Digesting DeepSeek's AI Impact
42:59 Share

89% of business leaders say AI is a top priority, according to research by Boston Consulting Group. The right choice is crucial, which is why teams at Fortune 500 companies use Grammarly. With top-tier security credentials and 15 years of experience in responsible AI, Grammarly is how companies like yours increase productivity while keeping data protected and private.

See why 70,000 teams trust Grammarly at grammarly.com slash enterprise.

Open Source AI is available to all, not just the few. Meta's Open Source is free to use, enabling startups like Nanome to innovate. Here's CEO Steve McCloskey. With Meta's free Open Source AI model, Lama, we built an AI tool that allows us to collaborate with other scientists to discover treatments for diseases. Learn how others are building with Meta's free Open Source AI at ai.meta.com.com.

From the heart of where innovation, money, and power collide, in Silicon Valley and beyond, this is Bloomberg Technology with Caroline Hyde and Ed Ludlow. ♪♪

Live from New York, I'm Caroline Hyde. And I'm Mike Shepard in San Francisco. This is Bloomberg Technology. Coming up, Deep Seek Day 2. We dive into the ripple effects in tech markets and the impact on U.S. chip curbs. Plus, we discuss Deep Seek's AI model and the future of open source with Hugging Face's chief AI scientist. And President Trump says Microsoft is in talks to acquire the U.S. unit of TikTok.

Details later this hour, but first we check in on the markets and we claw our way back. Mike, not nearly offsetting yesterday's significant sell-off. We're up just a percentage point on the overall Nasdaq. Look, Apple once again doing heavy lifting to the point side. Meta at a new record high, but not enough because Nvidia bounces back but 2% after a 17% sell-off yesterday, but $30 billion being added from the market cap that was wiped out by $600 billion yesterday. This is all

all surrounding China coming with clearly a very powerful, very efficient, very cheap open source model, Mike. And yesterday, NVIDIA called DeepSeek's new model, quote, an excellent AI advancement that complies with U.S. technology export controls and that DeepSeek's work illustrates how new models can be created. Let's bring in Bloomberg's Ian King now to tell us more about this. Ian, this was a rough day yesterday for NVIDIA.

It bore the brunt of the deep sea shock across the markets, yet their statement really was curious. Tell us what your interpretation was. Yeah, I mean, there are a couple of things going on here. I mean, they could have obviously said nothing. They could have just ignored it. They could have said maybe it's kind of dodgy technology that we don't really believe in. But they actually endorsed it and said, no, this is good.

And the sort of underpinning message we saw here was that maybe they're telling the U.S. government, your export controls haven't worked. We should be allowed to ship whatever chips we want to China because guess what? They'll do good stuff with whatever they can get hold of. Let's just go to that narrative on chip exports potentially not working. We heard from President Trump just yesterday about what the impact could be. Just take a listen in.

In the very near future, we're going to be placing tariffs on foreign production of computer chips, semiconductors and pharmaceuticals to return production of these essential goods to the United States of America. They left us and they went to Taiwan, which is about 98 percent of the chip business, by the way. And we want them to come back.

Of course, a lot of NVIDIA chips are manufactured in Taiwan, Ian, so there's going to be an impact there. What do you think the regulatory impact will be, let alone whether inferencing is really going to be the future for GPU demand?

Yeah, I mean, this is the big question here. The expectation, I think, before the inauguration was that this administration would, if anything, double down on this kind of crackdown of chips and the flow to China. And here we're getting a different aspect of it. So I think we'll have to see how much attention the new administration is placing on this issue. But the expectation is that they're not going to let up.

Ian, back to NVIDIA. How much is this complicating their efforts to get their customers to keep buying all those premium chips like Blackwell and the next generations that you've been writing about so much? Yeah, well, it creates a massive question. It's like, well, if a Chinese startup can fine-tune their models to this extent using sort of two-, three-year-old technology on the cheap, why can't you, right? So if you're Microsoft, if you're AWS and you're running all of these systems, then

Why wouldn't you do it cheaper if you can? So obviously that has implications for what NVIDIA is trying to sell. So we'll see how they answer those questions. Thank you. So far, the statement is all we get. Ian King, we thank you. Now let's just talk about how the tech market is, of course, roiled by China's seemingly cheap, powerful AI model. Here's what some Bloomberg television guests had to say about DeepSeek.

I don't think it's about China versus US. It's really about closed source versus open source. And more and more American developers should leverage the open source and what DeepSeek had done and build on top of it and make it greater. It's really amazing what they have done. Fraction of training costs, fraction of interference costs to produce a model like that in DeepSeek.

but it will unravel basically the path of spending. Of course, I think a healthy dose of skepticism is very, very good. It's needed. But even if you...

Multiply their spend and their cost by 10 times, 20 times. It's still an order of magnitude in terms of the cost reduction. They're not really a company that's out there trying to compete directly with the big U.S. players. So in that sense, I think it was an overreaction to the release of this model.

Let's get more reaction. Martin Orton now, Chief Investment Strategist at Empower. The question for many is, will capital expenditure remain for NVIDIA chips, for AI hardware writ large? Is that what you were tackling yesterday?

Well, I think that is one of the critical questions, right? Because this was a very surgical sell-off, a very surgical development in the sense that it's attacking really not the case for using AI, but really the supply chain around AI and all roads on supply chain lean back

to Nvidia. And so this question of whether companies really need to spend this much, I think that is the question that markets are grappling with. But it's hard to imagine that we're going to see, and of course we'll see in earnings this week, but it's hard to imagine that we're going to see companies massively change their spending patterns simply because of this news. I think

There's still this arms race going on. So, of course, that's something that we're going to have to watch especially closely. But we may need more than just this news to unravel that. But the unraveling could start when you have earnings expectations as they are and valuations at 41 times future earnings for a name like Nvidia. How much are you worried that this could have a similarity to what we saw in the dot-com era?

I mean, that's really something that we're watching really closely. When you're looking at earnings expectations for the broad market overall, you're seeing an acceleration of earnings. And, of course, maybe not the same for MAG7, but certainly a continuation of the powerful earnings that we've seen there. And then you couple that with valuations that are extreme. When we do our analysis of valuations and we break the time series for a price to expected earnings down into deciles, we're looking at the tech sector that's in the 90th percentile.

evaluations. Now, associating that with forward one-year returns, you can still see valuations grind higher and companies perform. But ultimately, that's where you also start to see some of these more meaningful sell-offs. And as pedantic as it sounds, I think a lot of these developments over the past few days relate more to a valuation argument than anything that relates to technology in the sense that

This is big news in technology, but the bigger news is that there's no room in the narrative to account for anything but world domination when it comes to Mag7. And this is allowing us to think about the range of outcomes.

Martin, to play the movie forward a few years, where do you see this taking us? Is this a parallel to the dotcom era, which, you know, eventually after, you know, some troubles right around 2000 really set us up for a long period of growth? What about AI? Are we in that moment?

Well, that's the corollary that I referenced. I know a lot of us are pointing back to the dot-com era and looking at the valuation argument. I think that's important. But I think the other important element is that this is an innovation super cycle, it feels like, just the way the dot-com era was. And what we saw in that period was a massive increase

expansion or growth for the broad market, not just for the tech area. So if we're looking at the trend growth from the 1990s and comparing that to the realized trend growth of the subsequent years, we saw massive expansion, whether that's margins or enterprise value. The internet world had a huge impact on the broad economy. And I think that's what we can expect to see with AI. And in many ways, this news around deep seek is part of that.

in the sense that it says we can economize, we can do this a lot better, and we can get deployment a lot more quickly than maybe we had otherwise thought. And on that idea of getting cheaper AI out to companies and businesses across the economy, maybe a little bit faster, what sort of impact do you see and where do you see growth potential from that?

Well, I do think there's this broad market where you have to understand that we can't perfectly anticipate. I think we're seeing a lot, hearing a lot more conversations around retailers and the deployment of kind of agency, capital.

or find products more quickly with fewer steps along the way. I think that's one of the applications. We're also seeing software stocks hold up quite a bit better. I think the idea that software can easily deploy or maybe more easily deploy AI within their world, I think that's a possibility. So I think that's...

Applications are somewhat endless. The imagination has to be pretty broad here. And so there's some argument for kind of an equal weight exposure to try to capture that range of experiences. But with a long-term horizon, I don't think this is necessarily something that we're going to see in Q2, Q3 of 2025.

We're taking a few tech hits to your line, but we're going to stick with it, Marta, because I have this crucial point for the here and now with retail actually buying back into NVIDIA. Institutional is not right now. What about a so-called black swan event or ripple effect? We talked, of course, to Nassim Taleb yesterday. Just have a listen to what he said. Right.

A little chip in the glass of the valuation question, a little chip in the glass as to whether this is akin to a dot-com crisis. But where could be sustained? Will energy survive this? We saw such significant sell-off to certain of those names.

You know, I think when we're taking a look at NVIDIA or energy, I think there is more vulnerability there simply because this is getting at the heart of the thesis there, that they are the, you know, especially NVIDIA, the dominant provider of this capability that all roads lead back to NVIDIA when it comes to supply.

And this story is really tackling that idea that potentially there's more ways to skin a cat. You can focus on efficiency rather than sheer run first. So I think the idea of a chip in the glass, I think that's important and it's getting at the heart of NVIDIA. But whether this is the beginning of the unraveling of all things AI, I think that's a little bit more extreme or a bridge too far today simply because

That unraveling would depend on the use case of AI being attacked. And I don't think that's what we're seeing here. It's more of the supply chain question. And really, at this point, I think there's still more that we would need to see to really suggest that companies are going to pull back on their capital expenditures in any meaningful way.

Martin Norton, Chief Investment Strategist at Empower. Thank you. Coming up, we're going to discuss DeepSeek's open source approach and what it means for competitors. The Chief Science Officer at AI platform Hugging Face will join us on that. This is Bloomberg.

89% of business leaders say AI is a top priority, according to research by Boston Consulting Group. But with AI tools popping up everywhere, how do you separate the helpful from the hype? The right choice is crucial, which is why teams at Fortune 500 companies use Grammarly.

With over 15 years of experience building responsible, secure AI, Grammarly isn't just another AI communication assistant. It's how companies like yours increase productivity while keeping data protected and private.

Designed to fit the needs of business, Grammarly is backed by a user-first privacy policy and industry-leading security credentials. This means you won't have to worry about the safety of your company information. Grammarly also emphasizes responsible AI so your company can avoid harmful bias. See why 70,000 teams and 30 million people trust Grammarly at grammarly.com slash enterprise. That's Grammarly at grammarly.com slash enterprise.

This show is sponsored by BetterHelp. BetterHelp has been revolutionary in connecting people to mental health services. Using BetterHelp can be as easy as opening your laptop or your phone and clicking a button, and the session begins.

Clients are able to choose in what way they would like to communicate with me, whether video or on the phone or chat texting. BetterHelp is there when you need it, and that's what makes all the difference. Visit betterhelp.com slash podbusiness to get 10% off your first month. Therapists were compensated. ♪♪

China's deep-seek AI model upended markets. Largely, though, those AI infrastructure plays. What about model providers? Clearly, open source is impacting AI development. We're joined by Thomas Wolfe now. He's the chief science officer at Hugging Face. An open source and collaborative platform for AI builders. Is R1 the huge breakthrough that the market clearly thinks it is?

I think there is definitely a bit of an overreaction. We've seen a steady increase in open source model performance, but we also have to be honest, it's the first model, I would say, that really reached the performance of closed source. So the gap is closed now between closed source models and open source model performance.

Let's just talk about how the derivatives have expanded because you are all about open source. You're about community, you're about the innovation that one model can provide to many others. From what we understand from Clem, the CEO over with Hugging Face, he's saying that there have been 500 derivatives already created from DeepSeek. How are you seeing an exponential growth of the use of this model?

Yeah, it's even growing, right? So I was just checking, I can give you fresh new stats just now. We're over 670 models already created by the community. More than 3.2 million downloads of all these models with $700,000 for the original model. So I think that's really the power of open source, right? We see basically a growth of more than 30% of downloads day to day.

which kind of is a testament to all this ecosystem that is already starting to build around DeepSeq, with basically all these companies, all these teams, all these organizations that are already taking this open source model and fine-tuning it, adapting it, testing it on the R news cases. So that's maybe the most beautiful thing about these open source capabilities, which is seeing this ecosystem grow live, day by day, around the original DeepSeq model.

Thomas, we can't separate the story of deep-seek from the geopolitics here. What concerns do you have that a reaction from Washington or other governments might be to restrict open source use in some fashion as a way of keeping technology out of China's hands?

Yeah, I think there's a lot of, I would say, a lot of way of reframing these stories as a U.S.-China, but really the more general story here is the open source effect of this model. It could have come out of almost any countries, and I would expect actually open source model to keep coming out of China, but we also have European companies like Mixtra starting to open source model, and very soon more teams.

More generally, I think hopefully we'll see a move from a geopolitical interpretation of this story to really interpretation as open source versus closed source. And our deep belief at Hugging Face is open source is really the way to foster development, to foster breakthrough technology, to foster growing communities and a lot more business use cases, basically.

Thomas, as a software creator yourself, you must have some admiration for what DeepSeek has pulled off, apparently at a much lower cost than rivals here in the U.S., but I bet you have some questions, too. What would you like to learn more about what the company did and how it pulled it off?

So what is interesting is we already started a tugging phase to actually explore, you know, can we reproduce this model? Can we reproduce it in particular in the open, right? So we have, right, the deep-seq model. I was trying it last year, last day, yesterday.

And it does really offer what we see in the benchmark, which is it's a very powerful model. But what we would like to know is exactly, can you retrain it? Can you actually apply the same recipes to more models? So we've started this project called Open R1, which is basically an open reproduction of the DeepSeq pipeline. Thankfully, they shared a lot of details on how they train the model, much more than we've seen recently. So I'm

quite confident in the coming months we'll be able to basically understand all the breakthrough that went into making this model. What does it mean for a closed source focused open AI or for Anthropic in this moment?

Well, I think it's quite positive. And you saw probably the reaction of some outman, right? I think for good competition in the field, it is something that is just a net positive for developing a technology. And in particular here, because a lot of this model, and as I was saying, a lot of recipes are open. It means I would be very surprised they don't directly take the recipes that can be useful for improving both OpenAI, Anthropic, Google.

or the coming LAMA models, and that these models basically quickly catch up. So I think it's a good thing. That's also the good thing about open source is because you share a lot of information about your models. You actually lift the whole field up with you by explaining how you make your breakthrough. But some people don't want to see everyone rise up. They don't want to see China rise up.

Ultimately, is that a false narrative, a straw man to be even saying that China is behind the U.S. when there's open source communities such as yours?

Yeah, the open source community generally don't really know any border, right? As I was saying, today it's a Chinese team, but generally, you know, it's just a very smart H-form team, right? There are teams like that in many countries. And that's why we keep saying, you will keep seeing new actors in AI, right? So I think ultimately, I believe in fair markets. I think having fair competition, sharing more, actually, you know, is a good way to make progress together. So...

You know, in the future, we would like to see much more actors active in AI. We would like to see, basically, large AI companies a bit everywhere in the world. And I think that's maybe the first step, you know, in this direction. Are we expecting to see similar innovations coming from Europe? There have been questions and concerns raised in the European Union about the level of regulation and that it might be restricting development in AI.

Yeah, it's obviously a big discussion, right? I was at Davos a bit last week, and definitely a lot of the discussion was, okay, what is about Europe regulation? I just think the teams are really good. I would be surprised, for instance, in the UK, in Germany, you know, in France, that we don't see all these teams that basically help train, you know, the models of Mistwell, at Meta, at OpenAI, at Anthropic. We've seen a lot of people also leaving OpenAI to start their own startup. So,

I think one takeaway from DeepSeq is that basically the recipe to build a very good quality large language model nowadays is something that's almost accessible to everyone, right? So I think all these people leaving the big tech companies will start out here. And because you kind of just need a few millions, I would be surprised we don't see a lot more teams coming out of Europe in particular, but also the region this year. Thomas Wolff, Chief Science Officer at Hugging Face. Thanks for joining us.

♪♪♪

Microsoft is in talks to acquire the U.S. arm of TikTok. At least that's according to President Trump last night, who did not elaborate further. Microsoft declined to comment. Let's bring in Bloomberg's balance of power co-host Kayleigh Lyons for more. Kayleigh, is this another case of Trump trying to make a deal happen? He casts himself as the consummate dealmaker. We are so short on details, though. Walk us through what more we know, if anything, beyond last night.

Well, the details are pretty limited, Mike. He was asked directly by a reporter if Microsoft was in talks to acquire TikTok, and the president said, quote, I would say yes. And that's really all we got on that specific matter. As you said, Microsoft isn't commenting on this either. Keep in mind, though, that during the first Trump administration back in the summer of 2020, Bloomberg did report Microsoft was looking at acquiring TikTok's U.S. operations back in the day.

back then when the first kind of pressure around divesting or banning it over national security concerns was arising Oracle was reportedly looking at it too and we know just last week when the Stargate project was being announced at the White House and Larry Ellison of Oracle was in the room Donald Trump suggested he'd be open to Ellison buying TikTok or Elon Musk when he was asked about that we know there are other players

involved as well, including Frank McCourt, who you speak with frequently on this program, who have made bids. And it does seem that is what Donald Trump's preference really is here, a bidding war. He said as much last night that he thinks bidding wars result in the best deals. As for what that deal ultimately looks like, that remains unclear. What he was clear on yesterday when speaking to the House Republican Conference at his Doral Club in Florida is that he does not want China involved in whatever happens here. Of course, China is going to have some say in that ultimately.

Would the US government have involvement still, Kayleigh?

Well, that's the question. He's talked about this 50% ownership structure, whoever buys it, sharing equity with the United States, essentially, so it would be some kind of half-and-half deal. But it remains unclear whether or not that can be done by the now April 4th deadline, which has been extended by Trump's executive order. There's also a massive question around whether antitrust concerns would be raised by some of these individuals or companies like Microsoft acquiring operations of this size, which could be tens of billions of dollars. And obviously, as you guys well know, 170 million users are

in the U.S. While Donald Trump and the people he's instilled at places like the FTC or at the Department of Justice may have some alignment on this, it still could face scrutiny, especially from big tech skeptical lawmakers on Capitol Hill. And where the bite dance would ever sell it. Kayleigh Lyons. Yeah, thanks so much. Welcome back to Blue Bag Technology. I'm Caroline Hyde in New York.

And I'm Mike Shepard in San Francisco. Let's discuss the impact that DeepSeek is having on U.S. tech companies now with Bloomberg's Shireen Ghaffari. Shireen, thanks so much, and thanks for all your reporting today. There is a lot of soul-searching going on in Silicon Valley over the past 48 hours. Tell us a little bit more about what's happening. They set up war rooms. What is the conversation like?

So the top AI labs right now in the US are trying to figure out how the Chinese startup DeepSeek was able to catch up so quickly, right? You have their latest R1 model really exploding on the scene in the past week. And everyone is astounded by how competitive it is. It is actually leading by some metrics on these AI models that have taken Western companies years and a lot more money, the Chinese company DeepSeek claims to build.

We can all debate as to how much it was copying, building off Western technology, but the key question is, will CapEx remain the same? And ultimately, what it means for OpenAI and Anthropic in terms of reducing the cost of their models. What do you think the ripple effects are, Shireen?

Yeah, I think we're starting to see people really question whether we need these astronomical budgets for building the most advanced AI models and whether, you know, because DeepSeq was able to make some gains in how efficiently they use the computing power, if those gains can be applied now to U.S. companies and sort of questioning why U.S. companies didn't come up with that first.

It's a fascinating big take. Go and read it from Shireen Ghaffari and the meta war rooms that are currently being set up. We appreciate it. Now, we've got to assess what President Trump had to say about the competition from Deep Seek. I think if it's fact and if it's true, and nobody really knows if it is, but I view that as a positive because you'll be doing that too. So you won't be spending as much and you'll get the same result, hopefully.

The release of deep-seek AI from a Chinese company should be a wake-up call for our industries that we need to be laser-focused on competing to win. Let's talk about that competition. Jacqueline Rice-Nelson is with us, co-founder and CEO of Tribe AI, provides AI services for leading enterprises. And I have a feeling a lot of them are rushing to understand how they can use this cheap and innovative AI model.

Absolutely. First, thrilled to be here. I think this 2025 is the year for enterprises to get ROI from their AI efforts and the ROI calculation just changed dramatically. So I think that kind of is the headline for businesses, which is use cases that weren't possible just a few weeks ago or even last week are now possible.

And I think we're about to see an explosion of AI experimentation and also value delivery to organizations and enterprises alike. Have any of those organizations been reticent to use it simply because of where it was born in China? It's a great question. I actually had the same question and dove into this myself.

And I think it all comes down to how you use these models. So there are lots of ways to be using LAMA or open source models today in ways that have lots of guardrails that are set up to be more secure and can be run locally on your own environment. In many ways, actually, it's almost easier with open source than it is with closed source models.

Everything comes down to the implementations of these models. I think today a lot of businesses have these same questions, but that's in many ways why they're sort of coming to Tribe and trying to help navigate the landscape across all of these different models. What are best to use and when and for what types of use cases and how to set it up in the optimal ways.

Jacqueline, one of the key questions that's come up in relation to DeepSeek is the one on spending. How is that factoring into the way you are guiding people through the AI landscape now? Is it a different message that you're having to deliver about efficiency and about bringing a product to market for a lot less?

So the short answer is yes. I think that we are, I compare it to a highway. We have just added a lane or multiple lanes to the highway. When you do that, the traffic does not go down. Actually, it only increases. And that's because driving and going on road trips gets more attractive. The same is true right now for AI. So I think we're about to see really an explosion of activity.

And that's sort of across companies, across enterprises, within the large, and to your point on efficiency and sort of product innovation, I think we're going to see proliferation across these use cases. And then that doesn't even include the large hyperscalers who are still on this quest for AGI. And nothing changes there. I don't think anyone is taking their foot off the gas. And if anything, competition has just increased dramatically.

And the open source model, is this something that is feasible to be scaled at large enterprises, big corporations that also may have security considerations and other factors to bring in that might make them question whether open source is the right choice for them?

Yeah, I think there are a lot of considerations that go into what is the right model selection. And so I think it's really clear that we have entered an era and actually I've always thought we were in this era, which is that we are in a multi-model world.

businesses need to be able to build to swap between models seamlessly, to optimize for performance, for latency, for accuracy, and for cost. And what that means is they also need the guardrails, the security, the infrastructure to set those things up in proper ways. They also need the evals, the evaluation frameworks to actually be able to compare and contrast across models so they know what models to use.

And so I think while there might be some hesitance, we have done lots of development on closed source models and lots of development on open source models. And the considerations still come down to what is that company trying to achieve and what are the things we're optimizing for? That's what determines the model. So I, oh please go ahead.

I'm interested in that because I want to go back to sort of the Jevons paradox piece that, you know, the more AI innovation there is, ultimately the more people on those highways, the more innovation. But also the same amount of compute. Can you speak a bit to like the inferencing part of this? Because that seems to be the silver lining that indeed NVIDIA is trying to see from this. Look, you're still going to need our GPUs, just not for training, it's for inference.

I think that's fair. Look, I think we have not cracked the code on the hardware that's needed. It's clear that compute is still needed at really high levels. There's no way that we have, even with these new deep-seek models, we haven't found the answers yet, right? We're not stopping. And I think that that's the bull case for how and why you're going to see continued investment here, is that we are in a continuous innovation cycle, and ultimately I think that drives utilization.

And I think the question is for what? And then the other piece is that ultimately consumers are the biggest winners here. And I think that is the most exciting story. There's a lot of fear. There's a lot of China versus US dynamics. I think the reality is that competition is good.

And what I'm hearing from the US companies is, you know, nothing has kind of, the US large AI companies, nothing has kind of blown their socks off, right? This is innovation they at least feel that they already have. And the difference is that there has been an optimization for open source and an optimization for cost in ways that they have not done. And so I think we're about to see a lot of fast follow. And Sam Altman signaled as much last night. Yeah, he said we're going to drop models faster.

Is it going to drop the cost? Because that must be something that puts off your clients. It's definitely going to drop the cost. And I actually think we are likely to see a cost curve decrease that's dramatic from them, but also a latency decrease. And I think that that's how they're going to play the game, is to try to now one-up completely on both dimensions.

Jacqueline Rice-Nelson, CEO of Tribe AI. Thank you. Caro. Let's just check on these markets because these are all the questions that investors are trained on at the moment. What does it mean for compute going forwards? What does it mean for the competitors in the generative AI model space? And what does it ultimately mean for software too? We're clawing back some of our losses of yesterday, but not much. Eight tenths of a cent higher on the NASDAQ 100. Meta at a new record high. We'll talk more about

alarmer there. But Bitcoin up nine tenths of a percent. We're going to delve into the world of crypto and the ripple effects in a moment where 102,000 move on. The individual movers that you've got to keep your eyes trained on have been, of course, some of the points contributors. We've seen Meta, new record, as I mentioned. Apple's been doing significantly well. Nvidia bounces back, but hardly at all, up 2.8 percent after a 17 percent sell off. We're only up about 30 billion. It lost 600 billion yesterday. Tesla off by 2 percent. Remember, earning

EARNINGS ARE ALMOST UPON US. TESLA TOMORROW OUT OF THE GATE. COMING UP, MORE ON MUSK'S WORLD AND QUESTIONS AROUND HIS APPROACH TO EFFICIENCY BECAUSE HE'S RATHER GOOD AT IT. THIS IS BLOOMBERG TECHNOLOGY.

89% of business leaders say AI is a top priority, according to research by Boston Consulting Group. But with AI tools popping up everywhere, how do you separate the helpful from the hype? The right choice is crucial, which is why teams at Fortune 500 companies use Grammarly.

With over 15 years of experience building responsible, secure AI, Grammarly isn't just another AI communication assistant. It's how companies like yours increase productivity while keeping data protected and private.

Designed to fit the needs of business, Grammarly is backed by a user-first privacy policy and industry-leading security credentials. This means you won't have to worry about the safety of your company information. Grammarly also emphasizes responsible AI so your company can avoid harmful bias. See why 70,000 teams and 30 million people trust Grammarly at grammarly.com slash enterprise. That's Grammarly at grammarly.com slash enterprise.

Grammarly. Enterprise-ready AI. In 15th century Florence, the great inventor Leonardo da Vinci dreamt of creating a flying machine. But something kept getting in his way. Admin. Piles of it. Luckily, Leo used the smart buying tools on Amazon Business, so he could work more efficiently. With the extra time, he not only invented the flying machine, but actually built it. Magnifico. Incredibile. Splendido. Whoa. Easy there, Leo. Splendido.

Amazon Business, your partner for smart business buying. Elon Musk's X has announced a new partnership with Visa. The social media platform is tapping the payment company to enable digital wallets on its app and website in what marks the first step for the company to create the so-called Everything app. Now the move comes as bankers are now looking to offload some $3 billion of X's debt.

According to sources, PIMCO Apollo are now said to be among the asset managers looking to purchase a portion of the debt being sold by a group of banks led by Morgan Stanley. Some interesting sweeteners in that offering too. And staying with Elon Musk, many are now scrutinizing the tech billionaire's approach to efficiency with his own companies. This as he gears up to cut spending and waste in the U.S. government with, of course, the Doge initiative. For more, believe it or not, Craig Trudell is the perfect person to discuss this to

How good is, you know, market efficiency in the private sector? Well, I think it's worth kind of, you know, taking a look at his track record and maybe sort of, you know, being open to the idea that there's maybe some more similarities in how he has run his companies and, you know, how he's been doing it.

how his criticism of the government is characterized, then he is possibly led on. It's been the case that Tesla, for example, took a good decade before it was making its investors any money.

And there was plenty of waste, even according to him, or inefficiency, according to him, within the company that was sort of allowed to fester for some time before he turned things around. And so no one's disputing that Tesla and SpaceX aren't quite efficient and now quite lucrative companies for their shareholders, but you can find your fair share of inefficiency or waste

within his own companies and even on his own earnings calls, him sort of talking about ways in which his companies haven't necessarily been as efficient as you would expect from somebody who's about to run Doge.

Craig, your reporting pointed to some pretty visible examples of maybe a profligate approach, flying tires from the Czech Republic to the U.S. to ensure their delivery. That had to have been expensive. How does that kind of approach translate to government, where the consequences and stakeholders are very different?

I think what Dana Hull and San Francisco pointed out in this story is essentially that there's a willingness to sort of allow for some inefficiency if it means other deliverables.

Within the government, I think the answer that we're clearly going to get is something different in terms of not so much forgiveness of, okay, yes, we were inefficient, but we delivered X or Y.

What we're seeing from Doge early on is this relatively new X account that is just highlighting all the ways that the government is blowing taxpayer dollars and all the ways in which Musk and his team are coming in and righting those wrongs in his view.

I think he's not showing a whole lot of willingness to cut slack in terms of trying different things or trying things and then not going well as he maybe has been in his time as CEO of many of these companies.

Bloomberg's Craig Trudell, thank you. Turning to crypto, deep-seek sell-off posed a threat across all markets, including crypto, with Bitcoin seeing its biggest intraday drop in more than a month. All this comes just as President Trump signed an executive order last week calling for the creation of a White House advisory group on digital currencies. Joining us now to discuss all of this is Meltem Demiror's Crucible Capital Group general partner and founder,

Meltem, we really have to get right to it. We were talking about Elon Musk, Department of Government Efficiency. One of the biggest questions facing your industry is regulation. David Sachs, the new AI in cryptos are. What are you looking for him and the Trump team to clear away?

Look, I think I'm going to say something controversial. The last four years, Biden administration, the lack of clarity on regulatory policy in many ways, I think was actually constructive for the crypto industry in the sense that there was not a lot of external pressure.

Clarity, I think, can be really challenging for markets. Yes, there are obviously bright spots. The rollback of SAB 121 will now allow banks to hold crypto on their balance sheet, make it easier for institutions to hold crypto. Sure, that's great. But clarity, I think, also creates more perspective on what will work and what won't work.

In crypto so far, the view has been number go up, right? We sometimes joke Bitcoin is number go up technology and there's a meme in the industry, "Wag me, we're all gonna make it." I think what's happening in week one of the Trump administration is this clarity is helping people and particularly markets understand that we are not all going to make it. There are going to be winners and there are going to be losers. And this administration has made it very clear that they're perfectly content picking winners and losers.

Are AI coins going to be winners in the future? There's been this dovetailing of the AI trade within crypto and they've taken a brutal hit, of course, in the current narrative.

Look, I think the AI crypto narrative is a confusing one. Two things I look at: one, how does crypto make AI better or safer? And the jury is out there. And then the reverse is, how does AI make crypto better or safer? There, I think we've used AI to create new casinos in crypto in the form of agents or online accounts tied to LLMs that are creating their own coins and launching their own coins.

To me, that's not what's interesting. What I'm looking at is how can crypto and some of the opportunities around aggregation, optimization, and financialization help crypto make AI better. And there I'm really looking at what Apple is doing. Smaller models run locally on devices, I think is a very interesting opportunity for crypto. Obviously, the big story around DeepSeek is what's going to happen to all of this energy and compute capex.

And I think there are some of the early efforts we're seeing in crypto to aggregate the long tail of compute into these marketplaces where you can pay in stablecoins or dollars is really interesting. It goes all the way back to Filecoin of old. But I'm interested, Meltem, in what the regulatory spirits have meant underlying all of what is meant and felt like number go up on Bitcoin. What has the trading told you?

Yeah, if we look at markets, right, what I think is always so interesting is you see a lot of sentiment online and it's easy to say something, but if you want to know what people are truly thinking, you have to look at markets and you have to look at flows. The biggest week of inflows we had

in the last year into crypto ETFs, which were a great proxy, $4 billion the week that Trump won the election. Last week, with all of the regulatory, quote unquote, clarity, or at least setting of direction with Trump administration and the executive orders being signed, $2 billion in inflows. So the expectation is always better than the reality, right? Markets not reacting positively. CME Bitcoin futures contract

had the biggest drop in open interest in its entire trading history yesterday. So it is very clear that on crypto, traders' markets are feeling overextended, just like they did on the big AI names, and there has been a pullback. The question is, how much of that is going to come back with clarity, and how much of that is going to come back when we actually start to see reality catch up with hype?

Trade the rumor, sell the news, Malcolm Demers. It's so good to have you on. Crucible Capital Group, general partner. We thank you. Meta is set to report fourth quarter earnings after the bell tomorrow. For more on what we can expect from the social media giant, Bloomberg's Kurt Wagner joins us now. Kurt, Meta somehow escaped the big sell-off from DeepSeek yesterday. Why is that and what does that tell us going into the results tomorrow?

Yeah, it might be two things. I think one is we you may recall, Mike, at the end of last week, they announced their big plan for the year, all the spending they were going to be doing on AI. You know, they announced that Threads was going to start running ads like they kind of front run to their own earnings a little bit with some of the biggest news that they have planned for 2025.

And so I think maybe some people just saw a little bit less uncertainty with Meta because some of that was out there. I think the second is that this, you know, deep seek stuff in a way sort of validates the AI strategy that Meta has been pushing toward this whole time. You know, Mark Zuckerberg has been arguing for more than a year that where this was going was open source and that that is where Meta was going to go as well, was to create an open source model that other people could build on. Now we saw that is what deep seek did.

is building as well. So there's obviously more competition for Meta here. But I think the strategy that they were employing might be seen as the best way forward here. So perhaps some investors were less spooked because they thought the road ahead for Meta made a little bit more sense than some of the others. Gene Munster, Deepwater Asset Management reflected that exact issue. So too did Citi. And they're also thinking that maybe the AI models becoming cheaper and cheaper are also going to just really push on the overall profitability of the advertising improvements. Is that what we're going to have to hear? Are

return on AI investment from Mark as well as his investment going forward in CapEx? Yeah, this has been one of the biggest questions, not just for Meta, but for all these tech companies is when is all this investment going to pay off? I think in 2024, actually, Meta did probably a better job than a lot of its peers at sort of conveying where AI was impacting the product, right? Not only in making the ads more efficient, but in dispersing their AI assistant across all of their different apps, having to

the Ray-Ban smart glasses, right? They had sort of tangible products that they could roll out and show people. Obviously, with this massive investment they're going to be making in 2025, it will be even more important for them to continue to show that. But given what they did in 2024, at least it seems like the street sort of knows what to expect from them a little bit. So we'll see if they're able to continue that tomorrow for earnings. Shares at a record high, currently trading $677. Kurt Wagner, we thank you. That does it for this edition of Bloomberg Technology. You do not want to forget...

our podcast. You can find it on the terminal as well as online on Apple, Spotify and iHeart. This is Bloomberg Technology.

Join Bloomberg in Chicago or via live stream on March 11th for the Future Investor, Finding the Opportunities. This 2025 event series will examine how companies are investing in their businesses to create efficiencies, innovating their products and services, and improving the customer experience. This series is proudly sponsored by Invesco QQQ. Register at BloombergLive.com slash Future Investor Chicago. That's BloombergLive.com slash Future Investor Chicago.