Open-Weight AI Debate Takes Center Stage

0:00With the highest number of young STEM graduates per capita in the EU, Ireland has the people and skills your company needs to succeed here. IDA Ireland, the National Investment Development Agency, can help you find and nurture the people you need to internationalise and thrive. Our talent is just one of the extraordinary benefits Ireland has to offer. Learn more at idaireland.com. Invest in extraordinary.

0:30Aging is real. And so are the benefits of adding Vital Proteins Collagen Peptides to your daily routine. Because around the age of 30, your body needs backup to keep your collagen up. To help support healthy hair, skin, nails, bones, and joints. Available in the classic collagen peptides. Collagen and protein shakes. And new Vital Proteins Collagen Sparkling Waters. So you can stay vital, stay you. Visit VitalProteins.com to learn more and where to buy. These statements have not been evaluated by the Food and Drug Administration. This product is not intended to diagnose, treat, cure, or prevent any disease.

1:00Get the news you need in just 15 minutes. Start your day with Bloomberg Daybreak, the podcast with a global view on the stories that matter. I'm Nathan Hager. And I'm Karen Moscow. Join us each morning for curated stories on current events, politics, business, and foreign relations. Plus one conversation on the day's biggest developments, all in just 15 minutes. Subscribe to Bloomberg Daybreak for a precise, thoughtful take on the stories that matter. Listen to Bloomberg Daybreak each morning on Apple, Spotify, or anywhere you listen. Bloomberg Audio Studios. Podcasts. Radio. News.

1:43Bloomberg Tech is live from the heart of Silicon Valley with Ed Ludlow in San Francisco.

1:52Ed Ludlow:This is Bloomberg Tech. Coming up, tech's biggest names are making the case for open AI, arguing open weight models are key to innovation, competition, and even security. Plus, Intel showing signs of an AI-driven comeback, forecasting stronger sales as its data center business accelerates. And AMD and Cerebris are teaming up to challenge NVIDIA, promising some of the fastest AI systems on the market. We speak with Cerebra CEO Andrew Feldman. Our top story this Friday, a major push to shape the future of AI. A coalition of technology companies, CEOs and venture firms is urging Washington to embrace open weight AI models, arguing they're critical to American competitiveness, innovation and national security.

2:38Ed Ludlow:The signatories include Microsoft Satya Nadella and NVIDIA CEO Jensen Wang, who used his very first post on X to share the letter. It argues, quote, the United States should lead in building an open AI ecosystem because it expands opportunity, strengthens competition and extends American technological leadership. Bloomberg's tech managing editor Sarah Fry is with us. The idea of open weight and open source, that's not new. But it was very, very interesting, the timing of the heavyweights of this industry coming out almost in unison and saying we need to think about this area. I mean, it's a huge issue because right now, and they didn't say it in the letter, there is a really intense discussion happening about Moonshots Kimi.

3:26This is a model that just came out of China that is pretty equivalent with U.S. models, and it's changing the way people think about it. Whereas previously, the Chinese models were good open-weight models and able to undercut U.S. models on price, and they were very popular. But they didn't actually get to the point where the innovation looked like it was about equivalent with what you could get from U.S. And that's caused a lot of concern in Washington, maybe some policymakers threatening to change things, restrict things. And, of course, U.S. startups really depend on those open weight models.

4:05So if the U.S. were to have something that was equivalent, that would really help a lot of a lot of the companies. But, you know, everything's getting built right now. And we have a lot of dependence on these Chinese models. I'm very curious what's going to happen in Washington as we digest what this means for the market.

4:25Ed Ludlow:I don't want to trivialize this, but it was fascinating that Jensen Wang elected to create an X account and use his first post as a mechanism to communicate this. For Jensen, this isn't new. He's talked about the importance of open models, open weight, AI is open source. The distinction is if you use a closed model, you go via the API, but you can't see the model. You can't download it. In open weight, you can download and run the train model. The data code aren't necessarily public. On open source, all of it's there available. You were absolutely right to frame this post Kimmy K3. But also with what happened with the open AI mistaken hacking of Hugging Face, it also raised some of the limitations of U.S.

5:14Ed Ludlow:open models because Hugging Face wasn't able to defend itself in that respect. What do you think happens next? How serious is this? Well, I think that it is serious because the whole market is resting on this idea that all of this tremendous capital expenditure investment, the build-out of data centers, it's all going to end up having a return on investment for U.S. companies. There's a whole other factor we haven't talked about yet, which is some concern that these Chinese models are built in part using unauthorized distillation of U.S. models, including Anthropics Cloud. So if all of that investment from the U.S.

5:53companies in their future innovation is then just being used to boost what China can accomplish, whether it's through distillation or through unauthorized chips use, that's a matter of big debate too. And that could really affect the ROI for these companies on the many, many billions that they're pouring in to build out of the future of AI.

6:19Ed Ludlow:That's probably the bit we don't know. You know, why did they do this and why now? Was it to counter distillation? Was it to try and get ahead of what happened with OpenAI hugging face? We'll find out. Bloomberg Sarah Fryer, who leads the team here at Bloomberg Tech. Thank you very much. Let's turn to earnings. Shares of Intel. It's interesting. Now down 3 percent. When the earnings hit last night, we saw the stock in after hours go as high as a gain of 13%. The chip maker delivered a stronger than expected forecast as demand from AI data centers is fueling a surge in its service business, particularly CPU.

6:56Ed Ludlow:CEO Lit Bhutan telling me CPU demand is now outpacing what is improving supply. Joining us now is Antoine Shkaivan, head of global technology infrastructure research at New Street Research. research. He has a neutral rating on the stock with a$115 price target. I want to get into the technology. I want to get into what Intel is actually doing and well. But the stock, you know, I was sat here last night and the gains were massive. Then they eroded. Now we're down 3%. With time, what has the market changed its mind about? I think, I mean, thanks for having me on the show, Adam. I think, you know, what's happening is the market is realizing that actually that's the, even the bull case is already priced in.

7:43Ed Ludlow:Even if you take like an Intel in 2030, where manufacturing is fully turned around, which by the way, we had indications of that yesterday on the print, you have a foundry EBIT margins up nine points sequentially. So they're headed really towards break-even. Even if you assume that in 2030 Intel is fully recovered, they're going to still have some structural cost disadvantages compared to TSMC. And gross margins are never going to reach the same level as TSMC. And in products, even if they're doing extremely well, and even if the overall agentic tide is lifting all boats, Intel is probably going to keep losing share to AMD.

8:28Ed Ludlow:It's probably going to lose share to ARM. You know, you have these hyperscale in-house designs. You have NVIDIA ramping, you know, Grace, Vera, etc. All of that is going to take share from Intel. And if you put all that together, even the bull case, you know, doesn't make much room to the current levels the stock is trading at. Okay, Ed, calm down. remind yourself and the audience that going into the print, the stock was up 170 % year to date. So that might be a part of it. What's interesting as well is that Intel never said they'd fixed everything, just that they were making progress. So the way that Lit Bu explained it to me is that, you know, their processes and production for their own products is improving.

9:10Ed Ludlow:But they're still in a place where demand for CPU in particular is outpacing what is improving supply, right? Right. And you'll remember last quarter, they basically left money on the table because they couldn't meet the demand. Do you recognize that kind of operational progress? Yes. I mean, we heard a lot of very encouraging data points on, you know, capacity increases on the call. First of all, you know, you have yields, of course, that are improving, that directly increases, you know, the output of the good chips that you can produce. On top of that, Intel has a lot of shells that they've been building over the last few years that they can now fill up with equipment.

9:48Ed Ludlow:All of that puts Intel in a good position, you know, to, well, number one, as you said, you know, address the huge demand that there is for CPUs in this agentic era. And two, potentially start addressing, you know, demand from external customers as well. You know, they targeted 15 billion external foundry revenues in 2030. And I think there's increasing evidence that it is actually pretty tangible. So let's talk about CapEx, which will be more than$20 billion now this year and grow over the next couple of years. Lit Bhutan, the Intel CEO, is cautious. He won't deploy CapEx unless he thinks there'll be a return.

10:25Ed Ludlow:How did you interpret all that? Well, I think it's very reassuring for investors to hear that. I think in order to generate that demand, Intel still has to execute. Now, let's keep in mind that EBIT margins for boundary are still in negative territory. Gross margins are probably also still in negative territory. So they still have a lot of progress to make to end up in a situation where they can actually really invest that capex with the confidence that they're going to be able to sell those wafers and turn the profits out of selling those wafers. um now in addition you know like the sort of you you alluded to the 20 billion dollar comment a lot of that is going to be of course for for equipment because as i mentioned they have a lot of tools not space yeah tools not space yeah exactly for equipment and now the question is what happens in 2027 and for that i think uh it's still open-ended you know it depends on whether we can uh you know keep improving those yields keep improving the cost structure of these where it fers and get into positive margin territory for boundary.

11:29Ed Ludlow:Let's widen this out to what's happening in infrastructure. What do we learn from Intel's print about the AI cycle? Well, of course, like the DCAI print is impressive. Like they're growing nearly 60 % of the DCAI business. I think an interesting data point is that Intel also still has a lot of room, to improve their roadmap on the AI XPU front. We're talking about a 2 billion run rate today for everything that's not CPUs in the DCAI segment. They think they have visibility to 4 billion. That's orders of magnitude smaller than, for example, like an NVIDIA that's approaching a$400 billion annual run rate for their data center business.

12:20Ed Ludlow:So I think that's an interesting data point also that Intel still has plenty of opportunities maybe to gain some momentum there. On the CPU front, I mean, it's an excellent 3D cross for AMD. If Intel is growing that quickly, despite, I think, a roadmap that's not that competitive, like if you look at the density of the chips that they're putting out, like the number of threads, the number of cores, which nodes they're manufactured on. I think AMD has a much stronger roadmap. So it's definitely a strong read-a-class for AMD on the CPU front. Anton Schraiban of New Street Research back on Bloomberg Tech.

13:01Ed Ludlow:Thank you very much. Coming up on the show, Scribe Therapeutics, the company behind gene editing CRISPR technology IPO today. And look who's joining the show, Scribe Therapeutics co-founder, scientific advisor, Jennifer Doudna. That's next. This is Bloomberg Tech. Thank you.

13:47Invest in extraordinary.

14:17Subscribe today on Apple, Spotify or wherever you listen.

14:29Ed Ludlow:Jennifer Doudna won the Nobel Prize for pioneering gene editing technology CRISPR. Today, Scribe Therapeutics, the company she co-founded, just raised$128.7 million in its IPO for a much bigger ambition, using CRISPR to prevent heart disease. Scribe Therapeutics co-founder and scientific advisor Jennifer Doudna is with us now. I've been reading so deeply about this. I think the best place to start is that your lead therapy aims to lower LDL cholesterol with just a single treatment. But distinct and unlike earlier gene editing approaches, you say it does not permanently edit a patient's DNA. Explain why that's such an important distinction and move forward in the technology.

15:14Well, great to be here, Ed. It's an exciting moment in gene therapies because, as you just said, Scribe Therapeutics has a strategy that involves what we call epi-editing. It means making changes in DNA that aren't permanent but alter the production of proteins. We think this is an incredibly important moment in the field because it means we can now use this therapy safely and effectively for common disease.

15:41Ed Ludlow:Jennifer may I ask you know in any IPO there's obviously a reason to do it and raise the proceeds right now where are you in in the cycle and the process and what will you use those proceeds for well Scribe Therapeutics is a clinical stage company we're already in our phase one trial for the first indication the funds that are raised today will help move that forward and progress these therapies so that they can be deployed much more widely. At the start of our conversation, you explained the technology difference. When it comes to regulators, how important is that distinction going to be versus the sort of permanent editing of one's DNA?

16:24Well, Scribe is a company that was really founded on great science and has an amazing scientific team. They've invested the last several years in improving the technology to the point where it's safe and effective to treat disease as a strategy for making non -permanent changes in DNA. So we think that this is going to be an effective way to prevent cardiovascular disease in the future that will be so safe that you can provide it much more broadly.

16:58Ed Ludlow:That's what we're talking about. Let's bring it back to basics, heart disease. And it's amazing, millions of people across America and across the world take statins, right? Every single day. But then there's a lot of evidence, a lot of data to support that then they stop. This is a one single treatment that you're working towards. And I wonder just how important that will be for the behaviors of the patients receiving it. Right. We think about this a lot. I think it's very interesting to think about a way of treating or preventing disease that doesn't require taking a pill every day, right? That is really effectively a one or few time treatment.

17:39That means that people are free of having to remember to take a pill and they don't have to worry about those kinds of side effects. It's a really interesting moment where I think we're going to see a real change in the way that medicine is delivered in the future.

17:54Ed Ludlow:When the Bloomberg Tech audience hear the phrase gene editing, inevitably people have concern. What is it from a safety perspective that they need to understand? Well, the CRISPR technology is built around a strategy for targeting DNA precisely. With any technology, there's always risk. But I think what's exciting about Scribe's approach is that they've invested several years in ensuring that their strategy, this epi-editing approach, is really, truly safe and effective for the indications they're going after. And as we discussed before, this is not a technology now that makes a permanent change to DNA.

18:34And I think that's a real difference.

18:37Ed Ludlow:You had one of the most interesting and engaging conversations I've listened to this year with Bloomberg's Emily Chang on the circuit. And you talked about AI. and you had a level of skepticism about where AI really is today across drug discovery, its utility in your field. I know it's not been that many weeks or months, but have you changed your mind? Do you still hold that position on the limits of AI in your field right now? Look, I think AI is an incredible tool. It's an amazing technology that allows scientists now to accelerate the kinds of work that we do. That being said, it doesn't replace scientists.

19:16We still need innovators. We still need creative people. We still need the new ideas that come out of left field. And we don't see that with AI. We see it being a great way to accelerate our ideas and make it possible to do things faster. And I think it's a great partnership.

19:32Ed Ludlow:Jennifer, bear with me if you would, but we have some headlines that are just crossing that right now Scribe is indicated to open at about$29 a share. The IPO priced at 15. So interesting because this was a difficult environment for biotech IPOs. We're also learning that the allocation, about 75 % of the shares to just 10 investors reflect on that. I mean, this must mean a lot to you, but clearly, you know, investors are very receptive to this. I think our investors really get this technology. They understand what's different about what Scribe is doing. They understand the quality of this team, its leadership and its commitment to real science.

20:14And that's what we're seeing, I think, reflected in the marketplace.

20:19Ed Ludlow:Scribe Therapeutics co-founder Jennifer Doudna, thank you very much for your time here on Bloomberg Tech. A story I just want to bring you real quick in the markets. BlackRock began marketing$12.3 billion of high-grade bonds to fund a meta data center project, testing investor appetite. This is concerns grow about excessive AI infrastructure spending, a source told Bloomberg. The bonds are being offered by a superpilot investor holding company tied to BlackRock and JP Morgan Chase and Morgan Stanley are running the offering, which is expected to price next week.

20:59Ed Ludlow:It's time for Talking Tech. First up, the EU escalated a probe against TikTok, accusing it of failing to protect the safety and privacy of teen users. The commission said TikTok's custom accounts for users under 18 can be easily found and viewed by other people online, which goes against the block's content moderation rulebook. Plus, SoftBank is considering an acquisition of Gravis Robotics in an effort to target the AI technologies underpinning robotics, according to sources. A deal could ultimately value Gravis at more than$500 million. And SAP reported stronger than expected cloud growth as more customers adopt AI-powered enterprise software.

21:42Ed Ludlow:Bloomberg spoke with CEO Christian Klein earlier about that, but also how AI tokens are now managed like any other business expense. In the R &D cost, you see a higher token consumption as we are using AI to code additional features, but especially now additional agents. But vice versa, you also see a huge productivity increase of 30%. Obviously, we are also managing tokens as part of our cost budget. You have a headcount budget, you have a party budget, and you have tokens. So our managers in R &D, in the go-to-market space, they have one budget, and they have to manage that also according to the productivity assumptions we reflected in the budget.

22:24Ed Ludlow:After two postponements, all eyes are on SpaceX's Starship test flight today. The mission includes upgraded Starlink satellites that will intentionally burn up during re-entry. The launch comes as SpaceX makes a major strategic bet on Starship. Bloomberg's learned the company is already turning away some potential Falcon 9 customers beyond 2028 as it shifts resources to its next generation rocket. That's all according to sources. Bloomberg Sana Pashanka broke the story and joins us now. Really interesting to work with you on this one. There's a lot of detail in there, right? Tell us about what we've learned SpaceX's attitude is to those satellite companies coming to them wanting either a dedicated ride on Falcon 9 or a rideshare ride.

23:12Ed Ludlow:What do we know? So what we know and what we've been hearing from sources and customers is essentially that SpaceX is fully booked through 2028 and they're turning away customers for dedicated Falcon 9 launches past that time point. So a dedicated launch is when a satellite company will buy the entire rocket and launch multiple of their satellites to orbit. And they're also turning away customers for their ride share missions, which is when multiple satellite operators can all put one or two satellites and hitch a ride to orbit to test satellites or if they only need a couple to get to space. And yeah, this is a really big deal for a long time.