Microsoft Points to Chinese Hackers in SharePoint Breach

0:00The most effective people at work aren't working harder than everyone else. They're working smarter, inside better systems. Superhuman Go, from the makers of Grammarly, is the AI assistant that works inside every tool you already use, always ready, already aware of what you're working on. It's a teammate whose only job is to help you be better at yours. With Go working with you, you can show off what you do best. See what Superhuman Go can do at superhuman.com. Some people treat ChatGPT like some kind of smart search engine, and some use it to get work done. ChatGPT Work is a new way of working in ChatGPT that can take action across your apps and files, stay with a project for hours if needed, and turn a goal into finished work.

0:46It's designed to help you move from a chaotic starting point to a reviewable first version. So all the source materials, briefs, and scattered information that you have to grind through to turn into something useful can just become something useful. Put ChatGPT to work on your most ambitious ideas and projects. Get started at ChatGPT.com by selecting Work Mode, available on Plus and Pro plans. Healthcare doesn't always work great. If you've ever waited on a refill or couldn't schedule an appointment, you get it. That's the kind of stuff Optum is changing. They're using data and technology to integrate patient care, pharmacy, and everything else.

1:25So healthcare is connected, not complicated. What's that look like? Cheaper prescriptions that are easier to get and care that looks at the whole person. How you need it. Optum is helping make healthcare work as one for everyone. Learn more at business.optum.com. Bloomberg Audio Studios. Podcasts. Radio. News.

1:53Bloomberg Tech is live from coast to coast with Caroline Hyde in New York and Ed Lovellow in San Francisco. This is Bloomberg Tech. Coming up, Microsoft accuses Chinese hackers of exploiting vulnerabilities in its SharePoint software that have led to breaches worldwide. Plus, OpenAI and Oracle deepen their Stargate data center partnership to power AI workloads. But no sign of SoftBank to help finance the US infrastructure. And NXP gives a not-so-bullish third-quarter forecast, suggesting the chip-makers still contending with a turbulent industry in the face of tariffs. Microsoft's our top story in a blog post saying that state-sponsored Chinese hackers are behind what's happening, and in particular, the stock reacting negatively early in the session, Caro, normalising a little bit.

2:43But we've got to get into it. We do. Let's stick with that Microsoft story. Bloomberg's Jake Blyberg joins us for the latest. And these are known state actors, two of them, and then one Chinese-based actor, it would seem. Yes, that's exactly right. Microsoft said this morning that two groups backed by the Chinese government, Silk Linen Typhoon and another group, have been responsible for some of the exploitation of this vulnerability in its SharePoint software. We've also heard from other cybersecurity firms that they're seeing multiple different organizations sort of hack into businesses and governments based on this vulnerability.

3:23Jake, the top priority right now seems to be identifying the victims. Do we know what the kind of net result of this hacking activity is? Who's been impacted? We know that the impacted organizations are ones that hosted SharePoint on their own servers. and that seems to be a real smattering of different groups, government agencies, businesses, other organizations all over the globe. We know that governments have been hit in the Middle East and in Europe and here in the United States. We also know of businesses and state-level agencies that have also been hacked into. Microsoft, for its part, say investigations into other actors are also ongoing when it comes to the exploits.

4:07We can currently see what they've put out in a blog post. But more broadly, they acted swiftly to try and patch the issue. But once in a server, that is the key concern. Exactly. We've seen warnings from cybersecurity companies that you don't just need to patch. You need to sort of hunt for whether your systems were penetrated and what they might have been taken. A source told Bloomberg yesterday that the hackers once getting in, in some instances, are stealing login credentials, usernames, passwords, tokens. And that suggests that they may be trying to exploit that information to launch other attacks.

4:46Bloomberg's Jake Blyberg out of New York City. Thank you very much. Meanwhile, shares of NXP Semiconductor are trading lower today. The company posting third quarter forecasts that were less bullish than investors anticipated. Then they went on to talk about the state of play in the automotive industry, which is 50 % or more of their revenue. Bloomberg's Ian King, who leads our chip coverage, is here. It's actually a pretty simple story, right? away from all the AI stuff. There's been a glut in inventory on automotive chips, NXP saying, look, it takes time to work through them. But for that industry and that key customer, the world's a bit uncertain as well.

5:20It absolutely is. And we'll hear from Texas Instruments in a similar vein later today. The cross currents are everything's supposed to be getting better. This kind of glut that we've had over the last couple of years is supposed to be going away. So we want, you know, if you're an investor, you want them to be more bullish. You want them to be a bit more reassuring. That's not what happened from NXP. They said things are still hanging around. And Ian, almost bullish is the wrong word for it, when revenue is down 6 % for NXP. And actually revenues have been down quarter after quarter, right? Yeah, I mean, what we'll have to look at, and this is why I think it's going to be useful to have Texas Instruments reporting today sort of back-to-back, is this a customer-related thing?

6:02Is NXP basically supplying automotive companies that are struggling more than others? Texas Instruments is a much broader supplier, has a bigger customer list and a bigger set of products. So that'll give us a sense of whether this is a general automotive malaise, whether this is China related, whether this is Europe related, and we'll have a much clearer picture later today. I am going to dig into this a bit later in the program, but the Philadelphia Semiconductor Index is underperforming generally, and NVIDIA and AMD are big drags on that. the read through between nxp and those two is difficult for me to see but generally speaking away from those kind of ultra specialized and even analog chips things are quite robust still yeah i mean we've seen i mean you talk about the index the index has done pretty well this year there's a lot of money has been put to work everything we've heard so far and we'll hear more on that vein as the earnings season unfolds is there's a lot of spending still going on on these large AI infrastructure systems, and we've heard nothing that's telling us that that's not going to continue.

7:08And that's really what's going to be the most important thing. And that's what's driven the SOX so far. Analog has been a slightly different story. But again, analog benefits from this server build out as well. You need these power-related chips. You need these microcontrollers as well. Ian King, always with all the context. So appreciate it. Thank you. Let's dive deeper into the tech earnings, the vibe in the market right now. Anna Rathman is with us, founder and CEO of Grenadilla Advisory. Anna, your take here is that, yes, we're perhaps pulling back on the day somewhat. NXP, not a resounding feeling of optimism.

7:43But we're still near all-time highs, and we've still got a$4 trillion Nvidia on our hands. Yeah, good morning. Markets are definitely, the valuations are high. There's no slicing that in a different way. But I guess, you know, you have to look at tech differently than we used to. So a lot of allocators and wealth advisors out there, we've been taught to look at tech as a growth sector. Perhaps it's more than that. So the AI component is definitely tech. But maybe there's also a quality factor there as well. Got healthy balance sheets, great cash flow, especially those mega companies, mega cap companies.

8:20It's also got a defensive sector characteristic as well. We learned that we cannot operate this economy and our society without tech infrastructure as well as software. So I think that when we're looking at tech, we have to look at it from multiple angles in order to perhaps try to explain some of this valuation. It's not just about AI fraud. It's much more than that. But what's been interesting is the AI winners have powered forward on 2025 and those winners have been deemed, well, the metas of this world, invidias of this world, Apple left for dust. In some ways, Alphabet has been a laggard too.

8:58How do we set up into Wednesday when we start to get a couple of those Mag7 names? Yeah, I think there's definitely a bifurcation in those Mac 7 names. And right now, it looks like the people who have really invested and gone forward with AIs are definitely winning. Now, Alphabet, for example, will have different challenges, right? Not only do they have some legal challenges, but, you know, I don't know about you, but when I talk to my friends, they've stopped using Google as a search engine. They're using ChatGPT. And I know that's just a few people around me and perhaps you, but those things can actually catch up.

9:32And I know Gemini is improving, but I think that's a real competitor there. And so AI is going to be more important going forward for some of these MAC7 names. And I wouldn't expect disappointing report from them, but certainly I think investors can read between the lines. Anna, what's curious about this week is that on Wednesday, Alphabet and Tesla post their earnings results, right? Somewhat analogous companies, some differences. And on the same day, we'll hear from the president about his vision for AI and American AI. It's going to be a wild ride for investors this week, isn't it? Well, you know, anytime anyone from the government talks about technology and AI and even Bitcoin or blockchain, I get a little nervous because I think we have to tread very carefully here.

10:27We don't want to get in the way of innovation, and yet we want to protect consumers, right? So it's going to be a balancing act for anybody from the government, from the White House, talking about AI and how America is going to be involved in the development of AI. I would go in with a little bit of a cautious view. And frankly, I'm expecting a lot of lofty statements. But the question always is, how does this actually translate down to the day-to-day in active software and make difference in our productivity? Let's bring it back to today's news. What did you learn about the near term and quantifiable impact of tariffs?

11:10You had names like General Motors and NXP show us that actually something's happening here. Yeah, I mean, something is happening there. We're talking about actual money, right? I mean, so for last week, if you think about the CPI and consumer sentiment, you didn't really see it, right? But you are seeing it in earnings. This is actual cash flow impact. So I think this is going to trickle in as the quarter goes and give us a better view of exactly what areas of the economy are being impacted by tariffs. Who is actually assuming some of those costs? Because GM actually came out and said, we don't want to pass down all the costs to the consumers.

11:48So it'll be interesting to see how those things get, I guess, diversified into the supply chain and how GM and other companies actually manage those challenges. And going back to tariffs, tariffs is all about the geopolitics that is currently involved in technology investing. I think of the geopolitics between the U.S. and China. I think of the latest news out of Microsoft blaming China for a hack. How much do you factor that into your current investment thesis? How much can you price that sort of geopolitics in? Yeah, I mean, I never thought that this trade talk with China was going to be simple or smooth, right?

12:21And this is yet another twist in the plot. But I would hold that August 1st deadline date really loosely in your hands because, you know, it was April 1st and then July 9th and then now it's August 1st. I think that the Trump administration, as well as China, I mean, they have a lot at stake here. So I don't think the deadline really matters. I think we're going to be negotiating until we actually get to a good place for both countries. And August 1st might come and go. Anna Ruffman, founder and CEO of Grenadilla Advisory. Thank you very much. So coming up, Oracle and OpenAI are set to expand their partnership, opening up more data centers to satiate AI demand.

13:07We have the details next. This is Bloomberg Tech.

13:16The people who seem to get more done than everyone else. They're not working longer hours or running on more caffeine. they've just stopped wasting time on the stuff that doesn't move work forward. Switching apps, re-explaining context, hunting for files. Those aren't small inefficiencies. They're hours wasted every week. Superhuman Go gives you those hours back. From the makers of Grammarly, Go is an AI assistant that sits inside every tab and tool you already use. Always available and already aware of what you're working on. Ask it to draft something, summarize a long thread, pull up a file, or prep you for a meeting.

13:58Go handles it without you ever leaving the page you're on. This is what it looks like when AI actually fits into your work, instead of adding to it. It's like having a teammate whose only job is to help you be better at yours. Go keeps up so you can move forward. With Go working with you, you can show off what you do best. Superhuman Go. Find out more at superhuman.com. That's superhuman.com.

14:47It actually pays off deep in the work that moves the business. Let's create smarter business. IBM. This is Matt Rogers from Las Culturistas with Matt Rogers and Bowen Yang. This is Bowen Yang from Las Culturistas with Matt Rogers and Bowen Yang. You know when people try on new food and suddenly it's like, wait. That's the reaction a lot of people are having when they first try Kewpie mayo. It's the one with the red cap and the little baby on the bottle. You've probably seen it in the grocery store before. And if you've ever just walked past it, some people would say that's a huge mistake. Because this mayo is different.

15:20Most mayonnaise uses whole eggs. Kewpie only uses egg yolks, which gives it this rich umami flavor. It's smoother, deeper, and almost buttery. Once people try it, they start putting it on everything. Egg sandwiches, fries, burgers. Some fans even swear by dipping pizza crust in it. And once you notice it, you start seeing it everywhere. Chefs use it. Restaurants use it. People who really care about flavor use it. Never tried it? Grab the bottle with the red cap next time you're at the store. Put it on just about anything. Then you'll understand. QP, the original Japanese mayonnaise.

15:58Oracle, along with OpenAI, have announced that they're set to develop four and a half gigawatts of additional U.S. data center capacity. Let's bring in Bloomberg's Brody Ford now on how the evolution of Stargate is going And who's involved and who isn't? Because we'll get to SoftBank in a moment, the fact that they don't seem to be on the page. But this is this commitment to what have 10 gigawatts by four years time. They really seem to be building it out now. Right. And we can't forget a gigawatt. I mean, we're throwing these numbers around now. But, you know, a couple of years ago, you would never hear about a gigawatt data center.

16:30And now we're talking about four and a half. I mean, each one of those about equivalent to a nuclear reactor. And so a four and a half gigawatt data center deal that's across the country. This is an unprecedented cloud deal. I mean, this is likely the largest single cloud deal ever signed. And so, right, if you're Oracle, if you've been trying to tell people for many years that you're going to be a real cloud infrastructure player, this is a pretty resounding sign of confidence from the most important AI company at the moment. Brody, the way we look at this is it's confirmation of our prior reporting on the projects, but also, as Kara alluded to, some clarity on the bits that SoftBank is and is not financing.

17:10Absolutely. So it's very interesting because in January we saw Larry Ellison, Sam Altman, and SoftBank sitting at the White House. Today SoftBank doesn't appear to be too involved. They're not saying we're out for good, but this 4.5 gigawatts, which likely gets us to around$30 billion a year, SoftBank is not involved in financing. Where does the rest of that money come from? That's going to be a big question in the following months, right? Because OpenAI makes a lot of money. They're not making that much money to be able to just spend all this themselves. We're going to dig into more news from Oracle in a minute, but I just want to understand where we think these data centers are going to be.

17:49Do we have much detail on the timescale here as well? Every time I ask a source where the next data center is going to be, they tell me a different state. They're evaluating a lot of different states, a lot of different sites. If you've read an article that says, hey, they want to develop a multi-gigawatt data center here, Odds are that Stargate has evaluated this site. So I'm keeping my eyes out. Slightly smaller scale, but Oracle's also in talks around$100 million a year. Paramount Skydance deal. What do we need to know about that one? What you need to know is if your dad is Larry Ellison and he helps you buy something, you're going to have to kind of help buy some stuff from him as well, right?

18:26I mean, Larry Ellison's son, David Ellison, has been in the works to buy Paramount for a while. and they're already having negotiations that assuming that deal goes through, Paramount will be a major OCI customer. They're going to put all of their CBS, MTV. It's a lot of assets, a lot of gigabytes, and they're going to be on Oracle servers. Bloomberg, that's Brady Ford. Thank you. Really on top of this story. Thank you very much. Okay, we have another story. Some updates in the$4 billion fraud case against the estate of deceased autonomy founder, Mike Lynch. A London judge ruled that HPE lost close to$1 billion after buying autonomy.

19:04Lynch's estate will face a claim for much of that sum after he lost a London fraud case along with the former CFO of autonomy. The ruling suggests HPE will likely regain just a fraction of its total claim. The ruling comes less than a year after Lynch drowned when his yacht sank in a storm off the coast of Sicily.

19:30poseidon has announced it's raised 15 million dollars in a seed round led by a16z crypto as the company looks to tackle bottlenecks in ai training data including ip safe resources this particularly aimed at training data for robotics multimodal models and physical ai here with more is s y lee poseidon president and ceo story the incubator behind poseidon state of play is really clear, right? AI models have become commoditized. Compute costs have come down. But the bottleneck as you see it and many others see it is the data. You have a solution. Yes. As you just wanted to put that into context again, because now, you know, the previous generation of models were scraping the internet's data, which is, you know, you can scrape it all.

20:18But now AI is going to the if it's the physical world, it's the multi-model models, right? It's the robotics model, it's the video model, audio model. And the really real bottleneck is the data, as you said, because, you know, you can't nest rape it illegally on the Internet. And a lot of times this data is very private. You know, there's a lot of IP safety concerns. And in that world, how do you coordinate the collection, labeling, curating of the data? And we're using actually crypto incentives to actually coordinate that massive labor coordination. So my next question was, why is Andreessen's crypto team backing you and not an AI specialist partner or a different firm?

21:04First of all, my co-founder is a robotics at UT Austin. And he's a big AI specialist who's been kind of working on this data licensing and data collection for the last 10 years. And even he thought about how to incentivize data collection with data dividends. So, you know, he is an AI expert in this realm. But the reason why we think this is so timely is, you know, 40 days ago, actually Meta semi-acquired, actually, as you know, Scale AI. And that left a very huge vacuum in the space to actually for competitors and new players to come in. And if you look at scale AI, I mean, I'm not trying to criticize, but they have to make a lot of shell companies work with shell companies in developing countries to actually coordinate this kind of digital swash shop to make that happen.

21:53But the beauty of crypto is that it's still really good at coordinating incentives at a global scale and pay them instantaneously with stable coins or other type of crypto tokens. And that is a very, very huge advantage, number one. And number two, not only that, what crypto and blockchain does is it is good at putting this data and IP on an immutable ledger and track and attribute it and license it with small contracts. And those two are very, very important advantages of crypto, even where my co-founder, AI specialist, wants to actually use this technology. That's right. Let's just go to the basics of the data that is needed here.

22:32I can see the crypto read-across. I think of Axie Infinity, the game that went bananas in the Philippines years ago, how people wanted to play because they earned money and tokens via playing this crypto-native game. You're going to use people, incentivize people to sort of video themselves in the real world that then ultimately helps Tesla build Optimus in the real world? Yes, actually, you hit on the right comparable. X-Infinity was an example of coordinating this massive number of people in the Philippines and other developing countries to actually play games to earn tokens. In the same way, now, when Tesla wants to train his humanoid robot, even if you script the entire Internet's data, you can't actually have a Tesla robot holding a coffee cup.

23:21Right. Because, you know, there's a you need a million angles of data, egocentric video footage and a million different types of settings. And that kind of data, instead of actually hiring a shell company and having the share company hire, you know, tens of thousands of people, you can coordinate this at a mass scale and instantaneously with a crypto incentive. That's what we're doing. And let's talk about the end demand here. How many companies? Who is your ultimate provider? How do you then sell this IP safe AI training data? Yeah, the reason why it's IP safe is that everyone who are actually contributing data, anyone can contribute this data anywhere around the world, right?

24:05And then actually set their licensing usage terms on the blockchain. And we are at the moment, we're not completely decentralized. And we are now going to launch a decentralized Deepin app, a Decentral Physical Infrastructure app in fall that actually collects data, you know, puts this on the blockchain with the licensing terms in place. But at the moment, we're now actually acting as an agent to actually secure these contracts. And we already have, you know, we've already signed a contract and we cannot disclose it at the moment. Right. You know, and then, I mean, in the robotics realm. Well, quickly, if ScaleAI, the scenario, left a vacuum and you look to occupy that vacuum, how many phone calls have you had from the other Frontier model makers that you could fit into their organization in the same way?

24:53Just 30 seconds. You know, actually, many, many organizations, and I was really surprised at C-Stage Company, they were willing to work with us because a lot of them were telling us that they were dropping contracts from ScaleAI. But I meant in terms of you being acquired or you joining a bigger group? We don't have any plan. We don't have any plan. And we think the advantage of this open system is that we can work with any AI farmers. S.Y. Lee, Poseidon, and Story, great to have you on today. Now, let's just turn back a minute for what's going viral, because Tesla's futuristic diner in L.A. is complete with a drive-in charging experience.

25:30Now, the sleek retro-vibe diner is also replete with two 66-foot mega screens and a rooftop skypad And a prototype of Optimist, the robot is serving popcorn. Apparently that was at the pre-launch event for the Hollywood venue. Tezza claims it's the largest urban supercharger in the world. The people who seem to get more done than everyone else. They're not working longer hours or running on more caffeine. They've just stopped wasting time on the stuff that doesn't move work forward. Switching apps, re-explaining context, hunting for files. Those aren't small inefficiencies. their hours wasted every week superhuman go gives you those hours back from the makers of grammarly go is an ai assistant that sits inside every tab and tool you already use always available and already aware of what you're working on ask it to draft something summarize a long thread pull up a file or prep you for a meeting go handles it without you ever leaving the page you're on This is what it looks like when AI actually fits into your work, instead of adding to it.

26:36It's like having a teammate whose only job is to help you be better at yours. Go keeps up so you can move forward. With Go working with you, you can show off what you do best. Superhuman Go. Find out more at superhuman.com. That's superhuman.com. The thing about AI for business, it may not automatically fit the way your business works. At IBM, we've seen this firsthand. But by embedding AI across HR, IT, and procurement processes, we've reduced costs by millions, slash repetitive tasks, and freed thousands of hours for strategic work. Now we're helping companies get smarter by putting AI where it actually pays off, deep in the work that moves the business.