Nvidia Earnings In Focus; SpaceX Eyes Historic IPO

21:44Ed Ludlow:Jay, case study, wire harness, right? That's something we talked about a lot in this program. Heavy dependence in that case study on China. give me a similar product a similar sort of component that you are solving for where there might be a reliance on an external country to get it yeah um you know i'll give you two examples um one example are the sensors that we put on to our you know put on put on to almost all of our industrial platforms um today a lot of those sensors uh are sourced the sensing elements are sourced um from offshore uh you know offshore companies um and nations allied with them uh You know, that's one example.

22:24Another example would be, you know, capacitors and other passive electronic electrical components. These are often components that have been offshored to Asian countries over the past few decades. And America today has a very, very low dwindling supply base for these types of engineering components. And so those are two examples of areas where we are designing in the country and then manufacturing into the country those components for the warfighter.

22:49Ed Ludlow:Jay, I just have a few rapid fire questions, if I may. What year did you found this company? 2024 in November. 2024 in November. And is this$1 billion valuation post-money or pre-money? That is post-money. So in about 18 months, you founded, launched and scaled up this company and got to a billion dollar valuation. What should I infer from that? The pace at which you've done that? You should infer that, you know, we're living at, I would say, the largest gap between what America needs and what America is able to produce in generations. It's a generational opportunity for folks who are looking to build in this country.

23:31I think over the next five to 10 years, you will see a lot more manufacturing companies in America building for America. and we're just sitting at the forefront of this, I think, immense growth curve that's coming and is already happening in the nation, both for defense and aerospace customers as well as in other categories like AI infrastructure, energy, robotics, etc.

23:53Ed Ludlow:Jay, who are your customers? Due to the sensitive nature of the work that we do, I can't disclose any specific names, But some of the names that we have been able to disclose are Boeing, Airbus, Honeywell, Lockheed Martin, and then certain branches of the military. But all of our components today that go on to platforms we all know about, 737 MAX, 777, F-35s, F-15s, F-16s, our tanks, MK1 Abrams. So there's plenty of different platforms that we're on today that the country actively uses. You know, it's interesting. We've probably spent a lot of time understanding some of the problems that those customers that you just outlined have.

24:42Ed Ludlow:But probably of Amco, it's better to talk about what your constraints are, where the bottlenecks are for you in your own supply chain. You know, in our own supply chain, there are certainly, you know, big gaps right now at the sort of like infrastructure level. Like I mentioned, you know, for sensors, for example, you know, we need sensing elements. These are highly precise elements that go into sensors that the country doesn't have a lot of suppliers for. So in those cases, we're taking a new approach. We're vertically integrating a lot of our supply, designing that in-house and then manufacturing it in-house to solve gaps where the domestic supply chain is not able to supply those components.

25:24Ed Ludlow:Jay Malik of Amca, raising$300 million,$1 billion valuation. Thank you very much. Plenty of other news headlines for the show. it's time for Talking Tech. First up, OpenAI is planning its first international flag. The ChatGPT maker committing$234 million to launch a new facility in Singapore, its first so-called applied AI lab outside the US. Partnering with local authorities, the hub will hire over 200 engineers to embed advanced models straight into healthcare, finance, and public infrastructure. Plus, Alibaba's matching NVIDIA's breakneck pace, its chip unit, T-Head, just unveiled a new AI Accelerator, packaging 144 gigabytes of memory for complex autonomous agent tasks.

26:04Ed Ludlow:Alibaba plans to publicly list the business to tap investor appetite for Chinese silicon alternatives. And Google is redesigning its iconic search box and adding new AI coding tools, the biggest update to the product in more than a quarter of a century. At Google I.O. event on Tuesday, the company also rolled out several new tools for developers to write code using AI and to manage agents. Coming up, we're counting down to NVIDIA earnings later today after the closing bell. And up next, we'll discuss its role in the wider chip ecosystem with Bailey Giffords, Paulina McPadden. This is what markets look like right now.

26:46Ed Ludlow:And NVIDIA is central to that story. Nasdaq 100 rebounding up 1.4%. Semiconductors after three straight days of heavy declines, almost in correction territory, rebounding up 4%. NVIDIA is now near session highs, up more than 2%. Half time in the program. Stay with us. This is Bloomberg Tech.

27:15Ed Ludlow:Welcome back to Bloomberg Tech. Nvidia earnings on deck. The company reporting its earnings results after the market closed. We're near-session highs, up more than 2%. And for many watchers, China is top of mind. The Chinese government has to decide how much of their local market do they want to protect and how much of their local market do they want to expand with more AI capacity. My sense is that the demand in China is so incredible. Just like it is here, agentic AI is also making enormous progress there. My sense is that over time, the market will open. Two themes, what's happening with China and demand outpacing supply around the world.

28:01Ed Ludlow:Let's discuss where NVIDIA sits in the global chip ecosystem with Paulina McPadden, Investment Manager of International Concentrated Growth Strategy at Bailey Gifford. I'm so excited to talk to you because, yes, NVIDIA is in your strategy, but so is everything else around the world. And what Jensen outlined there on China is so interesting. The base case assumption right now for NVIDIA is zero access to China. Maybe we'll get some clarity. There's everything else that comes with that. And I just wonder how much attention you really pay to NVIDIA's ability to service the China market or not. I mean, it's a really good question.

28:36And I think the honest answer is it's not really factored into our upside cases for NVIDIA for a while because the problem there is you're having to try and prejudge what politicians might want to do at some point in the future. So H200s have been approved for export to China since last year, but it's the Chinese government's decision not to approve purchases by Chinese companies that has stopped that sale going through for NVIDIA. Now, I think the really interesting long-term question for us is actually what that means for the broader supply chain in the long term, because is there a case that China could build up its own domestic chip supply chain?

29:11Now, long term, maybe that's possible. Maybe it does threaten NVIDIA eventually. But two things give me comfort on that over at least the next five years. One of those is that the existing high power chips in China are still a fraction of the performance of NVIDIA's leading chips. So I think that gives non-Chinese companies and AI researchers a leg up for the long term. And the second thing is it's actually really, really hard to build up a scaled and complex semiconductor supply chain. And it's not as easy as just putting capital into it as China has done in the past with things like EVs and immersion to global dominance in the battery supply chain, for example.

29:48These are extremely complex bits of equipment that require cooperation around the world. And that's just not something that I think you can do year on year if you're only relying on the domestic demand that you get from China and you're not able to sell internationally.

30:03Ed Ludlow:We just showed SK Hynix is a part of your strategy. And, you know, what we played to you a moment ago was a conversation I had with Jensen Wong on Monday. And it seems memory is still the bottleneck. But the way that Jensen outlined it to me is three years ago, he went to the memory makers, including SK Hynix. And he said, this is what I think the world will look like in a few years time. And he tried to convince them to invest in permanent capacity. You know better than I do, Paulina. Memory has been boom and bust, cyclical historically. In your strategy for SK, is that different now? Is there permanency to the capacity that's needed?

30:39So I think bottlenecks is a really interesting framing because people tend to think, oh, a bottleneck is extremely investable and that can lead to a lot of growth and a lot of upside. And that's true to a certain extent. But I think markets are actually quite good at, in some cases, recognizing bottlenecks well ahead of time and investing in them before the fundamentals even come through. So as long-term investors, what we're trying to do is identify these structural opportunities or companies that become even better as they scale and deepen their competitive position in the ecosystem. So I'd contrast SK Hynix here on the one hand with TSMC because to some extent it's a portfolio construction question.

31:16So I think that there is a potential that SK Hynix becomes a substantially better company over the long term. The fact that they are able to extract these long-term purchase agreements from customers now is really a substantial difference over history. The industry has consolidated. It's not as irrational as it used to be. And I think that there is a case to be made that with the rise of HBM and in particular custom HBM, that creates a degree of lock-in with customers that means that these are no longer commoditized products as they have been in the past. But that's just a hypothesis. And I think we're going to see how that develops over time.

31:50And I'd contrast that with something like TSMC, which is actually the largest position in our portfolio, where I think they have demonstrated that extreme control and dominance over their entire supply chain and the ability to coordinate multiple actors in an increasingly complex ecosystem. And that means that as semiconductors become ever more complicated to produce, TSMC just keeps getting better.

32:11Ed Ludlow:Sometimes you can't say what the story is from just a few market sessions. But Friday through Tuesday's close, we were almost in correction territory. particularly in the chip sector, right? And you're coming to us from sunny Scotland. I was trying to make sense of what the net outcome was of the president of the United States trip to China. And without putting words in your mouth, I think that you would say investors now start to look outside of the US, look internationally. I would certainly hope so. There's such a rich hunting ground for AI-related names outside the US. I think there's a really interesting dichotomy that chips and AI is sort of designed in America, but it's manufactured internationally.

32:52Whether it's TSMC actually making the chips themselves, whether it's SK Hynix putting the memory inside the systems, or ASML that builds the EUV machines, the lithography machines for producing the chips at TSMC Fabs. They do that in the Netherlands. All of these companies are extremely interesting, again, extremely dominant in their particular industries, and that makes them great investments for a concentrated strategy like ICG. I think the other thing that we can do being based in sunny Scotland is take that step back and say, yes, the market is doing its thing. There's a tendency, I think, for the market to price a lack of a catalyst as a negative thing.

33:30The fact that there wasn't really any news about chip exports coming out of the Chinese summit was seen as a bad thing. But I think over the very long term, none of that affects the fundamentals of any of these companies. They are still extremely strong growing. And in fact, the fact that TSMC raised its guidance for AI growth over the next five years to 56 percent is really encouraging because, again, they have this insight into the entire supply chain. And so the fact that they're able to do that with a great degree of confidence gives us the conviction as well.

34:04Ed Ludlow:Paulina McPattern of Bailey Gifford from sunny Scotland putting a pretty sunny disposition on this reporter's face. Thank you very much indeed. From public markets to the private markets again, today's big take focuses on SoftBank's OpenAI bet. It's committed more than$60 billion to OpenAI, with founder Masha Yossi's son reportedly convinced Sam Altman is leading the most important technology shift of the century. But as rival Anthropic gains ground, questions are growing, even inside SoftBank, about whether the company may be too heavily tied to OpenAI's success. Bloomberg's executive editor for Global Tech, Peter Elstrom, has the story.

34:41Ed Ludlow:No surprise that around the world, the big tech take gets clicked on, right? Take us inside it. What are the details we're reporting? What do we learn about this dynamic between Masayoshi Sun, his fixation on Sam Altman, and what the rest of that organization is thinking? Yeah, so this is a deep dive into the situation at SoftBank. Masayoshi Sun, as you pointed out, has been investing in startups for a long time. He's had several enormous hits, including Alibaba was one of his biggest successes, the Chinese AI company. And then he went into the Vision Fund and he made hundreds of bets on small AI, small technology companies, some successfully and some not so successfully.

35:21And now with OpenAI, he's committed more than$60 billion in capital, his biggest bet ever. It's a very big concentration of the money that he's investing in a single place, more money than he's ever put in a single company before. He's not only selling assets, including some Nvidia stock, by the way, he's also borrowing some money to be able to make this commitment to open AI and give Sam Altman the kind of capital that he wants. Now, as you mentioned, there are concerns that are being raised, including inside SoftBank, that maybe Masayoshi-san is a little bit starstruck here. Maybe he's being persuaded by this very charismatic founder to put in more money than really should be at this point.

36:00So it goes into some of the details of those concerns. And as you point out, this is at a time when OpenAI is facing some strategy challenges, some business challenges, and also reputational challenges, as we've seen with Sam Ullman.

36:14Ed Ludlow:We're showing it$65 billion cumulatively through 2026. That's the reporting on the insider concern. What was SoftBank's response to the big take, Peter? Now, SoftBank and OpenAI, we should say, pointed out that they have a great relationship. They feel like it's a very strong partnership at this point. But what some of the people have told us is that there are a few areas of concern. First of all, SoftBank and OpenAI talked about this big Stargate venture that they were going to do in the United States. You'll recall it was last year when they got together with President Trump and they talked about how they can invest$100 billion, maybe even$500 billion in the U.S.

36:53Now, they have begun to make some of those investments, but they've gone very, very slowly. And in the meantime, OpenAI has struck some other deals under the Stargate name. They've cut some deals with other data center operators. So Masayoshi-san, who sort of viewed himself as like an equal partner who's going to be able to play a very key role at this company, is not getting the kind of stature and the attention that he would like. He doesn't have a seat on the board. He doesn't even have an advisor seat on the board like many of the other companies there. But there are concerns inside SoftBank and outside SoftBank about how this relationship is going so far.

37:28Ed Ludlow:Bloomberg's Peter Elstrom with the big take. Thank you very much. An update on the situation with Samsung's workers union, which has decided to suspend its planned strike. South Korea's Yonhap News Agency says the union will put a tentative wage agreement with Samsung management to a vote. This comes after days of stop and start negotiations, with the union threatening an 18-day walkout that was supposed to start tomorrow. We'll keep tracking that story. Now, coming up, Forum AI CEO Campbell Brown will be joining us to discuss how the major AI models are fundamentally unready to handle news and geopolitics.

38:06Ed Ludlow:And NVIDIA, earnings after the market close. That is the big story and factor in markets right now. This is what markets look like. At the index level, Nasdaq 100 up 1.4%. Big rebound after three days of declines in the socks, up 4%. NVIDIA is off its session highs, but up 2%. There's a lot riding. There is a beat and raise kind of expectation that Jensen Wong is going to tell us something big later today. This is Bloomberg Tech.

38:54ideas 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 build a one-of-a-kind index. 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.

39:31Ad 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. Sonesta Travel Pass makes traveling more rewarding. Designed to help you get more out of every stay. Sign up at Sonesta.com to enjoy instant savings, bonus points, and valuable perks like early check-in, late checkout, room upgrades, and free stays over time.

40:06With Sonesta Travel Pass, every stay brings you closer to your next reward. Choose from more than 1 ,100 hotels across 13 distinctive brands and unlock the best available rates when you book direct with Sonesta Travel Pass. Here today, roam tomorrow. Join now at Sonesta.com. Terms and conditions apply. 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.

40:51The Cincinnati Insurance Companies. Let them make your bad day better. Find an agent at c-i-n-f-i-n.com.

41:01Ed Ludlow:A new study by Forum AI says, quote, AI companies are grading their own homework. The report reveals a staggering 90 % failure rate on election questions across four major chatbots. and they are routinely serving up Russian and Chinese state media as authoritative sources. Joining us now is the CEO of Forum AI, Campbell Brown, who believes AI companies have to be different. The headline, Campbell, is that chatbots struggle with news accuracy, if we're going to make it as simple as we can. But I think the best place to start is with the methodology, what you went and tested and how you tested it.

41:41Sure. Ed, we looked at essentially three dimensions, which is factual accuracy, bias generally, and then the quality of the sources that they used. And the way we did it was training judgment models with senior domain experts who architected benchmarks. This is an area, I think it's worth pointing out, that hasn't gotten a lot of measurement. Most of the measurement and benchmarks that you see around the models is focused on coding and math and model capability, which makes sense. I mean, this is where the model companies are making their money. But they're also marketing the chatbots as consumer products.

42:17People are using them for all kinds of questions. And certainly in an election year, political information is going to be really important. And they're not right now where they need to be.

42:26Ed Ludlow:I want to get to the midterms. Just very quickly, the eval process is called NewsBenchWy, right? Yep. Do we have the why on why chatbots struggle with news accuracy? What's the cause? um i honestly i think it hasn't been a priority uh as i said they're they're leaning into other areas in terms of the measurement um i think that's going to change uh not only because consumers are are looking for it and and going to demand it eventually but i think enterprise is starting to demanding it to starting to demand it and that's again where um where their focus is from a business perspective and being just okay on accuracy overall on news and politics and geopolitics, it's not going to cut it.

43:08Ed Ludlow:Let's talk about the midterms then. The Bloomberg report states the chatbots answers about elections failed on accuracy, bias or source selection 90 % of the time. What I found interesting reading about that is there must be evidence, therefore, that the electorate, everyday Americans, go to chatbots for news in the first place. You're seeing those numbers increase. I think more and more. We've seen a couple of studies showing people are using them for news more and more. And we found that about a third of the questions we asked on election-related questions had factual errors in them. But all of the chatbots failed on bias.

43:46Claude and Jim and I gave left-leaning answers on election-related questions 100 % of the time. ChatGPT, 95 % of the time. Grok was the only right-leaning chatbot that gave right-leaning answers on these questions. I think about 85 % of the time. So you were by no means getting a straight up answer or a balanced perspective. And people are asking, you know, who are my candidates? What are their positions on the issues? And who should I vote for? So improving this between now and then. I mean, the one thing I am positive, hopeful about is there was a broad difference, I think, overall. Jim and I handled a lot of the questions better than some of the other models, which shows there's area for improvement.

44:30but you can't fix something if you haven't measured it.

44:33Ed Ludlow:Campbell, I know you previously, of course, at Meta, vice president of news and global media partnerships, but also as an industry colleague, award-winning journalist and anchor, CNN, NBC News. Both sides you see. Do you see action on the news side and the likes of Meta, other frontier labs, on actively fixing this, taking action on it? Yeah, I do. I mean, I talk to people in the labs on a regular basis. at all of these companies, and they do care about this and are beginning to approach it with a sort of different way of thinking. I think one of the challenges we had, and you said it in the opening, is today these model companies are essentially grading their own homework.

45:16There's not independent evaluation. And I'm not calling for regulation. I have a private company, but I do think we need an ecosystem of companies and nonprofits that are doing independent evaluation. And I think the companies, the model companies that lean into this are going to make ultimately more trustworthy products. And I do think you're going to see the demand move in that direction, again, from enterprise. You're already seeing some states pass laws where they're requiring independent evaluation, but I think it's increasingly going to become more important.

45:50Ed Ludlow:Campbell Brown, CEO of Forum AI. It's great to have you on Bloomberg Tech. Thank you very much indeed. Now, coming up, we're going to get back to the top story. What to expect from NVIDIA's earnings? The chipmaker off-session highs, still up just more than 2%. Top-line growth 80%. EPS growth 85%. The market wants more than that. From Jensen Wang, this is Bloomberg Tech.

46:26We've been licensed for many customers in China for H200. We have received purchase orders from many customers, and we're in the process of restarting our manufacturing.

46:42Ed Ludlow:That was Jensen Wong in March, singing a pretty positive tune about selling AI chips to China. This was the NVIDIA CEO striking a slightly different note when he spoke to us on Monday. The Chinese government has to decide how much of their local market do they want to protect and how much of their local market do they want to expand with more AI capacity. That needs to be reconciled. Some analysts are hoping to get more insight into NVIDIA's future in China in its earnings later today. There's also the battle NVIDIA faces with Custom Silicon. Kunjun Sabani from Bloomberg Intelligence here to talk us through those expectations, writing that NVIDIA has extended its AI lead even as custom ASIC competition grows.

47:24Ed Ludlow:And so I'm trying to understand this, right? We're talking about a world where we go from training to inference. And then the custom ASICs also, just as with the NVIDIA GPUs, are in a world where demand is outpacing supply. So we're trying to work out the market dynamic. What do you think we'll learn tonight? Yeah, I mean, look, tonight the standard beat and raise numbers, given where we are in the time and where the stock is, don't matter. unless they are able to pass through a$90 billion hurdle mark, which is the high bar for the guide for the next quarter. Now, that should really make the buy side bulls happy.

48:00I think tonight the key focus is going to be reassurance. There has been some noise around supply chain issues or liquid cooling with this Rubin ramp. We really need to hear that the Rubin ramp for the second half remains intact. And we would like to see that$1 trillion demand pipeline raise up the numbers from here.

48:18Ed Ludlow:So give me that 90 billion figure again. I just put FA goal on the Bloomberg back up. The estimate consent. Yeah, go for it. Go for it. When the guide for the next quarter, the high bar, it seems on the buy side is 90 billion. So if they're able to clear that, that should be a very positive event. Fiscal 2Q27 consensus is 87.3 billion. The high side for the guide in the current period being 90 billion dollars. So fascinating. You know, I think Jensen Wang's talked about the performance of an NVIDIA-based system against TPU, against other ASICs. He may well get asked again, is there any data set out there in the world, Kunjan, that evidences your argument that the Vera Rubin platform at least is extending NVIDIA's lead rather than seeing its market share shrink?

49:10there have been some third-party you know benchmarks and as we as the systems do roll out which they haven't yet there will be more benchmarks and we have seen consistently nvidia system the latest nvidia systems outperform what exists in the market today but now it's no longer just about your performance uh you outperforming on the specs right there's a use case there's a concept of dollars per token per performance so when we look at something like a tpu where the entire architecture of the data center is controlled and driven by the company using it itself, they are able to achieve a much lower dollar per performance, which matters to them a lot.

49:50Ed Ludlow:Real quick, the Bloomberg intelligence thesis on China. Yeah, we didn't hear anything positive coming out of the trip with the Trump administration. So, we are not going to increase or adjust any of our China revenue estimates anytime in the near term. Kunjan Sabani, Bloomberg Intelligence. Thank you very much. That does it for this edition of Bloomberg Tech. One last look, the countdown to NVIDIA's earnings. NASDAQ 100, session high, 1.5%. Sox rebounding up 4%. And NVIDIA is kind of the heart of that story. There is a lot of optimism, but Kunjan Sabani just said, raise, beat, doesn't matter. It's all about the outlook for the current period.

50:31Ed Ludlow:We're just over four hours away from NVIDIA's fiscal first quarter, 2027 financial year earnings. Recap on the podcast, you know exactly where to find it. And I would. A lot of quality conversations throughout the hour. Thank you very much for joining us. This is Bloomberg Tech.

50:49We buy insurance for peace of mind, but every year millions of claims are denied. Not because people did anything wrong, but because their policies quietly excluded what happened. Insurers know every detail. Policyholders rarely do. That's why my policy advocate exists. For just 27 cents a day, their platform reads your policies and explains where you are vulnerable. They don't sell insurance, they deliver transparency. Before you trust your policy to protect you, let My Policy Advocate tell you what it really says. Go to MyPolicyAdvocate.com. Sonesta Travel Pass makes traveling more rewarding.

51:23Designed to help you get more out of every stay. Sign up at Sonesta.com to enjoy instant savings, bonus points, and valuable perks like early check-in, late checkout, room upgrades, and free stays over time. With Senesta Travel Pass, every stay brings you closer to your next reward. Choose from more than 1 ,100 hotels across 13 distinctive brands and unlock the best available rates when you book direct with Senesta Travel Pass. Here today, roam tomorrow. Join now at senesta.com. Terms and conditions apply. This dog salon? Operational excellence. Thanks to genius from Global Payments. Scheduling?

52:02Personalized. Checkouts? Instant. Absolutely genius. Big League reliability for any business. That's genius.