Broadcom Follows Oracle in Disappointing AI-Focused Investors

12 Dec 2025 · 44 min

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

Podcast Episode Notes: Bloomberg Tech - "Broadcom Follows Oracle in Disappointing AI-Focused Investors"

Episode Summary In this episode of Bloomberg Tech, hosts Caroline Hyde and Ed Ludlow discuss the recent performance of Broadcom and Oracle, the implications of large state-backed incentives for China's chip sector, and insights from White House AI Czar David Sacks regarding President Trump's executive order on AI regulation.

---

Key Topics Discussed

  1. Broadcom's Sales Outlook
  2. Broadcom's stock declined significantly after failing to meet investor expectations for future sales, despite an impressive AI-specific backlog of $73 billion.
  3. CEO Hock Tan refrained from providing an annual revenue forecast, citing uncertainty and labeling the future as a "moving target."
  4. Analysts noted the company's AI sales exceeded expectations, but market reaction was negative due to concerns over margins related to the Anthropic deal.
  1. Market Dynamics
  2. The Nasdaq 100 index falls by 2%, influenced by Broadcom and Oracle's disappointing announcements.
  3. Investors show anxiety about the AI market bubble, linking current trends to broader market performance.
  1. China's Semiconductor Incentives
  2. China is considering the largest state-backed financial support for its chip sector, potentially amounting to $70 billion.
  3. The aim is to bolster domestic AI chip makers, competing directly with U.S. firms like NVIDIA and Oracle.
  4. China faces challenges with tech restrictions which limit its access to advanced manufacturing technologies.
  1. Oracle's Data Center Delays
  2. Oracle reported delays in the completion of data centers for OpenAI, pushing timelines from 2027 to 2028 due to labor and material shortages.
  3. The report prompted significant market reactions, with Oracle's shares declining sharply.
  1. Interview with David Sacks
  2. David Sacks discusses the executive order aimed at unifying AI regulation at the federal level to counteract the confusion created by disparate state laws.
  3. The order seeks a standardized compliance framework to facilitate innovation and protect businesses from excessive regulation.
  4. Sacks addresses concerns regarding job displacement due to AI, citing that studies have shown no significant job loss attributable to AI yet.

---

Key Takeaways

  • Investor Expectations: The market is highly sensitive to forecasts, especially in the AI sector. Broadcom's reluctance to provide clear guidance exacerbated investor concerns.
  • Governmental Support: Significant state-backed funding from China poses competitive challenges for U.S. tech companies in the semiconductor space.
  • Regulatory Landscape: The need for cohesive federal regulation of AI is emphasized, with the goal of preventing a patchwork of state laws that could hinder innovation.
  • Long-Term Outlook: Despite short-term volatility linked to current tech giants, experts highlight the sustained growth potential in the tech sector, especially in semiconductors and AI.

---

Conclusion This episode of Bloomberg Tech presents a nuanced look at the challenges and dynamics in the technology sector, particularly focusing on major players like Broadcom and Oracle, as well as the strategic maneuvers being made in the global semiconductor market. The conversation with David Sacks underscores the administration's effort to streamline AI regulation, reflecting broader concerns about innovation and economic implications in the face of rapid technological advancement.

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

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.

0:01Bloomberg Audio Studios Podcast Radio News.

0:11Bloomberg Tech is live from coast to coast with Caroline Hyde in New York and Ed Ludlow in San Francisco. This is Bloomberg Tech coming up. Broadcom suffers a stock slide after its sales outlook, or lack of, failed to meet investors' lofty expectations. Plus, China eyes the largest ever state-backed chip incentives with as much as$70 billion of state money considered for the pivotal sector. And our conversation with White House AI czar David Sachs on President Trump's executive order aimed at limiting state-level regulation of AI. First, Ed, we check on these markets that are dictated by tech.

0:53Dictated by them. And we say that where we're off by 1.8%, let's look at it, on the Nasdaq 100. We are under pressure as we consider not just the macro perspective and Bon Yield's move, what's happening underneath the surface. I'm looking at the Golden Dragon Index. Look, this is showing what Chinese names trade in the US in terms of the ADRs are up to. What a dichotomy we've got going. What a juxtaposition. China higher. We'll get to that news later. US lower. And in large part, it's by one key stock you're looking at. Some breaking news. Bloomberg reporting, citing sources that Oracle has pushed back some of the completion dates of data centers that it's developing with and for OpenAI.

1:32That pushback is to 2028 from an earlier plan of 2027. The market reacted. Look at the right-hand side of the squiggly line. It's paired some of that decline, but at one point, session low well beyond 6%. I think, Carol, later in the show, we're going to get an opportunity to talk to Brody Ford. Yes. You broke that story. Get more of the details. Right now, the top story is Broadcom. It is on track for its biggest decline since January of this year. a$73 billion backlog, AI-specific. But what the street wanted was a very different number. During a conference call, CEO Hock Tan held off giving an annual AI revenue forecast, saying it was, quote, a moving target.

2:13Listen to this. It's hard for me to pinpoint what 26 is going to look like precisely. So I'd rather not give you guys any guidance. And that's why we don't give you guidance, but we do give it for Q1. Carl Ackerman, Managing Director of Semiconductors and Networking Hardware at BNP Paraba, joins us for more. His new price target,$475, among the highest now for analysts covering the stock. So that's the point. You have different sets of data. They gave us a figure which was a$73 billion backlog. Yep. This morning, quite clearly, the market would like to see a revenue number that's forward-looking.

2:52Yep. Yep, that's true. So you're right. I think what's interesting here, the reason why Broadcom is down is not about how revenue has perhaps missed expectations. First and foremost, the company actually beat and raised guidance. AI sales were above expectations. The outlook for next quarter of$8.2 billion. Farx City consensus of$6.8 billion of AI sales, which doubled on a year-to-year basis. When queried, the company indicated that, Hawk indicated that revenue could perhaps accelerate into fiscal 26 versus that 100 % accelerated growth number. We have over$50 billion of AI sales in fiscal 25.

3:30I think the reason why the stock is down, Ed, however, is in part because there are some investors a bit worried about the margin structure of the Anthropic deal. So of that$73 billion in AI sales, $21 billion will go to Anthropic as part of a TPU system sale. And that TPU system sale is going to be lower margin than what Broadcom generates today. It's about content. How much of the content of a server do you own beyond just the chip? But you like this name, right? Like you've just raised the price target to$475 from$385. Is that right, Math? Well, what is it that you like about Broadcom? Sure.

4:07I mean, Broadcom, I think, is in a league of its own akin to NVIDIA, really, where NVIDIA and Broadcom dominate two of the three main buckets of AI infrastructure spending. Those include compute, memory, and networking. They, of course, control two of them, networking and compute, more than anyone else. And I think what is happening here, what the investment community, I think, is overlooking today on this sell-off is the fact that Broadcom is moving toward a full solution stack akin to what NVIDIA is doing on providing AI compute and AI networking as we move to a scale-up domain where you're going to need optical for both the networking ASICs and the compute.

4:53And they offer that entire IP stack in-house. And that is what I think people are missing today. You're calling it short-sighted, Carl. I know you don't cover NVIDIA as a name, but does Broadcom's success come at the expense of others in the market? Thanks, Caroline. It's a good question. You know, I think it's in a vacuum. You know, we try and pigeonhole Broadcom versus NVIDIA or Broadcom versus AMD or GPUs versus custom accelerators. I think what we found out is that it is not a winner-take-all. It is not a singular approach. What you're seeing is that hyperscalers are adopting custom compute compute as well as GPUs for frontier model training.

5:41And that will continue to progress as we move toward inferencing. So it's not so clear-sighted there. I think what's important here is that Broadcom offers the full solution stack across networking and compute. And similar to NVIDIA, those are about the only two companies in all of tech, including semis, that offer that capability. Of course, Hock Tan's very much aligned in terms of his own pay package with hitting certain revenue numbers for AI in particular. But from your perspective, is there a supply-side headache going on at the moment? We're hearing reports that Dell, for example, is having to jack up pricing because of the pricing that it's feeding on memory.

6:25More broadly, we're seeing the breaking news that Oracle's having to delay some of its data centers. There is a tussle for the infrastructure that's necessary right now. And is that just not going to happen in the overnight way that the market is anticipating? Great question, Caroline. I think what's interesting to hear is I would like to tie in this idea that are we in a bubble of AI infrastructure? I think that's no. The answer is definitive no. I think gigawatt capacity and announcements take time to ramp. That ramp and that qualification and filling up that fab is now extending the visibility across the entire supply chain, whether it's hard drives, whether it's memory, whether it's networking, whether it's compute.

7:11And that is giving companies like Broadcom, companies like NVIDIA, companies like everyone across the supply chain, a very long visibility. And right now, things are very tight. And so some of the areas that are the most tight in our AI ecosystem coverage include lasers for optical transceiver components. There are ways to ameliorate that tightness among the supply chain. And Hawk and Broadcom announced how they have the sufficient capacity for their chips to meet the demand that they see in fiscal 26 and in fiscal 27. Who doesn't love talking lasers? Carl Ackerman of BNP Paribas. It's been great getting your take as we do see stocks fall through the Nasdaq.

7:56More broadly, now under pressure of some 2%. We want to keep up the discussion that's happening between the US and China because chip markets are in the eye of the storm there. China planning to pour up to$70 billion of state money into the sector, being pivotal to its technological conflict with the United States. Bloomberg's Maggie Eastland joins us now. And$70 billion would be an extraordinary amount of government support that we've ever seen worldwide, right? Yes, this is a huge number. This would be China's biggest semiconductor-specific packaging. And according to Bloomberg reporting, that number could be anywhere between, as you said,$70 billion, and it could go down to$28 billion.

8:36So it depends here. We'll see how large it is. That compares to the U.S. effort at industrial policy for semiconductors, which is$52 billion. So this is a similar scale to what the U.S. has undertaken in recent years. Maggie, there are some questions to which we just don't have answers, and that largely relates to what China specifically will do with the funds and where they'll go. But within that Bloomberg reporting, where do we think China is going to prioritize using the funding? Which champions is it going to elevate? Yes, so of course, China is always really looking to support its AI chip makers that compete with NVIDIA.

9:13So that includes companies like Huawei and CanberraCon. Con, but there are other chip companies, including SMIC or SMIC, which does the manufacturing. So, that would be the corollary to TSMC. So, there's a range of companies that could see this investment, but China will face headwinds because it still has restrictions on many of the equipment technologies and the foundries that can't access TSMC for those advanced node chips. You have many worrying about just what the yields are like from an SMIC, from a SMIC, when they're actually producing these homegrown domestically made chips. But I'm looking at NVIDIA under pressure again today, Maggie.

9:51More broadly, are we getting any sense whether this is meaning that China will say no thank you to the H200s in the same way that they did the H20s? As you know, China has certainly never been shy about saying no to U.S. technology when they have their own local champions. But what I will say is we don't know the exact number of H200s China will accept, if any. What we do know is they're clearly not backing down on their commitment to their own chip supply chain, and they're not going to readily give way to this U.S. strategy of selling advanced AI chips in order to undermine their local competitors.

10:25So we'll have to wait and see what happens on H200s. Bloomberg's Maggie Eastland, who's been across China's chip efforts all week long. Thank you very much. coming up. We're going to get more on that breaking news report from Bloomberg about Oracle delaying OpenAI-specific data center project by one year, back to 2028. It had an impact on the markets. The Nasdaq 100 is now 2%. Actually, Caro, as of Thursday night, the state of play was the Nasdaq 100 was flat for the week. So we are now down 2 % on the Nasdaq 100 on a weekly basis. Of course, the reaction to Broadcom and some other AI angst is part of what's weighing on this market.

11:04And we're going to get a lot more on that very, very soon. Stay with us. This is Bloomberg Tech.

11:21Okay, shares of Oracle are down almost 5%. They've hit session lows of 6.5 % decline after Bloomberg's Brody Ford broke the story that the firm will be delaying data centers for OpenAI to So 2028, instead of a previously planned 2027, the reason labor and material shortages, Bloomboast Brody Ford runs the set and joins us now. Important reporting for you, because this is what a part of Oracle's debate is. Okay, great. You have a great backlog of business, but we are very closely monitoring the ability to execute on it and then book revenue on it. And OpenAI is a big chunk of the exposure. Fill the blanks for us.

12:01More reporting. What other details have you got? These are unprecedented scale data centers, right? I mean, GigaWatt scale data centers unprecedented and Oracle's trying to do effectively five of them at once. And so what we have today is that the initial completion dates, the initial full delivery dates have been pushed back in some cases from 27 to 28. And, you know, I think if you've spoken with data center folks over the last couple months, it's not a huge shocker because, I mean, these are such crazy projects, right? And the idea of getting them done in two years was always going to be ambitious.

12:35So now getting them done by 28, it's still a tight timeline. It's still, frankly, an impressive turnaround, but it's maybe just not as quick as the company had initially hoped. We're looking at Abilene, Texas right now, where on the conference call during earnings, we had the co-CEO, one of them, saying that they had more than 96 ,000 NVIDIA chips delivered, giving you the sense of scale. But is it a hindrance to what the revenue is ultimately for Oracle here? Or is it just investors having to be like, oh, it's still jammed tomorrow, not today? Right. I mean, Oracle, I'm sure, will say that as you put, they deliver these sites in chunks, right?

13:15And so Abilene is already being turned on, this massive data center in West Texas. Once you get the servers running, OpenAI uses them. That's revenue recognition, right? And so really what we're thinking about is the further out sites beyond Abilene. We keep seeing these Stargate announcements for Michigan and New Mexico. These sites which are still being kind of put together and Oracle and other vendors too are finding that, wow, there's a lot of stuff out there that's backlogged. Labor for one of them. There's only so many electricians. I mean, you want to build in rural Texas. It's a smaller pool of people you have access to.

13:50It's a fascinating story. It's going to run and run. Bloomberg's Brody Ford with a real market moving bit of reporting there. Let's talk more about the market implications. We've got Margie Patel with us, Senior Portfolio Manager and Head of Capital Allocation over at Allspring Global Investments. She has$629 billion in assets under advisement. And Margie, are you worried these drip, drip bits of information that maybe the revenue streams aren't able to be booked tomorrow? It has to be weighted out a little bit more in terms of the returns on AI investment? No, I think the long-term trends are still in place.

14:24I actually thought that Broadcom's numbers were quite good and people were just, I think, very nervous at the end of the year, particularly with some bad news we've seen, such as from Oracle. And I think it's really more just end-of-year jitters rather than anything fundamental. I think when you look out into 2026, you still have to like the tech sector, especially the Semi, the Memory, those companies, I think you still have to stick with them, that they're going to continue to be high growers. And these little hiccups we have here and there don't change the fundamental trend of very, very strong growth in a year which should be very modest growth next year.

15:01So we still like the whole sector. So, Margie, on a day where Broadcom's off by 11 percent, when the Nasdaq's off by 2 percent, is that hiccup a buying opportunity? Well, I think it is, actually, because if you look at the leading stocks that have stuck to their plan, that have great growth, great profit margins, great innovation, and they're down 10%, 15%, even 20 % from their peaks of a few months ago. So I think that looks like a pretty attractive time to add to these names because there's end-of-year uncertainty. A lot of short-term traders want to preserve their gains, cash out. And I think that's what you're seeing is this pressure on the sector rather than the change in the fundamentals.

15:40The cash flow of the big companies is so large. This isn't going to be derailed anytime soon. So I think next year looks pretty good sailing too. Margie, it's good to see you. It sounds like at your end, you don't personally have many jitters. One of the best read stories on the Bloomberg Journal today is about the debt and the lending that is behind the build out in that infrastructure. where does the debt does debt as a factor in consideration sit for you when you are tracking all sorts of different hard and soft data sets to work out what's going on here? Well, I look at the debt as really a company's choice of how they want to allocate capital.

16:22Do they want to borrow? Do they want to increase the dividend? Do they want to do share buybacks? And particularly the large successful companies really have no need to borrow, even paying for the CapEx, they still have plenty of excess cash flow that they have to decide how to utilize. Whereas I think Oracle has really taken on a lot of debt compared to their cash flow. But the rest, I think, all look pretty good. It's a volatile sector, and that's why the returns are higher, because you have to be prepared for these little downdrafts in order to get the upside. Maggie, Brody's report on Oracle was very specific, that Oracle's delaying those projects by a year because of labor and material shortages.

17:05Elsewhere in the U.S. economy, we have all the chips we need, clearly. Are we good at the other stuff, and is the other stuff where it needs to be to support this build-out? Yes, I think it is. I think when you have this explosion in demand of these very complex centers, I think you should expect there will be short-term problems of supply and so forth. But really, that's a good problem to have rather than lack of demand or pricing pressure. And we're really seeing very strong pricing for the semis that are going into the data centers. So we think it's just nothing to worry about. And the fundamentals are still very strong.

17:43Briefly, though, there's reports today from other outlets saying that Dell is going to have to jack up its prices because of pricing strength in memory. Are there areas of this AI trade that are overvalued that you shouldn't be piling into from a margin perspective? Well, I think when you look at tech, it's like any other sector, is the best companies usually trade rather richly. The cheap companies usually have problems. They don't have the leading edge. They don't have the innovation. And when you look at tech, it's just, and you can see even here with the data centers and concern about what approach that companies are using, you know, Broadcom or NVIDIA, whatever, in the new products.

18:19It's really about innovation and who are the leading innovators. And so the companies that don't have that innovation are just going to fall behind. So I think this year up till, say, the summer, everything moved up. And now we're seeing a separation between the companies that have leading edge and the companies that are really falling behind and aren't going to catch up. So I think it'll be much more stock selection next year than we've had for the first part of this year. With a message that in tech there's nothing to worry about, Margie Patel from Allspring Global Investments. Great to have you back on the show.

18:50Thank you very much. Now, coming up, we're going to bring you Bloomberg's exclusive conversation with Uber's CEO on the company's international ambitions, particular focus on Asia. That's next. This is Bloomberg Tech.

19:07Uber CEO Dara Coswishahi says the company expects to offer robo-taxi services in more than 10 markets by the end of next year. He spoke with Bloomberg Tech Asia's Annabelle Droulers about why he's optimistic in particular about growth potential in Asia. The APAC market and in particular the North Asia markets as well, they are huge growth markets for us. And if you look, for example, in the rideshare business, over 30 percent of our global first trips coming into the category come from the APAC region. the area is growing very quickly including taxi as well which is actually one of our newest products on the platform so for me coming here seeing the teams meeting with local business people and regulators and talking about how we can be part of the future growth of the region is really what my agenda is early this year you also put out a statement on on the robo taxi push as well.

20:06And so the Middle East and Asia were the markets for 2025 to launch. And we've seen, of course, that initial deployment in the Middle East already. What's happening on the Asia side? Well, lots of discussions on the Asia side. I think what's really important is to set up a regulatory framework to go forward. For example, Hong Kong has various trials and pilots going on and in many other markets. We're talking to regulators about how we can be a part of shaping rideshare and autonomous rideshare going forward. The technology is absolutely getting there. These are the robot driver. It doesn't get tired, doesn't get distracted.

20:46And we very much look forward to working with various authorities to introduce rideshare into the markets. We're now live in four markets now as we speak in the U.S. and in the Middle East. And I expect to be in 10 plus markets by next year. And we want those markets to be in the Asia Pacific region as well. Where then in Asia do you think is the most likely place? We'll see. I think that certainly Japan has great potential. They are behind on their regulation. They are behind on their regulation. But I think that they also understand that with an aging population, there's a real need for transportation.

21:29not just in the large cities, but in the rural areas. And for example, I experienced that personally going to Kaga City, where we have communal rideshare and kind of took a rideshare trip and understood what the needs are there. So we're talking with various countries, regulatory authorities. I think Japan is going to be a part of it. I certainly hope that Hong Kong is going to be a part of it. Australia, where we were just talking about, is a huge market for us. So we're having those dialogues. And I think that the picture will shape up over the next two years because the technology is definitely getting there.

22:03Uber CEO Dara Khosrowshahi there along with our Annabelle Drowlers. Now sticking with Uber, the company along with DoorDash is suing New York City to block requirements that the delivery tipping option be available at the time of checkout and set to at least 10%. The two companies argue this would worsen sticker shock for inflation-weary consumers. New tipping laws are set to become effective January 26th. Caro? Want to watch, Ed? Meanwhile, coming up, you're reporting to watch how Rivian is replacing NVIDIA's tech in future vehicles with its own chips. From New York, from San Francisco, this is Bloomberg Tech.

23:00Welcome back to Bloomberg Tech. Rivian. Take a look at its shares, up currently 16%. At one point in the session, up more than 19%, trading at its highest level since January of 2024. So its highest level in two years. Yesterday, it fell quite a lot after it told investors and the world its plans for autonomous driving. Rivian's plan for autonomous driving is based on two big technology bets, and we went to see them.

23:33Rivian's developed its own artificial intelligence chip for its future cars. That gamble might pave the way to fully autonomous driving.

23:45A lot of the tech world imagines a future where we don't own cars at all. We're talking about fleets of robo taxis that are summoned through an app, maybe no steering wheel or driver controls at all. But Rivian's in the camp that does see people owning their own cars in the future and being willing to pay top dollar for a software platform that allows the car to drive itself. If this idea seems familiar, Tesla has been selling a version of it for years. Full self-driving supervised is not technically full autonomy if you read the fine print, but it can get you from point A to point B without needing to put your hands on the steering wheel.

24:25Rivian's path to autonomy is rooted in two big bets. The first, a custom AI chip developed in-house, which marks a big break from NVIDIA. This is a Wrap 1 chip. It's a multi-chip module. In the middle is a Rivian-designed custom silicon surrounded by memory on two sides. The decision to build an in-house was based on a very rigorous analysis of the benefits we could come. And those benefits are velocity or ability to get to market very quickly with it, performance and cost. The second, a major change in how Rivian vehicles see the world. The next generation Rivian R2 will have indented LiDAR sensors in it.

Read the full transcript

25:05It's not just about the compute, it's also about the sensors, and it's about how they all come together. This is Ruby on Mark 1. What the Gen 3 architecture does with compute levels that are dramatically expanded, so just to put some numbers to this, at the platform level, 1 ,600 sparse tops. You can process 5 billion pixels per second, beautifully integrated LiDAR. That raises the ceiling to allow us to take your eyes off the road. This will cause some debate. Tesla vehicles only use cameras as sensors for their systems. And what Elon Musk's company has always argued is that other sensors like LiDAR or radar are too expensive to scale.

25:48Until now, Rivian didn't really have an autonomous system. It had advanced driver assistance tools powered by cameras and radar. And Rivian used NVIDIA chips as the brain in the vehicle to interpret the world around it. Rivian says its new AI model will keep improving those older cars too, eventually adding capabilities like hands-off point-to-point driving. Rivian doing its own chip and ditching NVIDIA is a surprise. On the Rivian processor side, this represents a significant cost savings to us. There is a lot of margin, of course, in the semiconductor space and working directly with TSMC, we have a great relationship with them.

26:25The new hardware is the unlock. It should allow Rivian to go from that driver assistance software in its existing lineup to true autonomy in the next gen R2. The next big step is personal level four. And what I mean is the vehicle can operate empty. It can operate without anyone in the driver's seat. It can pick your kids up from school. It can drop you at the airport. It's a complete shift in how we think about the vehicle experience. Looking back, this was the company that pulled off the sixth largest IPO in US history, and it was first to market with full-size battery electric pickups and SUVs, beating out Tesla, Ford and General Motors.

27:03But today Rivian's struggling with the basics. Production of its EVs hasn't really scaled. Rivian's sole plant in Illinois is capable of building 250 ,000 units a year. But in 2025, it probably won't hit 50 ,000. The truest representation of that struggle, the stock is at a fraction of its peak. Right now in the world of tech, you have to have something to say about AI. and Rivian is diving deep into autonomy to appease its investors. An AI chip, LiDAR and a large driving model. Right now, it's a promise from Rivian that their next generation vehicles will have genuine autonomous capabilities.

27:48Extraordinary reporting, deep dive and Rivian, as you said, Ed, doing very well on the day, unlike the rest of the markets. Just checking what's happening to them in NASDAQ. We're down on the week, we're down on the day. NASDAQ 100 off by more than two percentage points. Tech in the line of fire today. Even dragging Chinese names now into the red. They started the trade and our show in the green as we understood that the Chinese government was going to be going all in on funding its own domestic chip supply, up to$70 billion worth of government incentives is the reporting coming out of Bloomberg News at the moment.

28:20But even China starts to dip at the moment. We see Broadcom in the red, NVIDIA in the red, Palantir, Amazon, Micron, some key names currently on the downside as we have that AI bubble anxiety all over again. Right, coming up. Actually, Caro, we're going to talk about the United Kingdom. We're going to be joined by British Business Bank CEO Lewis Taylor for his take on the UK's tech sector. He's in town in San Francisco in Silicon Valley to think about how technology might work across the Atlantic. That's next. This is Bloomberg Tech.

28:59The UK economy is at risk of its first costly contraction since Labour returned to power after growth, disappointed again by shrinking ahead of Chancellor of the Exchequer Rachel Reeves' tax-raising budget. Can the UK's tech sector come to the rescue? Louis Taylor, CEO of the British Business Bank, joins us now. It's great to have you in town in San Francisco. Really great to be here, Ed. Thank you very much indeed. You have an annual budget, essentially, to invest in and lend to and support the technology industry in the United Kingdom. Why are you in San Francisco, then? What brings you here?

29:39Why are we in San Francisco? Well, look, we're here pitching to USVCs a really great new growth opportunity for them, which is based on three things. Firstly, as you say, the quality of the UK tech industry. Four of the top ten universities globally producing great research. with some really excellent entrepreneurs and the ability to scale businesses as well, although that's where they come in. Secondly, a new pool of capital coming on stream domestically in the UK from pension funds, hopefully. And then thirdly, the opportunity to partner with the bank, which is the biggest LP in UK venture and growth equity, connected and knowing the landscape pretty well.

30:14Louis, the landscape we know so well is fintech. I think of Revolut, I think of Monzo, the standouts. But where else is really thriving? Where else should VC come in as port in the UK? Well, look, I think you're absolutely right. FinTech is very strong. And just today, GoCardless did a deal with Molly and that unicorn that we invested in 12 years ago. So you need a bit of patience on this. So FinTech is very strong. AI in different places is very strong, not so much on the hardware side, not so much on the large language models. But more broadly, if I think about companies like Synthesia or Eleven Labs, all of those companies coming out of the UK.

30:52So I think those tech areas are great. But also the application of AI into life sciences is incredibly strong as well. And the UK has an incredibly strong life sciences industry. So I think we see AI as being a theme across the sectors of the industrial strategy the government announced and the UK being strong in those. To be fair, many of those companies you named, they come on this program regularly, and they are making advancements in their respective fields. I want to go back to what you said, a big pot of pool of money coming online. A little bit more, please. How big, how certain is it? And that's important, right?

31:30Because the lesson of AI in this country, at least, is the capital requirements are much bigger. Much bigger. So the UK, we incubate companies incredibly well. we scale them less well and we haven't had the scale-up capital we need. It's not that we don't have the money because we have the second largest funded pension scheme in the world at around four trillion pounds but we have an allocation issue with that pension money which is changing. The government encouraging pension funds to invest more in the domestic economy and in the growth economy but I think what we're looking for is some of the expertise here in the VC industry in the US about scaling those companies.

32:05As I say we incubate well but it's that scaling stage and the expertise needed there, where, of course, you need capital, but you need other things as well, networks, capabilities, and many other things, mentoring of leadership teams. Louis, what is the state of brain drain when you actually do get a really successful company? Now, 11 labs are staying there since these years as well, but many up and leave and come to Silicon Valley or to the US. Is that something that's still happening? Well, look, I think it does happen to an extent. I think we really want to try and address that and try and stop it and actually capture some more of the value in the UK economy.

32:39As I say, the companies have largely come over here because it's where the capital is. If we've got a new capital stream with some expertise based on it in the UK, we're going to hope to retain those companies in the UK longer. And actually, the innovation ecosystem, I think, is quite self-perpetuating. Success breeds success. We've got a lot of the right things in place, but it's just this top end that we need to really make sure that we realize more potential and keep the flywheel going. So your strategy here, and one of the reasons you're in town is to go to the American VCs, the Valley VCs, and say, give us your capital, come to the UK, we have something to offer.

33:13But that strategy then carries risk with what you were just talking to Caroline about, right? Yeah, we're not quite saying that. We're saying, bring us your expertise. We'll help you raise capital locally. This is a growth opportunity for you. Raise capital locally. That's a key point of distinction. Absolutely. And the connections that we have with all the GPs, LPs, but also the pension funds. I mean, we actually are raising our own fund at the moment for a co-invest fund in the UK from UK pension funds. And we'll do the first close end of January, early February. And this is a thing. It's a real thing.

33:44It's a real thing. And we'll let you go out and have real conversations with those VCs in the Valley. Louis Taylor, thanks for stopping by. CEO of the British Business Bank. Coming up, White House AI and crypto czar. David Sachs joins us to talk about the President Trump executive order on AI regulation. This is Roombag Tech.

34:15Welcome to our global radio and TV audiences. President Trump signed an executive order aimed at limiting state-level regulation of AI. The move is supported by tech leaders who have argued local rules could stifle innovation. We're joined by David Sachs, the White House AI and Crypto Czar, or Senior Advisor. David, I think it's a really good place to start. In your work with the president in consulting and advising on the formulation of this executive order, what was the problem that you were trying to solve for? And what is it that you said to the president about why this EO was the right approach to focus on state-level laws?

34:58Well, Ed, thanks for having me. The problem that we see is that you've got a thousand different bills going through state legislatures right now to regulate AI and over 100 measures have already passed. Some of these bills are contradictory and you've got 50 different states running in 50 different directions. That type of compliance regime is going to be very hard for small companies and startups, especially innovators, to comply with. And so what we need is a single federal or national framework for AI regulation. And that's what the president has supported. And by the way, he supported this for a long time.

35:31If you go back to his July speech on AI, he called for a single national framework then. And what we've done with this EO now is to make clear that that is the administration's policy and to task members of the administration to work with Congress to try and enact that framework through legislation, because ultimately this needs to be a law, and in the meantime create tools that the administration can use to push back on examples of the most onerous and excessive state regulations. David, there is, of course, some pushback on the executive order from the states themselves, from other Republicans.

36:06And, you know, as you know, like I studied the July speech and strategy closely. A big part of it, you know, was infrastructure related and about deregulation. The concern about this latest executive order is that while it addresses your concerns about many different pieces of state regulation, it does not provide for a single federal framework. well at the end of the day that single federal framework has to be enacted through law and we need congress to do that and so the president has asked congress to do that and he's tasked members of the administration to work with congress to produce that framework in the meantime what we've done here is articulate a set of principles we've said what values are important to us we said that we want to protect child safety that's important we want to respect copyright we want to preserve the ability of local communities to choose what infrastructure is in their communities.

37:03We're not seeking to preempt the states in any of those areas. So this is an important set of principles that we have put forth. And at the same time, the EO provides for a number of tools that can be used to push back on excessive state regulation. And just, Ed, let me just illustrate why I think this is so necessary. Let's just, what we're really talking about here is regulation of AI models and algorithms. Well, think about how an AI model is developed. You can have developers in one state or multiple states writing the code. It can then be trained in a data center in another state. You then can have inference happen in another state.

37:41And the entire service is provided over the Internet using national telecommunications infrastructure. So you're dealing there with at least four different states. And all of them can lay claim to regulating those AI models. And those regulations can be in contradiction with each other. Even Democrat governors have admitted this is a problem. So just the other day, Kathy Hochul, the governor of New York, basically said that she might prefer to enact California's SB 53, which is a regulation that they just passed in California, rather than the bill that her own assembly gave her, the Rays Act. Because she sees that, wait, do we really want to create this patchwork of different regulations?

38:17So even Democrat governors are realizing this is a problem. And if they all run in different directions, then we're going to end up with a patchwork or a mishmash of regulations that are impossible for companies to comply with. What the president is calling for here is just common sense. We want to get to a single national framework of compliance as opposed to 50 states running in different directions. Meanwhile, Cathy Hochul actually is getting a bit of criticism, perhaps, for narrowing. And what some are saying is bowing down to business. David, I'm really interested in how you oppose that view because there is anxiety in the population, AI versus jobs, AI versus energy bills.

38:55How are you giving them the sense that we haven't seen federal government and indeed now state governments just handing over the reins to big tech billionaires, as people call them? Right. No, I understand there's a lot of fear out there about AI and job loss specifically. And a lot of those fears have been drummed up. Let me just say on the job loss question, because I think this is really important, that Yale just released a study, and it showed that in the 33 months after the launch of ChatGPT, there was no discernible disruption to the U.S. job market. None. They said no discernible disruption.

39:27And in fact, if you look right now, more jobs are being created than being lost. So this whole idea of job losses just isn't true. There was an article in the Wall Street Journal just last week talking about the construction boom that's happening that's benefiting construction workers like electricians, like plumbers, like workers who pour concrete or hang drywall. Their wages are up 30 percent because this infrastructure boom that's happening right now. And there's actually a job shortage in many of those trades, meaning we need more workers going into those trades. So what we're seeing right now is an overall AI boom that's benefiting the economy.

39:59You know, the GDP growth rate was tracking about 4 percent. And half of that, up to half of it, has been attributed to AI. So I just think that this narrative about job loss has been blown out of proportion. Certainly there could be job displacement in the future, but we haven't seen any of that so far. It's been quite the opposite. It's been job gains. David, final one on the EEO, if I may. You know, what this EEO allows for, is it the sort of hope that it will lead to the DOJ suing states like New York and California? And if that's the case, you know, the president and the administration's confidence that you'd win them.

40:40Well, that is one of the tools that is in the EO, is that the DOJ has been tasked to form a litigation task force that would have the ability to push back on excessively burdensome state laws, laws that may be unconstitutional, violate the First Amendment, things like that. By the way, the DOJ already had that power. So this is not a novel power. But what's being done here in the CEO is we're marshalling all the resources of the federal government behind the strategy of the president to create a national framework. Now, in terms of what laws we go after, that's a decision that has been made. We haven't decided whether California or New York should be targets in that way.

41:17The one that I think is probably the most excessive is this Colorado law that seeks to prohibit algorithmic discrimination. What that basically says is if an AI model has a disparate impact on a protected group, then that model is violating the law. Model developers, by the way, have no idea how to comply with this because they're not aware of all the downstream uses of their model. I mean, if a business decides to use an AI model in a hiring decision, for example, that business is already on the hook for discrimination. So how would the model developer know that it was being used in that way?

41:49But what Colorado is trying to do there is get their ideology inserted into the model. That's very concerning to us. We think there's a First Amendment issue there. But look, we haven't made any decisions in terms of how that litigation task force will be used. David, briefly, all of this is set in the context of U.S. versus China and are deemed to run forward on AI development. Meanwhile, it's been a busy week and H-200s might indeed be able to get to China. How many do you think you'll do in volumes and what do you think the appetite is of China to buy NVIDIA's more sophisticated chips? Well, it's interesting.

42:21I just saw an article that said that China was rejecting the H-200s, so apparently they don't want them. And I think the reason for that is they want semiconductor independence. The same way that the United States wanted to be energy independent, they want to be semiconductor independent. So they're rejecting our chips, and that's part of the calculation that goes into the decision of what we authorize to be sold to China. The U.S. policy has always been that we don't allow the leading-edge chips, and we're not. This H-200 chip was state-of-the-art a couple of years ago, but now it's been superseded by the newer Blackwell architecture and the Rubin architecture that's coming out next year.

42:56So this is now a lagging chip, not a leading chip. But what you see is China is not taking them because they want to prop up and subsidize Huawei. They want to create a national champion. And that was part of our calculation of selling not the best, but lagging chips to China's. You can take market share away from Huawei. But I think the Chinese government has figured that out. And that's why they're not allowing them. David Sachs, we always wish we had more time. White House, AI and CryptoZar, we thank you for joining us today on the Executive Order and indeed on NVIDIA's H200s. That does it for this edition of Blue Bag Tech.

43:27The market is in sell-off mode, Ed, as we wrap up this week. We're down by more than 2 % on the Nasdaq more broadly and indeed for the week. But really, all eyes on, well, Broadcom and its numbers. Yeah, Broadcom in the earnings context, investors wanted more. But Bloomberg reporting on Oracle is what moved the needle. Believe it or not, this is my last show of 2025, an astonishing year. And a lot of the themes in today's show are what we've been talking about all year long. Recap on the podcast. You can find it on the terminal and online. You know where to. Caro, I'll see you in 2026. Have a great break.

From the publisher

Bloomberg’s Ed Ludlow and Caroline Hyde break down Broadcom's plunge after the company's sales outlook failed to meet investors' lofty expectations. Plus, China eyes the largest ever state-backed incentives for its pivotal chip sector, and White House AI Czar David Sacks joins to weigh in on President Trump's executive order aimed at limiting state-level regulation of AI.

See omnystudio.com/listener for privacy information.

More from Bloomberg Tech

All 343 episodes
Broadcom Follows Oracle in Disappointing AI-Focused InvestorsBloomberg Tech · 44 min
Listen in VO