Anthropic Unveils Updated Opus 4.7 Model

16 Apr 2026 · 43 min · 21 chapters

Ask about this episode

Ask anything about it. ChatGPT or Claude reads this page and answers with the times it was said.

Connect VO and ask about every podcast you hear, including the moments you saved. Add to ChatGPT · Add to Claude

In short

Bloomberg Tech covers rapid AI model updates (Anthropic Opus 4.7 and Mythos), cybersecurity guardrails, AI chip demand and semiconductor supply chains, and broader market/geopolitical context. It also discusses political backlash to AI data centers (U.S. and Spain) and fintech/AI adoption.

Guests (and backgrounds)

  • Rachel Metz, Bloomberg reporter covering AI (reported on Anthropic’s model internals).
  • Peter Elstrom, Bloomberg global tech editor (Asia-focused tech reporting).
  • Scott Ladner, Chief Investment Officer at Horizon (invests in tech/AI hardware).
  • Margie Murphy, Bloomberg cybersecurity reporter (Mythos cyber risk reporting).
  • Eric Abrescia, Managing Director at Redpoint Ventures; Linux Foundation board member; cyber/AI governance investor.
  • Mark Niquette, Bloomberg reporter on politics/tech infrastructure impacts.
  • Clara Hernan-Slizaraga, Bloomberg correspondent in Spain (data center/environment reporting).
  • Pip White, Anthropic head of EMEA North (Mythos rollout in UK financial institutions).
  • Victor Cadenas, CEO/co-founder of Slash (B2B fintech; AI-enabled operations).

Key claims & notable examples

Opus 4.7 improves coding (larger, more complex tasks) and computer vision (higher-resolution images; better extraction). Mythos is limited-access due to cybersecurity risk; Anthropic experimented with reducing Opus 4.7 cyber capabilities. Mythos reportedly found vulnerabilities across operating systems and browsers; internal review showed it could exploit flaws autonomously. TSMC profit rose 58% and raised 2026 revenue outlook, driven by AI chip demand. Elon Musk’s TerraFab: sources say suppliers (Tokyo Electron, Applied Materials, LAM) were contacted for quotes; industry skepticism remains (e.g., Samsung reluctance). Data center backlash: U.S. states considering moratoriums; Maine reportedly first to impose one; Spain’s Aragon faces land/pollution concerns despite job promises.

Written by AI. May contain mistakes. Listen to the episode to check what was said.

Chapters

Tap a time to open that second in VO

Anthropic's Opus 4.7 Launch

2:26 to 3:22

Discussion on Anthropic's updated AI model Opus 4.7 and its implications.

“Anthropic introduces an updated version of its widely available AI model Opus 4.7 just a week after its limited release of Mythos.”

Market Reactions to AI Developments

3:24 to 7:40

Exploring market responses to Anthropic's Opus and related AI news.

“But we've got to be talking as well about Anthropic introducing an updated version of its widely available AI model, Opus, of course, barely a week after its limited release of a more advanced offering called Mythos.”

Elon Musk's TerraFab Project

7:42 to 11:03

Insights into Elon Musk's ambitious chip-making initiative TerraFab.

“Elon Musk is racing ahead with his chip-making TerraFab project.”

TSMC's Strong Market Performance

11:03 to 12:14

Analysis of TSMC's profit surge and its implications for the tech industry.

“And when we broke the story in Asia, those names that you mentioned jumped pretty significantly.”

Global Investors and Market Dynamics

12:14 to 14:03

Discussion on global investor behavior amidst current geopolitical tensions.

“Nobody knows exactly what that's going to do to the global economy, to trade.”

The Impact of AI on Productivity and Market Dynamics

14:03 to 15:10

Explore how AI is expected to enhance productivity and affect market pricing.

“We're talking about, like, the electrification of the country, like, you know, the widespread availability of electricity.”

Earnings Season and Its Influence on Tech Stocks

15:10 to 16:38

Discuss the implications of the upcoming earnings season for tech companies.

“And I do think as folks start to figure out the margin channel is really the thing that matters that's going to make AI even more important to the business processes.”

AI's Role in Hiring and Training in Finance

16:38 to 17:44

Understand how AI is transforming hiring practices and training in the finance sector.

“You probably want to buy individual stocks.”

Anthropic's Mythos: Cybersecurity Innovations and Challenges

20:43 to 22:45

Examine the implications of Mythos for cybersecurity in the finance industry.

“UK institutions will soon get access to Anthropik's powerful new model, Mythos, on preview.”

Concerns Surrounding AI Model Releases

22:45 to 25:06

Delve into the decision-making process behind releasing AI models and their risks.

“And obviously, I've already articulated our kind of approach with Project Glasswing.”
Show all 21 chapters

The Future of AI in Cybersecurity

25:06 to 26:31

Discuss the potential and challenges of AI in enhancing cybersecurity measures.

“And the executive team kind of had to sit around and say, how do we deal with this?”

Current Trends in Tech and Market Performance

26:31 to 28:00

Analyze the latest trends in the tech market and their implications for the future.

“Check it out on the terminal or on the website.”

TSMC and AI Infrastructure Challenges

28:00 to 28:58

Discussion on TSMC's performance and the challenges in AI infrastructure.

“and it's top-line growth and bottom-line growth.”

Political Backlash Against Data Centers

28:58 to 29:55

Mark Niquette discusses the political implications of AI infrastructure and data centers.

“we've talked about this in the context of energy prices and zip codes where data centers are getting billed outright.”

Community Concerns Over Data Centers

29:55 to 31:06

Exploration of local opposition to data centers and their impact on communities.

“And what we sought out to explore was the reaction that we've seen at the federal and state level to this backlash by having the politicians ask the developers to sort of bring or buy their own power.”

Spain's AI Infrastructure Ambitions

31:06 to 32:06

Clara Hanans-Lizaraga reports on Spain's push for AI data centers and local sentiments.

“the concerns that residents have about data centers?”

Balancing Economic Growth with Regulation

32:06 to 34:31

Discussion on the balance between economic growth and regulatory measures in Spain.

“Over in Europe, there are tensions flaring about the environmental footprint of AI in Spain in particular.”

Market Reactions to Political Developments

34:31 to 36:22

Analysis of market reactions to political news regarding the war in Iran.

“to start building these data centers for Microsoft or AWS with less regulation because it's needed for the economy.”

AI Security and Investment Perspectives

36:22 to 39:48

Eric Abrescia shares insights on AI security and investment strategies.

“OpenAI has rolled out its new GPT 5.4 cyber model to select users, doubling down on security as competition with Anthropic heats up.”

Interview with Victor Cadenas of Slash Financial

44:58 to 49:14

Discussion with Victor Cadenas about Slash Financial's growth and funding.

“Slash Financial has just raised$100 million from backers, including Ribbit Capital, Kozla Ventures.”

IPO Watch and Industry Insights

49:15 to 50:39

Insights on current IPOs and market trends in the industrial sector.

“It's not the reason I'm in New York, but we're here for it.”
Hear the part that matters, and keep it.Open this episode in VO. Double tap your headphones to save a moment as you listen.
Get VO free

Transcript

Automatic transcript. May contain errors.

0:00The right technology can strengthen human judgment. That's why Deloitte brings together AI and data analytics with multidisciplinary teams who can help you connect the dots across your enterprise. From risk to operations to customer needs. So opportunities don't slip by and surprises don't spread. Because the smarter your systems, the sharper your instincts. That's how technology makes people better at what they do best. Deloitte. Together makes progress. Learn more at Deloitte.com slash Together Makes Progress.

0:58Caroline Hyde:deep in the work that moves the business. Let's create smarter business. IBM. When you're running a business, the best days are the ones where priorities stay on track. For midsize and large companies, that isn't always easy. Risk can touch multiple parts of an organization at the same time, often in ways that aren't immediately obvious. It might involve property, liability, or cyber. It could stem from regulatory requirements or challenges tied to a specific industry or the scale of an operation. At that level, managing risk becomes an ongoing discipline, not a one-time decision. At the Hartford, the focus is on helping businesses manage risk before it turns into something more disruptive.

1:37Caroline Hyde:That means working with companies to identify where they're exposed, decide what matters most, and put practical standards in place so risk is managed as part of day-to-day operations. And when losses do happen, the Hartford can pair that risk control work with insurance coverage grounded in underwriting, risk engineering, and claims experience developed over time. Learn more at thehartford.com slash risk mitigation. Bloomberg Audio Studios. Podcasts. Radio. News.

2:13Bloomberg Tech is live from coast to coast with Caroline Hyde in New York and Ed Ludlow in San Francisco.

2:23Caroline Hyde:This is Bloomberg Tech coming up. Anthropic introduces an updated version of its widely available AI model Opus 4.7 just a week after its limited release of Mythos. Plus, Elon Musk is kicking his terrified plan off into high gear, even as skepticism grows from the semiconductor industry. And TSMC reports a big surge in profit and raises its revenue outlook for 2026, driven by strong demand for AI chips. There is so much AI news out today, but there's still the geopolitical global context that investors are currently trying to navigate. And it's a tough one. Stocks, as we speak, at new record highs.

3:01There we are. Come closer. We're up almost two tenths of a percent. NASDAQ 100, therefore, up for a 12th straight day, the longest winning streak since 2017. But longer term peace processes needed in the Middle East all remains elevated. Concerns about the Hormuz straight, of course. But at the moment, tech has been leading the charge and it continues to as we see it on the day higher. And then it's looking at it on the 12 days higher. We're now seeing that extraordinary run, 14 percent higher ed more broadly. But we've got to be talking as well about Anthropic introducing an updated version of its widely available AI model, Opus, of course, barely a week after its limited release of a more advanced offering called Mythos.

3:42Now, Bloomberg's Rachel Metz joins us for a little bit more. And we want to get into a broader market context in a minute. But Rachel, for now, there hasn't been much market reaction to Opus. But what are you seeing at this current moment of really how it's improved 4.7 versus 4.6? Sure. I mean, it sounds like it's improved at a broad range of capabilities, at least that's what Anthropik has said so far, particularly at coding and taking complicated tasks that people might in the past cut up into a few pieces. You might be able to give it an entire more complicated task, especially as pertains to software engineering.

4:20It also sounds like it is better at computer vision, which is the task where a computer is trying to figure out what is in a picture. It can take on much higher resolution images, they say, and it can also better extract information from these images, which could lead to a whole range of different kinds of applications, like analyzing different kinds of data, more complicated data, that sort of thing.

4:45Caroline Hyde:I think the point that Anthropik would make is that out in the real world, everything is not plain text so that vision component was really important um generally speaking they are at pains to make distinctions between opus 4.7 being an updated and improved existing model from 4.6 and how it's different from mythos right i think they even talk about in the blog about how they experimented with ways to quote differentially reduced 4.7's cyber capabilities. Could you just explain a little bit about the relationship between 4.7 and Mythos that we've all been talking about for like a week now? Sure.

5:27So Mythos is a model that the company has said is so powerful that it doesn't want to generally release it and it is handing it over to a number of companies who it wants to be like using it in the earlier stages. is this is a model that's meant especially for spotting cybersecurity vulnerabilities so that people can then patch them. But of course, it could also be used for the opposite, if it is as good as they say it is at finding those vulnerabilities. So this model is not going to do all of those things. It is purposely not doing all of those things. For that reason, some people might not be happy with it, but the company says that it is also better at a wide range of things than its previous Opus model.

6:09So we'll see how it is used. For a lot of people, it will probably be considered an improvement, a step up from the past one. Maybe not the step change that everyone feels that mythos is, Rachel. But just give us the context. This is a company that we understand, according to Bloomberg Reporting, is pushing back on an extraordinary valuation, yet more funding because the venture community wants to get in on it so much. And we are seeing this global rollout a little bit more. Anthropik's just so hot right now. Is 4.7 going to be enough to maintain that momentum? Oh, that's a good question. I guess we're going to see how people are using it.

6:44I mean, I think the company is really on a chair right now due to like a bunch of different factors. And some of that is its technology and the capabilities of that technology. Some of it, I think, is vibes, right? But we'll see. I mean, this is like, as we know, in AI, as in any other market, there isn't usually like one winner. So I think that it's still anybody's game and there will probably be a bunch of winners in the end.

7:13Caroline Hyde:We're showing this chart, which is on the Bloomberg terminal. It's data that I think we keep our own. It's actually coming from Ramp, shows OpenAI ahead in terms of adoption. It's a sort of data set. But we remember that in the context of Opus 4.7 caro and with Mythos, the story of Anthropic from day one was enterprise. That is a more recent pivot for OpenAI. Bloomberg's Rachel Metz, you have done terrific reporting on the inner workings of Anthropic and what's going on with these models. Really appreciate it. Another big story from Bloomberg. Elon Musk is racing ahead with his chip-making TerraFab project.

7:45Caroline Hyde:Sources say that billionaires, the tenants, have already reached out to industry suppliers, including Tokyo Electron, Applied Materials, and LAM. And they're seeking price quotes, delivery times for an array of chip making gear. Let's get out to Bloomberg's global tech editor, Peter Elstrom. This was a big team effort around the world to try and work out how real is this with TerraFab. And we got a lot of detail, Peter, on where they're at and the pressure that Elon, Inc. is putting on the supply chain to get moving. Give it what it wants. What do we need to know? Yeah, just to set the stage here, when Elon Musk first laid out this vision of a terra fab a few weeks ago, nobody really knew what to make of it.

8:27Caroline Hyde:He was talking about this getting into chip manufacturing for the very first time, something he's never done before. It's a very competitive market, a very difficult market. He was talking about building the equivalent of hundreds of new fabs that he was going to roll out and he was going to invest. one analyst estimated that it would cost something like$5 trillion to$13 trillion to be able to do what he was actually talking about. So nobody knew how seriously to take him. There were a couple of analysts who poo-pooed it from the beginning and said he's not really going to go ahead and manufacture chips or even try to manufacture chips.

8:58Caroline Hyde:He's just trying to get the existing companies to move ahead more quickly. But, and I want to turn the tables around a little bit here, Ed, because you were part of the people on the byline in this story. It sounds like now we need to take them a little bit more seriously. What do you make of this? How seriously do you think we should take this effort now? Well, what I would say is the two data points that we had in the story. 3 ,000 wafers per month initially is nothing in the context of current capacity for lead edge fab. And then what we understand from sources, by the way, 2029 is when they start.

9:29Caroline Hyde:I don't know how far ahead you're looking on your calendar, but I don't even know what's happening next week. 2029 is quite far away. It is. And look, we had Berenberg analysts on yesterday, who I love that you quoted and put the video in the story of, really saying they're not factoring in these numbers yet. It's going to be two, three years really for them to understand when ASML gets involved in giving the equipment that's going to be necessary to make the chips more broadly, Peter. But just talk about the reticence that the rest of the industry has towards this. I think one line that really caught my attention and caught Ed's, I know, was that Samsung's like, no thanks.

10:03We don't particularly want to help in this in the way that perhaps you'd like us to.

10:07Caroline Hyde:Yeah, it's important to say that the leader in this market right now, the foundry business is TSMC. They're far ahead of everybody. We just got their earnings today. We can talk about those a bit, too. But Intel and Samsung are both trying to get into this foundry business because it has been so good for TSMC. They know that that business is growing. So if you have somebody like Samsung that's already expert in building these kinds of chips, they're going to be kind of reluctant to trade some of that expertise to somebody like Elon Musk, who has totally no track record in this area whatsoever.

10:36Caroline Hyde:The same applies to the chip equipment makers. So Tokyo Electron, Applied Materials, ASML, they need to decide who are we going to sell our equipment to. We're in this market right now where there's actually a limit to the amount that they can provide, and they're limited more by supply than actually demand at this point. So are they going to sell it to Elon Musk, who's never done this before for the TerraFab project? Or are they going to sell it to TSMC and Samsung and Intel, who are pretty experienced at making chips at this point? And when we broke the story in Asia, those names that you mentioned jumped pretty significantly.

11:07Caroline Hyde:And pre-market, some of the U.S. suppliers also did, then fell away. On Samsung, they basically counted, is my understanding, saying, we're going to ramp up in Texas. But the story in the markets morning is TSMC, right? On the call, TSM CEO was like, this is the reality of building a fab. Three years to build it, two years to ramp it. But things for TSMC in particular are going well, Peter. The data tells us what you said. They are the leader and demand is accelerating. Yeah, it was interesting to see. C.C. Way, the CEO, was asked about TerraFab and he said there are no shortcuts. You can't do this faster than what the leaders in the industry have been able to do so far.

11:46Caroline Hyde:But what TSMC's financial results show us is demand is very, very strong. Their profit was up 58%. That was well above analyst estimates. They said they boosted their revenue growth forecast for the full year to more than 30 % growth before they had had it below 30%. They also said that they're going to ramp up their capital spending. They're going to be at the top end of this range where they approach$56 billion. So they're moving very, very quickly. They say the demand for AI is very, very strong at this point. There were a couple of notes of caution in there. Of course, the Iran war. Nobody knows exactly what that's going to do to the global economy, to trade.

12:21Caroline Hyde:There's another issue that's kind of critical for the tech industry right now that we've talked about a couple of times, the crunch in memory chips. Memory chips, of course, go into all sorts of things. Because the AI hyperscalers are building out so quickly right now, they're taking up a bunch of that capacity. That's limited the ability of smartphone makers to be able to get supply. That could taper off some of the demand in the future. And then there's Peter Elstrom, who's been a busy man editing the great story from Ed and the rest of the team and keeping an eye on all things Asia. Let's keep an eye on all things broader markets.

12:51Global investors, they're continuing to navigate a precarious environment. We keep talking about it. But U.S. stocks have recovered to new record highs. And it's been led by the technology sector. In the studio with us now, Scott Ladner, he is Chief Investment Officer at Horizon. So do you buy into chips even more in AMD than Intel are leading on a points perspective on the day? Do you continue on the hardware AI theme?

13:10Caroline Hyde:Absolutely. I mean, I think it's really the thing that's driving the market as we came into this conflict, and it's going to drive the market as we come out of this conflict. I think the reason why we are seeing markets at all-time highs is because, you know, the market's decided that this thing's going to be short. We're measuring in weeks, not months. And the tailwinds we had coming into it are just going to continue to kind of wash over the markets, and that's why we're sitting where we are right now. Scott, it's good to be with you in person. I really enjoyed our conversations of late. So if you look at the NASDAQ 100, just as the case study, it's up for a 12 straight day in spite of the war in Iran.

13:40Caroline Hyde:So what is it? What's the data point or the thesis that's driving the technology sector higher if people are willing to look past a major conflict and a huge impact in energy markets? I mean, I think it's a couple of things, Ed. It's one, you know, the fact that this is sort of the next big thing. And when people talk about this as being like the next Internet, I think that's wrong. I mean, I think we're talking about this in terms of scale and importance. We're talking about, like, the electrification of the country, like, you know, the widespread availability of electricity. It's that scale of productivity enhancement that I think we're talking about.

14:10Caroline Hyde:with AI. So I just think people are actually still under sort of appreciating how important this thing's going to be. And then as the market kind of looks forward and kind of sort of figures out how to price AI, like, you know, how do you price the AI in this other company? I think that's the thing that we're going to have to figure out next. It's going to be through the margin channel. You know, like, you know, ultimately, like, productivity, you know, it should increase margins. And that's, and that's, that is going to be where the rubber is. There's no point in exaggerating or hyperbole, but the NASDAQ 100 is up for 12 days.

14:34Caroline Hyde:It hasn't done that since 2017. So something clearly is happening underneath. Now we have earnings season. How much could that derail that? Well, I mean, look, it could derail it or it could sort of verify why we are up for 12 days straight. We're not up 12 days straight in a vacuum though. We're up 12 days straight coming off of some pretty nasty declines as a result of the war. So the context for that stuff matters. But we actually think that the bar is beautiful for these tech companies as we come into this earnings season. I mean, I think people are trying to understand what is the next leg?

15:08Caroline Hyde:What is the next stage? And I do think as folks start to figure out the margin channel is really the thing that matters that's going to make AI even more important to the business processes. And I think it's going to just drive attention into the sector even higher. I hate to say it, but margin often equates to fewer people. What's interesting is we're looking at Opus 4.7 just getting announced. In it, it says, as well as state-of-the-art score and finance agent evaluation, our internal testing showed Opus 4.7 to be more effective finance analyst than Opus 4.6. Now, how's that make you feel about your role, about the people that you're hiring, and more broadly where you want to invest and how it's going to help finance?

15:49Caroline Hyde:I mean, frankly, it makes me kind of excited. Yeah. Because the thing that we've seen out of AI, and so we have some young people at our company, and the thing that has been the most impressive to me is how quickly they are adopting and sort of figuring out and getting up the learning curve, frankly. I mean, it's really tough for a 23-year-old to be very useful in this industry because you haven't seen very much. And they're using these tools to become much smarter, much faster, be able to try more things and do a lot more experimentation. And that's how you actually build knowledge and get to be a real good finance pro.

16:17Caroline Hyde:And so I'm actually kind of buffeted by the use case of this stuff. And it's not going to change the way we hire. It's going to change the type of people that we hire and how we train them. But in terms of it, you still need bodies to figure some of this stuff out. Scott, briefly, should we buy software at this point? Whew. Yeah, probably, but you have to be differentiated, I think. I don't think you want to buy the ETF. You probably want to buy individual stocks. I was scrambling. I've got the benchmarks here on Opus 4.7, and I was waiting to see if Scott said he was worried about his job. Scott Landon, Chief Investment Officer at Horizon, always great value on Bloomberg Tech.

16:51Caroline Hyde:Thank you very much. A federal jury has ruled that Live Nation illegally monopolized the live events industry and overcharged fans for tickets to music performances. The landmark antitrust trial follows years of regulatory scrutiny of Live Nation's dominance in the live events business. And it now sets the stage for a possible breakup of the largest U.S. concert promoter and ticket seller. Live Nation says it will appeal the verdict. Carrie. Now coming up, Ed, UK financial institutions will soon be getting access to anthropic mythos as concerns about cyber risks from the global regulators and enterprise continues to grow.

17:25That next, this is Bloomberg Tech.

17:34The right technology can strengthen human judgment. That's why Deloitte brings together AI and data analytics with multidisciplinary teams. People with deep industry experience who can challenge assumptions and help you connect the dots across your enterprise. From risk signals to operational pressure points to shifting customer needs, Deloitte helps you see what's coming sooner. so opportunities don't slip by and surprises don't spread. It's not just dashboards. It's real clarity in the moments your decisions are made. When models reveal patterns, people can ask better questions. When data and people are connected, leaders can move faster with confidence.

18:15And when your teams are aligned, smart choices can scale from the front line to the C-suite. Because the smarter your systems, the sharper your instincts. That's how technology makes people better at what they do best. Deloitte. Together makes progress. Learn more at Deloitte.com slash together makes progress.

18:34Caroline Hyde:Support for the show comes from Public. Public is an investing platform that offers access to stocks, options, bonds, and crypto. And they've also integrated AI with tools that can assist investors in building customized portfolios. One of these tools is called Generated Assets. It allows you to turn your ideas into investable indexes. So let's say you're interested in something specific like biotech companies with high R &D spend, small cap stocks with improving operating margins, or the S &P 500 minus high debt companies. Chances are there isn't an ETF that fits your exact criteria. But on public, you just type in a prompt and their AI screens thousands of stocks and builds a one of a kind index.

19:15Caroline Hyde:You can even backtest it against the S &P 500. Then you can invest in a few clicks. Go to public.com slash market and earn an uncapped 1 % bonus when you transfer your portfolio. That's public.com slash market. And paid for by Public Holdings. Brokered services by Public Investing, member FINRA SIPC. Advisory services by Public Advisors, SEC Registered Advisor. Crypto services by ZeroHash. Sample prompts are for illustrative purposes only, not investment advice. All investing involves risk of loss. See complete disclosures at public.com slash disclosures. Deadlines move. Plans change. And sometimes opportunities pop up out of nowhere.

19:52When you need branded gear fast, 4imprint is ready to deliver. 4imprint offers hundreds of promotional products in their 24-hour category. Everything from custom apparel, bags, and drinkware to writing tools, trade show staples, and high-tech gear. At 4imprint, they're focused on getting the details right, printing your logo with precision, packing your order with care, and shipping it out fast. And it's backed by their 360-degree guarantee. That's 4imprint's promise your order will show up right on time, just the way you planned it. That's what it means to be 4imprint certain. So, if you're prepping for a last-minute event or jumping on a big opportunity, you don't have to settle or scramble.

20:32With 4imprint, fast, reliable service and peace of mind are built right in. Check out their full 24-hour selection at 4imprint.com. 4imprint. 4certain.

20:49UK institutions will soon get access to Anthropik's powerful new model, Mythos, on preview. Now, the company's head of EMEA North, Pip White, spoke with Bloomberg's Tom McKenzie earlier. Take a listen.

Read the full transcript

21:00Caroline Hyde:Mythos found vulnerabilities in every single operating system and in every single web browser. And so we wanted to build this small cohort first to basically learn from what they found in a very contained and controlled manner. But you're right. Our intention is definitely to expand this out in a very intentional way. Typically, it hasn't been expanded to UK financial institutions yet. No, but I think that is in the very near term, like in the next week. Weeks away? In the next week. In the next week. OK, and these are top UK financial institutions. Yeah, we've been, I mean, as you would expect, the engagement that I've had from CEOs in the last week in the UK has been significant.

21:36Caroline Hyde:Customers firstly want to say that they appreciate our very thoughtful approach to the way in which we have released Mythos. But secondly, of course, they want to understand the opportunity to be considered to gain access to that model in a controlled way. And so we've been having a lot of conversations with CEOs about the reason why, firstly, we did Project Glasswing and the way that we thought about that intentionally. And then secondly, being very clear and to understand exactly why they want to get access to Mythos and then working in partnership with those customers to think about the next kind of phase two and three and four.

22:14Caroline Hyde:Does the system of release from Frontier Labs need to change? Some will be looking at this and thinking Anthropic has made this decision to limit the release of a model that can prove very dangerous. There's the upside, but there's also clearly the dangers that you've articulated and that the team have articulated. It's come down to decision-making at Anthropic as to whether or not and how to release this. That has to change, doesn't it? I think where I would position it is that this model in particular, we saw a significant leap in cyber. And as a result of that, we were very concerned about mass release and made the decision not to do that.

22:50Caroline Hyde:And obviously, I've already articulated our kind of approach with Project Glasswing. I think labs have to take a responsibility for thinking about safeguards, for thinking about safe and responsible AI as it gets into the hands of the enterprise. That was Anthropix head of EMEA North, Pip White. Let's dive a little deeper into concerns Anthropix Mythos has raised, even within the company. Bloomberg cybersecurity reporter Margie Murphy has been looking at exactly that in today's Big Take. This is an important body of work, timely, and packed with detail about first the origins of Mythos, why they started with the cybersecurity capability.

23:32Caroline Hyde:It's not the only capability. Take us inside that report and what we need to know. yeah so we were really keen to understand how anthropic came to this decision this really unusual unprecedented decision to not only hold back um you know what would be the the latest update to claude and be you know a great revenue spinner but also open it up to its competitors to use. And so to do that, we spoke with Nicholas Carlini, who in the cybersecurity world is known as being a kind of security researcher genius, and learned that he'd actually realized back in during testing a previous model that Anthropics Opus could actually figure out ways to potentially exploit exploit flaws in in tons of software so they they were kind of aware that there might be some issues with cyber security with the newer MIFOS model so they were keeping their eye on that as they were training the model and and testing and then around February when they released it for internal review which means that all the anthropic employees are allowed to sort of play around, find glitches, feedback.

24:49Nicholas realized that there was not only could he get help from Mythos to exploit these flaws, Mythos could just go and do it completely on its own. So that was a step up. And that's when they raised the flags to the executive team. And the executive team kind of had to sit around and say, how do we deal with this? can we release this? Should we put guardrails in? And isn't it interesting that they are implementing more safeguards when it comes to Opus 4.7 that's just been released, talking about being able to detect and block prohibited or high-risk cybersecurity uses. They've learned from Mythos already on what they are allowing us to have.

25:29But Margie, push us as to whether it is as powerful quickly as many make it out to be, because there is some quandary. Right. Look, I think a good question is, it was actually Heather Adkins, who's head of security engineering at Google, and where they have, you know, through Gemini, they have a big sleep LLM that they've released that does the same thing. It finds bugs. It helps companies know if they have vulnerabilities they need to fix. But they put guardrails in place so that it can't kind of run wild and exploit things. Anthropic has come to the decision to release in a limited way something that can go and do quite scary things so people can find them.

26:10So I think, you know, what it's found is quite incredible. Other models probably can do the same. It's a decision about putting the guardrails in place. Great context, always. And you've got a Q &A on all of this tomorrow. You've got to tune in to Margie Murphy. We appreciate it. And Bloomberg's journalists are going to be, therefore, answering those questions for you on Anthropics Mythos. I think it's 1.30 p.m. New York time. Check it out on the terminal or on the website. Time now for Talking Tech. First up, Tesla's Cybertruck has been struggling to drum up consumer interest, but it's getting a boost from inside Elon Musk's empire.

26:45SpaceX made up about 18 % of U.S. registrations in the fourth quarter, of just over 7 ,000 for the truck sold. Plus, SpaceX is reportedly lining up tours of its Texas and California facilities for prospective anchor investors, according to sources. SpaceX is aiming to showcase its operations to big-ticket backers like Sovereign Wealth Funds and has already filed confidentially for an IPO. And an early bet on SpaceX could deliver a huge payoff for Alphabet. A new filing shows Google held a 6.1 % stake in the rocket maker at the end of 2025, with roughly$1.22 billion if SpaceX hits a$2 trillion valuation in its IPO end.

27:25welcome back to bloomberg tech where we're both in new york and it's such a joy to have ed here to discuss what's been tech on a tear look let's back up again on the day we're at a record high we've been up for 12 straight days that is the longest winning streak for the tech benchmark since 2017 so at four tenths percent in fact we're sort of managing to be near or at session highs just coming off of them a little bit right i guess we made the point that's not a vacuum

27:49Caroline Hyde:That's where the war in the Middle East, energy prices going crazy, full conviction on tech. I mean, extraordinary. And conviction, because look at TSMC and look at these other numbers. We're getting fundamentals from Netflix, for example, after the bell. Yeah, so Netflix is a first read since walking away from Warner Brothers' story, and it's top-line growth and bottom-line growth. And Netflix will probably want to keep it to that, right? Because everyone's really focused on engagement right now, which prior to them walking away, I don't know that the engagement data was that good. Highest this December 2025.

28:18TSMC is interesting, isn't it? Down, maybe weakening a bit, but it moved ahead of its numbers. Its numbers are stellar. The desire to buy chips just shows the bottlenecks in AI and Compute right now.

28:28Caroline Hyde:Yeah, and going into earnings prints, and TSMC seems to report some kind of earnings every week, you know, there's a high bar. And I think we have that maintained for the rest of earnings. There's a lot going on in the world of AI infrastructure. The massive build-out of the infrastructure is hitting a political wall just months before the midterm elections. Bloomberg's Mark Niquette joins us now to share more about the high stakes balancing act between industrial progress and political survival. I mean, in the last week alone, Mark, we've talked about this in the context of energy prices and zip codes where data centers are getting billed outright.

29:04Caroline Hyde:For the president, that is an issue amongst the electorate. Now, going into the midterms, there is a NIMBY consideration. There are all sorts of things about how real everyday Americans read the build-out in AI infrastructure and whether they care. What are you reporting? Well, we've been tracking sort of the political impact of the growing opposition to data centers we're seeing at the local level, primarily because of rising electricity prices. We saw a big impact in the gubernatorial races in Virginia and New Jersey last year, last November, and a sense that this was becoming more of a political liability for politicians if they didn't address the rising electricity prices, in particular because of data centers and data center demand.

29:55Caroline Hyde:And what we sought out to explore was the reaction that we've seen at the federal and state level to this backlash by having the politicians ask the developers to sort of bring or buy their own power. The idea is data center developers would build their own on-site generation or have purchase agreements that account for their energy usage. This could be a way to address the concerns that voters have about electricity demand from data centers. I mean, the news just keeps on coming. Maine just this week has become the first state that's sort to put a moratorium on the build out of AI data centers.

30:40It's still got to be signed off by the governor. But I was speaking with a business owner in Maine just saying the relief that we just get a pause so we can understand what regulation is needed. So does it need more for the electorate than just Microsoft paying its way for its power?

30:55Caroline Hyde:I think so. Our story today sort of explored this issue. Does having a project sort of bring its own power or buy its own power alleviate any of the concerns that residents have about data centers? And the answer seems to be no, in the sense that there's still a lot of opposition to data centers because folks don't want to live next to them or they're concerned about pollution, either noise or emissions from the power sources that are being generated to power the data centers. and just overall concerns about AI and folks opposed to the advancement of artificial intelligence. So what we have today sort of explore the idea that if the politicians' outreach to get data centers to bring their own power was going to alleviate opposition to data centers, it probably won't, or at least it's not going to alleviate all of the opposition.

31:58Mark McKeat, great reporting. We urge people to go and read that story. And look, that's about domestic tensions in the U.S., but it's not just happening here in the U.S. Over in Europe, there are tensions flaring about the environmental footprint of AI in Spain in particular. It's actually still positioning itself as the AI capital of Europe, a$90 billion infrastructure push there. Lumo's Clara Hanans-Lizaraga is here with us to detail where you went in Spain. You discussed with families, with farmers, who suddenly had this sense of urgency to sell their land to AWS, for example. Go from there.

32:31What have you learned from being on the ground as to how people feel about these data centers? Thank you. Hi. So I traveled to Adarron. I spoke to locals, to farmers. And the sense I got is, you know, some people are in favor of the data centers, but some people are very directly affected because these infrastructure needs land. And that means either expropriating people or pushing them out of land. They've been leasing for generations. So for a lot of the locals, there is a sense that there's very little in it for them and that most of the benefits of this push are going to go elsewhere.

33:10Caroline Hyde:We're talking about the region near Barcelona, northeast Spain, and the pitch from the Spanish government and from municipalities and from industry is this is the model of how it should be to build out data centers. What is their pitch for how they think this will benefit the region economically? And what's the real terms build out that's happening on the ground? yeah so the the local government um argues that alaron will become the virginia of europe um they expect to it will create many tens of thousands of jobs um and it's been so successful that actually the the companies are investing there such as amazon and Microsoft, they're now pushing this idea that the framework that's in place there should be an example for the rest of the EU.

34:10Nationwide, though, the central government is more reluctant. There is a different political party and the government in Madrid is more keen to regulate and set some limits to, you know, who gets to plug into the grid and who gets to build. And there's the tension, that PIGA framework, basically an ability to fast track the ability to start building these data centers for Microsoft or AWS with less regulation because it's needed for the economy. But you sort of point out in the story that actually maybe it's not that necessary for this part of the economy. There's people who've got pretty wide, broad reach of jobs.

34:49and I'm interested as to whether regulation will come in and what more broadly you'll see from companies responding to this because they say PIGA should be used more broadly and it's necessary. Yes, so regulation is expected at EU level. I believe the commission is set to present a proposal in the first quarter of this year so it should be happening soon. In terms of Spain alone There is a law that came out in March that will set the, so the granting of the permits will be tied to meeting certain sustainability and efficiency criteria. But those new measures will only apply to the future projects, not to the already existing ones.

35:36Caroline Hyde:Bloomberg's Clara Hernan-Slizaraga, who joins us as a correspondent from Spain. Thank you so much.

35:47We want to bring you some breaking news on the war in Iran. President Trump posting on Truth Social saying a 10-day ceasefire has been agreed to with Lebanon and Israel. The markets react. Optimism there. We're still higher. We were higher throughout the day, Ed. But many had wondered how this would progress given Hezbollah had, well, not wanted to down their own tolls or arms.

36:07Caroline Hyde:Yep. And, you know, that's the market reaction in equities. There's some reaction elsewhere, but in the post, President Trump saying that he spoke with both Netanyahu and the leaders of Lebanon directly. We will continue to monitor it as well as the market impact. Let's go to private markets. OpenAI has rolled out its new GPT 5.4 cyber model to select users, doubling down on security as competition with Anthropic heats up. The focus on cyber comes as trust and governance continue to kind of slow down AI adoption in different sectors at scale. So what does this next phase of the AIS race mean for startups?

36:44Caroline Hyde:But it's investors too. Joining us now is Eric Abrescia, Managing Director at Redpoint Ventures, a backer of OpenAI at the Series C stage. This is really interesting right now. You know, defense technology generally is still a main focus of investors like yourself everywhere. But the relationship between the technology sector and the arms of government, the Pentagon, we're rethinking it a bit. What's your latest thinking on that? Well, look, I think we've always known that these models were going to get increasingly powerful. And we need to work together across the private and the public sectors to make sure that we keep people safe.

37:23Because whatever, you know, power we build into models and try to use responsibly, there are others in the world, unfortunately, that will use maliciously. And so it's important that folks have open dialogue. we're constantly sharing and collaborating and figuring out the best way to shore up defenses for those who need and deserve it to make sure we can always stay one step ahead of those who are meaning to do harm. It's so fascinating to have you on in particular because of your experience of, you know, when you were being purchased, your own company, Bitnami was being purchased by VMware. You think about what the experience you had at GitHub.

38:00But for me, it's the fact that you're on the board of Linux Foundation since all the way back to 2016. Look, our Margie Murphy, our cyber report is really talking about how Mythos has been a concern for the very underpinning of the internet, Erica. How much are these models giving you pause over at that foundation or more broadly an opportunity for some of the cyber companies that you back? Look, I'm a techno optimist, right? And I think that AI can do wonders for open source, which to your point is the underpinnings to the internet. Using these models can help us find vulnerabilities that we didn't know were there.

38:35It can help them patch vulnerabilities more quickly. So I think it's great for the world. And as you mentioned, we do have a number of cyber investments, one of which is called Jetstream, that is focused on AI governance and visibility. It'll allow enterprises to understand what's happening in their organization, create a dynamic graph of AI activity across the system and put in place controls to secure it. And so, I think there's a lot of opportunity both on the open source side of things to bring AI to improve open source, as well as a lot in private companies to deliver tools to enterprises to get them ready for what's to come as AI continues to proliferate.

39:22Caroline Hyde:So you're the investor and you have a conversation with the portfolio companies and they'll probably go to you, Erica, what do you think we should do? And the tension is this. Industry says these should be smaller open source models. That is the best way for everyone to work collectively in the cybersecurity context. Anthropic and OpenAI do limited access to what are much more powerful models. And they argue the limited access is a guardrail in and of itself. But it means that for researchers, bug bounty players, etc., it's closed to them. What is the right pathway forward? Look, I think we need to be responsible with the power of these new models coming online.

40:05And it reminds me of like responsible disclosure in cyber, right? When you find a zero day, you have to report it responsibly before you share your work with the world. I see a very similar pattern here, and I'm actually very supportive of what the model vendors are doing and trying to bring together the right folks, the Linux Foundation included in Project Glasswing, to make sure that people get an opportunity to see what these models can do, to put them to work, and to make sure that they're building up tools that can defend against them if they're used for malicious purposes. Again, I think you need a lot of different perspectives, organizations, and quite frankly, just brains around this problem to make sure that we're building the right controls to make sure that we can get all the value out of AI without sacrificing our defenses.

40:51Value out of AI. At this moment, Erica, how are other bets of yours, other portfolio companies managing to get the value? How are we seeing this being adopted within teams, embraced and pushed forward? Yeah, look, I think the best companies are adopting AI across the stack. Ramp has had a number of fantastic posts. They're completely changing their entire culture by enabling their company to build with AI across the stack. We have companies like Railway, for example, that run all of their go-to-market with AI. So it's not just about coding where, of course, we're seeing incredible acceleration in product shipping, evolution, new feature development, et cetera.

41:34But also the go-to-market teams are building their own tooling now with AI. I mean, I'm building my own tooling for my work in venture, right? And so I think the companies that get ahead are the ones who are going to drive AI usage throughout their organization across finance, HR, support, go to market, and not only code, which is obviously a huge opportunity in and of itself. Maybe have some bandwidth for some fun marketing, as we've all been liking the Lagora new ad of late. Mr. Jude Law, Erica Brescia, we thank you very much of Redpoint Ventures. Thank you. Great to have you. Coming up, the CEO of Slash joins to discuss his startup's newly minted unicorn status that aims to take on fintech rivals like Ram.

42:18We're just talking about them, like Brex. This is Blue Bag Tech.

42:27Caroline Hyde:Support for the show comes from Public. Lately, it feels like there are two types of investing platforms. Some are traditional brokerages that haven't changed much in decades, and others feel less like investing and more like a game. Public is positioned differently. It's an investing platform for people who are serious about building their wealth. On Public, you can build a portfolio of stocks, options, bonds, crypto without all the bugs or the confetti. Retirement accounts? Yep. High yield cash? Yes again. They even have direct indexing. Public has modern design, powerful tools, and customer support that actually helps.

43:02Caroline Hyde:Go to public.com slash market and earn an uncapped 1 % bonus when you transfer your portfolio. That's public.com slash market. Ad paid for by Public Holdings. Brokered services by Public Investing, member FINRA SIPC. Advisory services by Public Advisors, SEC Registered Advisor. Crypto services by ZeroHash. All investing involves risk of loss. See complete disclosures at public.com slash disclosures. Deadlines move, plans change, and sometimes opportunities pop up out of nowhere. When you need branded gear fast, 4imprint is ready to deliver. 4imprint offers hundreds of promotional products in their 24-hour category.

43:39Everything from custom apparel, bags, and drinkware to writing tools, trade show staples, and high-tech gear. At 4imprint, they're focused on getting the details right, printing your logo with precision, packing your order with care, and shipping it out fast. And it's backed by their 360-degree guarantee. That's 4imprint's promise your order will show up right on time, just the way you planned it. That's what it means to be 4imprint certain. So, if you're prepping for a last-minute event or jumping on a big opportunity, you don't have to settle or scramble. With 4imprint, fast, reliable service and peace of mind are built right in.

44:16Check out their full 24-hour selection at 4imprint.com. 4imprint. For certain. From coast to coast, unlock adventure at Red Lion Hotels by Sinesta, where restful sleep, friendly service, and trusted local knowledge are part of every stay. Red Lion makes it easy to feel welcomed, comfortable, and connected wherever the road takes you. Whether you're traveling for business or pleasure, you can spend less and make more of every trip. When you sign up for Sinesta Travel Pass, you'll get their best rates instantly. Go to Sinesta.com to book your stay and unlock the best rates with Sinesta Travel Pass.

44:51Here today, roam tomorrow. Join now at Sinesta.com. Terms and conditions apply. Slash Financial has just raised$100 million from backers, including Ribbit Capital, Kozla Ventures. The startup offers banking financial services, along with features tailored to specific industries. Medical spas, construction firms, sneakers is where it started. Slash's co-founder and CEO, Victor Cadenas, joins us now. I wish we could just talk about Allbirds with you, and it's resurrection into an AI company, Victor. But I'm interested as to what the$100 million helps you do. You're going global. Where are you leaning into?

45:25Caroline Hyde:Absolutely. First of all, Caroline, thank you so much for having me on as a long time watcher of the show. It's such a privilege to get to participate for the first time. Thanks, mate. Succinctly, there's two things this funding enables us to do. Serve more markets and serve more geographies. Quite core to the Slash thesis and strategy, as you mentioned, is this idea that business owners deserve banking products hyper-tailored to the needs of their specific industries. Slash right now is valuable to businesses in over a dozen industries, but we want Slash to be a household name. We want to bring it to as many possible entrepreneurs in America and beyond.

45:57Caroline Hyde:Right now, we're living in a super interesting time in fintech history where for the first time ever, it's very easy and possible for an American fintech company like Slash to serve businesses all over the world. Stablecoins are this really enabling force that allow us to have a default global product. And we want Slash to reach business owners worldwide. And that's what this funding is going to go towards. So you're raising$100 million. But I think you guys would be at pains to say we're a really small team or a lean team is how you'd put it. you're at$300 million ARR, right? Which is something to note.

46:28Caroline Hyde:If that's the case, where does that$100 million go? Like, why do you need to deploy it? Absolutely. I think you're absolutely right. We have a very lean team. Our company's been profitable since May of last year. But in the age of AI, because it's so easy for anyone to build software, brand is more important than ever. And banking is very much a relationships game. There's three reasons customers come to Slash. The first of all is we have a phenomenal product. The second of all is we offer phenomenal rewards. And the third is we've brought the community bank model to fintech, all else being equal.

46:59Caroline Hyde:People like to bank and spend with an institution they have the best relationship with. So if we want to reach and deliver an amazing level of support to business owners all over America and all over the world, that's going to take a lot of people and a lot of investing in our brand. Victor, just go back to how lean the actual startup is. I think you've only got 68 or thereabouts employees before you start adding more to this go-to-market and marketing strategy. How much are you enabled by AI? How much are you thinking about keeping your cost base low? Entirely, this business, the way that we run it would not be possible if it were not for AI.

47:33Caroline Hyde:I think that's the thing that Ribbit, that Kosa, that Goodwater, all the investors that invested in this past round found most attractive about Slash. It's this idea that over 50 % of the engineering hours at our company are spent actually building internal software to help us be much more efficient. So a lot of these tasks that the legacy banks or fintechs that came before us have thrown copious headcount at are just agent automated away at Slash. Everything from parsing documents when somebody is applying for an application or processing disputes or responding to requests for information from a bank partner.

48:05Caroline Hyde:All these back-end tasks are agent automated away at Slash. So, Victor, kind of as your elevator pitch for Slash, who are the incumbents that you're trying to displace in helping us understand, like, if you were a customer of Slash, how it would work and what it is you offer? Absolutely. Well, Ed, I think a very underappreciated fact is that fintech, particularly B2B fintech in the U.S., is extremely underpenetrated. Less than 5 % of businesses in America, bankers spend with a fintech. The vast majority of them still use a legacy bank. And the reason our offer is resonating so much is quite simple.

48:38Caroline Hyde:We live in an age where machines can think, where cars can drive themselves. But as I mentioned, the vast majority of entrepreneurs are still banking with an institution that has an interface that was last updated in the year 2003. Core to the slash thesis is this idea that your bank account can do much more for you. It can be the place where you invoice your customers, where you reimburse your employees for out-of-pocket expenses, where you get insights around how to cut wasteful spend and much, much more. We have ambitions to become the most popular business banking platform in America, and we're going to do it by building a very powerful platform that puts a lot of money back in the pockets of our customers and gives them back a ton of time.

49:13Caroline Hyde:Victor Carlenia, CEO of Slash. It's great to have you on Bloomberg Tech. Thank you very much. We're on IPO watch this week. It's not the reason I'm in New York, but we're here for it. Shares of Madison Air are indicated to open$30,$31 a share. The IPO pricing at$27 a share. We're waiting for trading to begin. It's a ventilation and filtration systems company, but it's seeing increasing demand for its services amid the boom in data center build out. Set to have, believe it or not, Caro, the biggest US IPO of the year so far. So far. Big asterisk. Two billion. But why don't CEO Jill Wyand? I mean, she's coming on the network later really to talk about the lead up, the U.S.

49:55industrial focus. And look, all of these companies suddenly de facto becoming AI companies because there's sheer scale of need for infrastructure right now.

50:03Caroline Hyde:So here's a data point for you that kind of explains where we are at in the world of AI. This will be the biggest U.S. industrial IPO for almost three decades. Wow. It shows you the transformation that these kind of legacy industries have had, HVAC being one example. But yeah, we'll keep an eye on it. Indicated to open$30,$31 a share IPO at 27. There's a few others too this week. Ed making HVAC sexy again. That's what he does. That does it for this edition of Bloomberg Tech. And I'm so pleased to be here for another day. Yeah, yeah. We've got a lot going on. It is a very busy week. Lots of news in the show.

50:36Caroline Hyde:Recap it on the podcast. You know where to find it. It's on iHeart, Spotify and Apple. and it's on the Bloomberg Terminal and on the Bloomberg platforms. From New York City this week, this is Bloomberg Tech.

50:49Caroline Hyde:If you follow markets, you know the value of long-term thinking. You plan, you diversify, you prepare for volatility. But even the best strategies can't prevent every bad day. For more than 75 years, Cincinnati Insurance has helped individuals and businesses navigate tough moments with expertise, personal attention, and independent agents who focus on relationships, not transactions. The Cincinnati insurance companies. Let them make your bad day better. Find an agent at c-i-n-f-i-n dot com. Deadlines shift, plans change, and sometimes you just need promo products fast. Turn to 4imprint. 4imprint has hundreds of promotional items available with 24-hour turnaround.

51:36From custom apparel and drinkware to trade show gear, writing tools, and more. And their 360-degree guarantee promises your logo will be printed with care. Your order ships fast, and it'll show up right and on time. That's the certainty of 4imprint. Check out the full 24-hour selection at 4imprint.com. 4imprint. 4certain. Find home wherever you roam at Sinesta ES and Simply Suites, where longer stays feel comfortable, flexible, and easy. Stretch out and enjoy spacious accommodations and home-like amenities designed to help you settle in and stay productive or relaxed for however long you need.

52:13And when you're a Sonesta Travel Pass member, staying at Sonesta ES and Simply Suites means earning points toward free nights, upgrades, and more with every eligible stay. Go to sonesta.com to book your stay and unlock the best rates with Sonesta Travel Pass. Here today, roam tomorrow. Join now at Sonesta.com. Terms and conditions apply.

From the publisher

Bloomberg’s Caroline Hyde and Ed Ludlow discuss Anthropic's updated version of its AI model, Opus 4.7, released just a week after its limited release of Mythos. Plus, Elon Musk is kicking his Terafab plan into high gear, even as skepticism grows from the semiconductor industry. And TSMC reports a big surge in profit and raises its revenue outlook for 2026, driven by strong demand for AI chips.

See omnystudio.com/listener for privacy information.

More from Bloomberg Tech

All 343 episodes
Anthropic Unveils Updated Opus 4.7 ModelBloomberg Tech · 43 min
Listen in VO