. All of these elements can be optimized using AI. And at any given time, probably Amazon is focused on maybe one or two of those.

34:10So I feel kind of over the longer arc, the bigger winner is probably going to be a company like Amazon, given the surface area of improvement they have, the potential. But in the near term, probably a one-trick pony like Facebook has executed better in the last 18 months, and they are demonstrating better returns today. I hate to talk about the near term, because I know you're a man who likes to give longer term projections and price targets. But when you're seeing a sell-off like we've had after the last couple of days, a lot of it tied up with worries about ROAI and the report coming out of MIT, the Gen AI divide.

34:46and the state of AI in business. Do you give credence to that? Do you think there's a lack of nuance? Or do you think actually we should be worried about how much we're investing in the returns on it? I think there are two ways to interpret MIT and just the overall questions that are being raised that, hey, by the way, we applied Gen AI algorithms, LLMs, we invested so much, but we are not getting the returns as much as we wanted. I think the way these large companies are trying to solve really large problems, the efficiency gains that they will provide over the next 12 months are something that we are very confident on.

35:24Whether that translates into revenue uplift, that remains to be seen. Only a company like Meta has demonstrated that other guys are kind of TBD. That's why from AI winner standpoint, Meta is the one to buy on this weakness today. Rohit Kulkani, Managing Director and Senior Analyst at Roth Capital Partners. It's great to catch up with you. This is Bloomberg Tech, and you're looking at a live shot of the principal room. Bloomberg is live at Jackson Hole this week. Tune in for continuing coverage leading up to a special episode of Bloomberg Surveillance on Friday. This is Bloomberg.

36:06Elyse AI, which automates complex healthcare housing systems, has just announced a$250 million Series E round, led by Andreessen Horowitz. The startup, which was founded back in 2017, offers conversational AI, a platform for users. And I'm pleased to say that the CEO and founder, at least AI, Mina Song, is here with me. You're based in New York. And just how was it raising these funds? How quickly are you able to raise such an enormous amount of money? Thank you for having me. It was pretty quick, actually, to raise the round of funding. I think investors are incredibly excited about investing in vertical AI, which is AI that solves specific problems for specific industries, which is exactly what we do, particularly focused in the housing and healthcare industries.

36:50And why those particular industries? Was it just the pain points were just so obvious and glaring to you? I think so. It's pretty obvious to all of us how inefficient and how limited the technology is when we interact with these industries. But the core mission of the company has always been to develop advanced technology for industries that serve fundamental human needs. So housing and healthcare are some of the most important decisions we make and the most basic needs that we have as humans. So the way in which the business works is you're selling your conversational AI in a B2B manner, but then it's me that ends up experiencing it as a consumer.

37:26How have you seen healthcare institutions, housing institutions' willingness to purchase these sorts of products thus far? Yeah, they're extremely excited to use it. So we serve, on the housing side, one in every eight apartments in the U.S. today are using our AI software. And we're using AI to automate manual processes to make operations more efficient and to make them more cost-effective. And these inefficiencies really affect all of us. When an industry is inefficient, then that drives up costs for owners, which ultimately trickles down to renters, which is really important and really affects all of our lives.

38:05So there's a lot of excitement about AI. Okay, and so with the money, you do what? A significant amount of the funds are going to be going to expanding our teams. So a lot of engineering team members, product team members, research, operations, and more. We're expanding our offices in New York City, San Francisco, Chicago, and Boston. So we're looking for people who want to have real-world impact and solve really important problems that impact millions of people's lives. We've talked a lot about talent wars and how expensive certain researchers have become. Is it the mission that therefore attracts them, or have you had to up the paycheck significantly and indeed the reward of equity?

38:44Absolutely the mission. I think people who want to apply these great advancements in artificial intelligence but want to see that that has real-world impact and that it can affect all of us is what attracts people to working at Elise AI. How do you tend to offer when you think about packages? Are people coming on board wanting to see a roadmap for you to go public? I mean, you're already at Series E stage. We're focused on building a generational company. That's what's really important to us. We want to change the face of housing and healthcare and want people to look back on it and say, I can't imagine what it was like back when all of these services were so expensive.

39:24Why were they like that? and people who are joining want to be part of making that change happen. How do you calm any anxiety that you're not going to do a windsurf? You're not going to do suddenly a big purchase where you as the founder get lifted, Alexander Wang style, but leave others behind? How do you ensure that this idea that you take the team with you remains resilient? I think we've been through and through dedicated to the mission of the company and there's so much more that we all want to be a part of to make this happen and it's a really exciting journey to be on. We have no plans to do that.

39:58Well, thus far, it seems as though the clients have loved you and speak very highly about the impact that your technology is having serving people in healthcare, people in housing. Mina Song, founder and CEO of Elyse AI, on that fundraise, we congratulate her. And meanwhile, soaring valuations are coming from the likes also of the major foundational models. Think OpenAI, think Anthropic, but also Databricks. They've also been boosting. But they're creating perhaps a bit of a false sense of strength among the VC markets. In reality, only the top 1 % of companies are pulling in the majority of capital.

40:28And now more investors are beginning to worry about the health of the young Silicon Valley startups. I mean, most Kate Clark joins us for more. And really, you've been uncovering, as you always do within tech and venture capital, some of the trends. So these big whopping valuations cover up a darker side? Yeah, there's a tale of two markets in venture capital right now. You have the AI mega companies, OpenAI, Anthropic, Databricks, raising rounds in what feels like in a monthly cadence. And then you have all of the other startups, and many of them are actually struggling or even raising down rounds.

41:00So who are those types of startups? Because we just had Elise AI on. They've raised$250 million, relatively young startup, and actually managing to, well, raise. But who are those who have been left behind, particularly those perhaps who did well in 2021 and don't really want to admit that they're going to have to face a down round? I mean, it's exactly companies like that. A lot of the unicorns that were minted in 2021, a lot of the companies that were founded before 2021. The VCs like to call these sort of the non-AI native companies that are maybe trying to catch up, trying to incorporate AI.

41:28But really, they were born before ChatGPT. Born before ChatGPT. Some of them have made the pivot. Like I think of a Figma, for example, managed to then IPO at a valuation that they were previously going to be bought at. and people have really managed to think of them as a generative AI business even though they've integrated it as they grew. Where are some of the pitfalls? Where are the industry exposures that we're not liking right now? I mean, I think you have to look at e-commerce, fintech, cryptocurrency. There are many areas in which investors are not as excited because they don't see how generative AI can really help that business.

42:00I think, though, the main point of weakness are the companies that raised at extremely high valuations during the last hype cycle and investors now just don't see a path to allowing them to raise money at an even higher price. So those companies are kind of becoming what, again, VCs call zombie unicorns and are just sort of slowly, slowly dying. Will they be allowed to die? Will they be forced to consolidate? What happens, Kate? I think there'll be some consolidation, but I do think you'll sadly see a lot of just companies quietly disappearing and slowly dying over time. But of course, there'll be M &A, and of course there will be winners.

42:36There will be companies like Figma and many others that were born pre-Jet-DBT and still find success. And you'll be there to document it all, Kate Clark. We appreciate it. So there's a lot of noise about AI, but time's too tight for more promises. So let's talk about results. At IBM, we work with our employees to integrate technology right into the systems they need. Now, a global workforce of 300 ,000 can use AI to fill their HR questions, resolving 94 % of common questions. Not noise. Proof of how we can help companies get smarter by putting AI where it actually pays off. Deep in the work that moves the business.

43:13Let's create smarter business. IBM. Whether you're planning a big tech event, launching a new campaign, or just stocking up on team gear, finding the right promotional products makes all the difference. 4imprint offers thousands of options. From on-trend apparel and premium drinkware, to tech, totes, and giveaways, so you can find the right fit for any audience, purpose, or budget. You can customize it all, your logo, your message, your look, and many items come with no setup charge to help you save. And if you're really watching the bottom line, you'll find standout choices at every price point so you can make a real impact while staying on budget.

43:48Plus, you'll get expert help, fast turnaround times, and their 360-degree guarantee. So you can be 4imprint certain your order will arrive on time and look exactly right. Whatever your goal, 4imprint makes it easy to find your perfect promo match. Explore the possibilities today at 4imprint.com. 4imprint. 4certain. Let's talk about healthcare for a second. It doesn't always work the way people expect it to. If you've ever waited on a prescription refill or had a hard time getting the care you needed, you know the feeling. The system should just work better for everyone. That's exactly what the people at Optum are trying to do every day.

44:26They're a healthcare company linking patient care and pharmacy services and using data and technology to drive the whole system so care is connected, not complicated, for patients and providers. Things like making it easier to get care that looks at the whole person, from primary care doctors to mental health support and even in-home care, and then using technology to make sure they all work together. Technology designed to help doctors spend less time on busy work and more time with their patients. And those prescriptions? Optum is working to bring costs down, save patients money, and make it easier to get refills.

45:00Little by little, Optum is helping make healthcare work as one for everyone. Head to business.optum.com to see how.

45:14Google, well, it's set to show off new Pixel hardware later today at its Made by Google event. Just bring in Manip Singh, who heads tech research at Bloomberg Intelligence, and talk about, well, how important is the hardware for Alphabet, the parent company of Google? I mean, everyone realizes Google has an advantage on the custom chip side because of their TPUs. I think that's going to show up in repeatedly, like they release Pixel every year, so you will see a much better Pixel version that can run LLMs natively. That's the one thing everyone wants Apple to do. I think Google is already doing that.

45:50Obviously, on-device LLMs is not that big of an opportunity right now, But look, whatever people think in terms of AI agents, to me, AI agent opportunity will come through on-device LLMs. And yes, there will be apps that run AI LLMs. Browser is very important. I mean, in the case of Google, they have the Chrome browser. For now, we know. For now. And then layering it on operating system, I think, gives them a unique advantage along with the chips. So that's what I feel they will emphasize at the event. that this asset is incredibly useful to them and really only useful to them and would be wither and die on the vine if sold to another.

46:32But compare and contrast how important it is to show off it within their hardware versus, say, a Samsung, which was also infused with Gemini, too. Yeah, and look, Android operating system is used by a lot of other OEMs. So from that perspective, what Google has done well is to showcase AI capabilities on all kinds of hardware. I mean, even the earbuds that they will showcase, I expect, at the event will have some integration of AI into variables because that's what they want to show. At the end of the day, your LLMs right now will be distilled into smaller versions that can do intelligent tasks, whether it's scheduling or voice assistance.

47:13and that's where Google has a unique advantage because it has the chips, it has the LLM, it has all the data, and it has the family of apps with over 2 billion plus users. And that is why it knows so much that it can do personalized AI that others can't. So to my mind, they're in a unique position. Obviously, they have a small market share when it comes to smartphones, but they're doing enough to showcase that embedded AI element, which could potentially drive up their share when it comes to the hardware side. Yeah, briefly, how much could they steal from a lackluster Siri offering? Well, I mean, look, Apple has their problems.

47:52They have to figure out a way to integrate LLMs, whether through partnerships or through their own. I think it will be through partnerships. In the case of Google, they will use all the OEMs. So really, the advancements will be at the operating system layer, but they have to showcase that in their own hardware first. So that's why this event is quite important. So too is as we await in anticipation of Judge Meta's ruling on whether or not they're able to keep Chrome as well. Bloomberg Intelligence, Mandeep Singh, we love having him as always as we anticipate some of that hardware coming out. But that does it for this edition of Bloomberg Tech.

48:25One last check on these markets before we leave because it's been a volatile hour of trading during this show. We're off by one and a quarter percent on the NASDAQ 100. Sox is feeling the pain. NVIDIA is down significantly. We're off by 2 % on the semiconductor index. Crypto, though, turning back into the red, there is risk and anxiety building about some of the valuations in the market. We see some profit taking. Don't forget, though, to check out our podcast. You can find it on the terminal as well as you do online on Apple, Spotify, and iHeart. From New York, this is Bloomberg Tech.

49:31Thank you. helping make healthcare work as one for everyone. Learn more at business.optum.com. Whatever your goal, trade show giveaways, client gifts, or team gear, 4imprint has the promo products to match. With thousands of options, from apparel and drink wear to tech and totes, it's easy to find the right fit for your brand and budget, with standout choices at every price point. And with their 360-degree guarantee, you can be 4imprint certain your order will show up just right, right on time. Explore more at 4imprint.com. 4imprint. 4certain.

50:33Michigan. Pure opportunity. Seize your opportunity at michiganbusiness.org.