US Launches Probe on Potential DeepSeek-Nvidia Link
24:00 Share

We've all been there. Your flight was cancelled and everyone is trying to rebook at the same time. Please hold. Estimated wait time is 25 minutes.

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. Amazon Business, your partner for smart business buying. Bloomberg Audio Studios. Podcasts. Radio. News.

Welcome to the Bloomberg Daybreak Asia podcast. I'm Doug Krisner. On today's episode, we'll be taking a look at the global macro landscape. In a moment, we'll be chatting with David Espel, co-CIO at Mount Lucas Management, plus a discussion on the Fed's path ahead with Bill Adams. He is the chief economist at Comerica Bank. But we begin in Tokyo.

Joining us now is Catherine Thorbeck. She is Asia Tech columnist for Bloomberg Opinion. She's been very busy this week writing about the story on DeepSeek, the Chinese AI startup. Catherine, thank you for making time to chat with us. I know it's been a busy week for you. What I want to begin with, the story that just broke late in the day U.S. time. American officials are investigating whether DeepSeek essentially is a

circumvented restrictions on advanced semiconductors from Nvidia. Now we know that the Biden administration cut off China's access to a range of some of Nvidia's most powerful products. Now, sand seems to be shifting a little bit. Can you fill us in on what you understand the story to be?

Right. So this news just came out, as you mentioned, just a few hours ago. And I don't think it really caught anybody by surprise that the U.S. would launch a probe into what sort of chips DeepSeq was using to train its models and whether it was able to circumvent export controls. And it's too soon to sort of know the outcome of that investigation. But honestly, it wouldn't entirely surprise me if they were able to get their hands on some NVIDIA chips, some of the more advanced chips that have

been cut off to China. These export controls have proven pretty porous in a number of instances. That said, we don't know at this point in time. And DeepSeq has, it is out there that DeepSeq was hoarding some of the older NVIDIA chips before these export controls came in and whether they were able to get their hands on the more advanced models, that's sort of the question now.

And I think a lot of people in the U.S. are just very curious how they were able to create these AI models with the technology that they had. So it makes sense that the U.S. would launch this investigation now. It wouldn't entirely surprise me if they were able to obtain some of these more advanced chips. I'm not sure that would necessarily change the narrative about how DeepSeq has sort of upended big tech and the AI race right now. I still think that they were able to prove that they, you know,

you know, are able to come up with AI breakthroughs despite all of these restrictions. And I think that that still is very much top of mind for, for big tech right now. So our reporting indicates that when it comes to the chip story, that American officials are looking at whether they were acquired these Nvidia chips through a third party or third parties, plural in Singapore. But there are a number of other investigations, uh,

going on in regards to DeepSeek. And that would include Microsoft and OpenAI looking at whether OpenAI's model was somehow used in training DeepSeek. Is that right?

That's correct. And so right now Microsoft is probing whether DeepSeq used outputs from OpenAI's models to train their own model, which would technically be a breach of the terms of service for using OpenAI's models. But to me, this really reeked of a lot of irony. You know, OpenAI has trained its models by basically scraping writing and art from the entire Internet,

you know, without really getting meaningful consent from artists and writers to do that. So it did seem a little bit odd and it has rubbed a lot of people sort of the wrong way. A lot of jokes have been made about, oh, now open AI, all of a sudden cares about, you know, intellectual property rights. You know, whether DeepSeek obtained this data improperly or not is still sort of the subject of investigation. But I think it does put open AI in a little bit of an awkward spot right now. So what does the DeepSeek story tell us or,

allow us to understand about the evolution of the AI industry in China? Right. So I think DeepSeek really caught the West by surprise when it released its most recent model on January 20th. And I don't think it necessarily had to catch so many people entirely off guard and, you know, cause this multibillion dollar stock market sort of route. And some of those losses have been pared back over the week.

But I first wrote about DeepSeq back in June, more than six months ago when it released its V2 model, one of its earlier AI models.

And I was just saying that this is actually a very impressive AI model. You know, for a long time, the West and Silicon Valley has sort of claimed dominance in the AI race, and they thought they were years ahead of China. And I think policymakers in Washington have thought, you know, the export controls, the restrictions on chips that they've been putting towards China, that would really hold China back. And DeepSeek really upended that.

that narrative, you know, they showed that Chinese AI, Chinese AI startups are actually really able to innovate and are able to do it at a fraction of the cost that these Silicon Valley tech giants have been doing. So it really sort of, I think people were really, really impressed by this R1 model that they put out

on January 20th. And this R1 model, it can really, it's a reasoning model. So it can really sort of think through questions the way, you know, a human would think through questions. It's been on par with, you know, OpenAI's models on a number of industry benchmarks. And it really, I think, you know, caught the sort of tech billionaires in the U.S., some of these very powerful companies, totally off guard this week. To what extent do you believe there is an entrenched bias between

that China doesn't innovate, that it merely copies? So I think we've seen this narrative for a while. And I think that there's a number of reasons why this comes up over and over again and why we're continually shocked by China's tech breakthroughs.

And, you know, some of it is, I think, you know, a language barrier, separate social media ecosystems and sort of a shrinking number of foreign reporters actually in China give the U.S. very little visibility into some of the more exciting sort of developments in China's tech sector right now. And I also think that there's

sort of an insular mindset in the U.S. that Silicon Valley and sort of these Silicon Valley leaders, a lot of them seem to think that they're the only ones who can really lead the charge here. And I think it leaves huge blind spots to what's going on in China. A lot of sort of the coverage of China that we read in the U.S. is very focused on geopolitical rivalries and a lot of voices in Washington see any sort of tech breakthroughs in China as a threat to democracy everywhere. And whether or not

That's true. I think it does create these big blind spots to some of the sort of actual people working in China and, you know, the more than one billion people that live there and a lot of times are innovating and sort of driving tech forward despite their government and not because of it.

I know the TikTok story is something that you and I have discussed in the past and the proprietary nature of that algorithm, something the Chinese would like to protect. I found it curious that in the case of DeepSeek, the developers opted for kind of an open system. Did that surprise you at all?

It didn't entirely surprise me. I think that they were sort of working with what they've got and open source models you can sort of tweak and a lot of people can work on them at once and you can sort of -- they're a little bit cheaper to run than sort of what OpenAI is doing which is just training bigger and bigger and bigger proprietary models. So it didn't totally surprise me that they did this.

I think it does sort of raise larger questions of sort of the future direction of the sector and whether it will be dominated by these like smaller startups that are really perfecting and tweaking these open source models or whether sort of the open AI route, which is just creating bigger and bigger and bigger models and spending more and more on building out these models.

whether that will sort of prove superior. And we'll have to see how it does play out in the long run. So we were talking a moment ago about OpenAI, and you referenced the fact that OpenAI essentially was trained using the U.S. intranet. And I'm wondering, from the point of view of DeepSeek, do we know much about how this model was trained at all? So I think it's interesting. I think when it comes

When it comes to DeepSeek, one of the big issues that sort of as this swarm of American and international users have discovered as they've logged on and downloaded this app and it's shot up to the top of Apple's App Store this week, one of the things that they've discovered is what's off limits behind China's great firewall and whether that's talking about Xi Jinping, talking about China's human rights record in Xinjiang, where there have been documented human rights abuses of Uyghur Muslim minorities,

And right now, when you chat with Deep Seek about that, you know, it'll initially say, let's change the subject. Let's talk about something else. There are sort of workarounds that I was able to play around with this week where I did get it to say and sort of encode, you know, that Xi Jinping has faced international criticism for his treatment of Hong Kong protesters and political dissidents.

So it does show that DeepSeq was trained on data beyond the Great Firewall, and I'm sure that's going to cause a headache for DeepSeq. I don't think Chinese authorities are going to crack down on this tool that has brought such a positive international spotlight to its tech sector and its tech capabilities.

But what I do think that this all shows is that there's a lot more at stake in this global race for AI dominance between the U.S. and China. You know, as more people turn to these tools for research, for homework help, they could also be used to sort of covertly influence our ideas and influence our ideologies, whether that's from Beijing or from the U.S.,

So I think that, you know, there's obviously a lot of money and stock market value at stake here. But these tools are also very powerful and they have the power to really influence people. And I think that we're going to see more of that sort of playing out in the long run. Catherine, we'll leave it there. Always a pleasure. Catherine Thorbeck, Asia Tech columnist for Bloomberg Opinion, joining us here on the Daybreak Asia podcast.

I'm Alpine skier Michaela Schifrin. I've won the most World Cup ski races in history. But what does success mean to me? Success means discipline. It's teamwork. It's the drive and passion inside of us that comes before all recognition. And it's why Stiefel is one of the fastest growing global wealth management firms in the country. If you're looking for success, surround yourself with the people who will get you there.

At Stiefel, we invest everything into our advisors so they can invest everything into their clients. That means direct access to one of the industry's largest equity research franchises and a leading middle market investment bank. And it's why Stiefel has won the J.D. Power Award for Employee Advisor Satisfaction two years in a row.

If you're an advisor or investor, choose Stiefel. Where success meets success. Stiefel Nicholas & Company, Inc., member SIPC and NYSE. For J.D. Power 2024 award information, visit jdpower.com slash awards. Compensation provided for using, not obtaining the award. The average time to hire for most organizations is 30 to 45 days. Are you tired of a costly and lengthy hiring process? Simplify and speed up your recruitment by using the experts at Express Employment Professionals.

Reduce time to hire, cut down on multiple interviews, and lower your recruitment costs. Whether you're looking for contract workers or a new team member, our streamlined recruiting, candidate screening, and hiring process is more efficient than hiring on your own. Visit ExpressPros.com today. Haven't used a staffing company before? We know you enjoy learning new things and keeping up with trends and efficiencies.

Welcome back to the Daybreak Asia podcast. I'm Doug Krisner. There was certainly volatility in the currency markets in late New York trading today, and the catalyst was President Trump. He was the first to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a decision to make a

He said that he will follow through on his threat to impose 25% tariffs on imports from Canada and Mexico this Saturday. For a closer look now at how these tariffs might affect the macro outlook, I'm joined by David Aspell. He is partner, also co-CIO at Mount Lucas Management, joining us from just outside Philadelphia. Thank you for making time to chat with us, David. Do you have a sense of what the tariff threat may mean to global markets, generally speaking?

I mean, it certainly adds a lot of uncertainty, doesn't it? You know, we've got a new administration that's just come in that, you know, clearly sees volatility as a feature, not a bug. And, you know, they're using tariffs, amongst other things, for negotiating leverage to try and

pushed through some pretty large macro policy goals. And these policy goals run directly through currency markets, through bond yields, and they're going to have large impacts on economies everywhere. It's going to be a fascinating four years, I think. Yeah, definitely. The dollar has been a big beneficiary. We can talk about Fed policy in a moment, but do you expect the dollar to remain very strong here against the majors? I

I think it's hard to say. I think what's going on is that we've gone through a period of US exceptionalism over the past few years where the US economy has been stronger than others. And so that's why you've seen it particularly against Japan or against the euro.

The euro has been weak because the European economy has been weak. It's the same with the UK. And if you think why that is, it's because the US has higher rates currently. It can survive higher rates because it's got a slightly different type of economy where mortgage rates don't feed through in quite the same way. And now you've got an administration that seems like it's unleashing inflation.

some animal spirits. So as long as that continues and the U.S. does better than others, then yeah, I think the U.S. dollar can stay fairly well bid. There's been some debate in Washington as whether or not these policies regarding tariffs are inflationary. Do you have a view on that? I don't think they are.

I think they might be for a short period of time. As a macro trader and how I look at inflation now, I think the Fed looks at inflation is maybe a little different to how people on Main Street see it. Inflation to me is an ongoing process of higher prices for a period of time, whereas tariffs to me are a one-off price shock. Now, it could be that they get embedded, particularly at a time when people are nervous about inflation. But I think

I think that the negatives of inflation will take, you know, as they slow down the economy in some places, if people are nervous about them, will take the edge off them. And I think generally the inflation path to me seems quite well set over the next few years, partly because of housing. Just a few weeks ago, the debate in the bond market was whether or not we were going to see a 5% yield on the 10-year. I'm looking at something now that's closer to 4.5%. What's your view on yields going forward?

Unless something goes very awry in inflation, which is possible but seems unlikely, I think we've seen the peak in yields. I think the economy can survive them, but I think you've got a Fed that wants to reduce rates and is able to reduce rates over time. If you look at where the Fed projections were at the end of the year, they thought that with a 2.5% core PCE at the end of this year,

they would have rates down at 3.9%, which to me seems like they're making it an easier bar to cut rates rather than a higher bar. So I think they offer good value. It seems that they're useful in a portfolio context currently because if the economy slows down, the Fed is able to cut, which I think is real helpful.

In terms of the story across the Asia-Pacific, it's mainly been about weak domestic demand in China and sluggishness in the industrial economy. Do you have a sense of what China is going through, where we are now in this cycle, and whether or not things are beginning to improve? Yeah, I think that's been...

one of the major stories in the global economy over the past few years. The extent to which China has been slowing, they've clearly got a consumption problem that their economy is incredibly unbalanced. They run very large surpluses and rely on other people to soak up those surpluses. And the rest of the world has decided that rightly, I think that it's not going, they're not willing to absorb those surpluses, which has led to some slowdown in China. I think what they're doing is slowly, they are trying to rebalance.

And I thought that the stimulus policies they put in around rates and the other sort of loosenings and the cleaning up of balance sheets, I thought they were quite powerful. And you're just about now starting to see those come to fruition. The data seems like it's starting to turn up. It seems like the stimulus was powerful around the consumer.

It's just I worry that it's going to run into a buzzsaw of the new administration and tariffs. And I hope that we can get a deal done that rebalances their economy and means that we don't have to, the rest of the world doesn't have to soak up their surpluses without too much collateral damage. I'm curious about whether you're finding any value across markets in Asia. Now, Chinese assets may be inexpensive on a relative basis. We can go there. But I'm also curious about what you're seeing in Japan.

I would say that we generally do systematic trend following. So what we're doing is just following prices. If things are going up, we're long them. If things are going down, we're short them, which means we've been in the yen trade for a long period of time. We've been long dollars and short yen. That's been good. I think that trade can continue because of the interest rate differential.

Alongside, on the discretionary side where we run macro strategies, I think the Chinese equity market is incredibly cheap as long as it can do the things. It just needs to be able to do the rebalancing to get through this period and come to a trade agreement and operate in a much more normal way as something closer to a regular emerging market.

If it can do that, then the equity market there is incredibly cheap. If it cannot do that, then I think we'll end up, this will be another false dawn of which we've had a few over the past years. But I'm pretty hopeful that the new administration is able to do a trade deal and we see some brighter times ahead in Asia. David, we'll leave it there. Thanks so much. David Espel, partner, co-CIO at Mount Lucas Management, joining us here on the Daybreak Asia podcast.

So let's turn our attention now to the U.S. economy. Today, we learned that economic growth expanded at a solid pace in the fourth quarter, albeit a little less than forecast. Fourth quarter GDP growing at a rate of 2.3 percent and consumer spending advancing at an annual rate of 4.2 percent. Let's take a closer look now with Bill Adams. He is the chief economist at Comerica Bank. Bill, thanks for making time to chat with us.

Give me your overall impression of this GDP print. What do you think it tells us? So the headline was a slowdown, but the details were strong. If you look at my preferred measure of core real GDP, that's real final sales to private domestic purchasers, that tracks GDP's trend by leaving out the effects of inventories, leaving out the effects of trade and government spending, that grew 3.2% annualized in the fourth quarter. So that's a

Good news, the U.S. economy is in good shape. Growth momentum is stronger than the headline in the fourth quarter and closer to the annual growth rate, 2.8%, which is excellent for the U.S. economy. So at the same time today, we learned that weekly jobless claims came in well below estimates. And I'm wondering whether when you look at the claims data on top of the GDP data, you come away with the conclusion that, yeah, the Fed's going to be on hold for a bit longer. It's

It sure looks like it. The labor market data in the second half of 2024 were worrying the Fed. You'll recall that was when we were all talking about the SOM rule, former Fed economist Claudia Somm's observation of when the three-month moving average of the unemployment rate rises by half a percent or more from its 12-month low.

Typically, the economy has been in recession. And so that that sound rule triggered. And I think that was one of the big contributors to why the Fed cut interest rates a full percentage point between September and December. But now we have this solid GDP growth data in hand. Labor market data like that claims data, like the jobs report, have stabilized.

And that means that the Fed is less worried about the economic outlook and they can they feel they have the latitude to refocus their attention on bringing inflation closer to that 2 percent target. So, Bill, I'm curious to get your take on how the policies from the Trump administration as they relate to deportations may impact the labor market. Do you have a sense of that?

So directionally, if the U.S. labor market has fewer workers, that would tend to push the unemployment rate down. That's just labor demand is not going to decline as much as labor supply would. And so it's a question of magnitude how large this effect is going to be.

but I think the Fed is clearly taking a wait-and-see approach to see how much the change in immigration policies affects labor supply over the next 12 months, next 24 months, as well as waiting and see for what exactly happens with tariff policies, tax policies. But directionally, it all kind of points to

either a tighter job market with changes in immigration policy or higher prices for consumers and for businesses if tariff rates go up, or just an overall hotter economy if we have tax cut

directed stimulus. So all of that would be reasons for the Fed to slow the move lower in interest rates or to just hold interest rates steady for an extended period. That seems to be the theme right now. So if you extend that then into the market's response, I'm talking about the treasury market here, is it likely that that 10-year yield could approach 5% this year?

I'm expecting that for most of the year, the 10-year is probably going to range between 4.25% and 4.75%. I think a lot of the concerns about upward pressures on inflation or pressures on the labor market or on the fiscal deficit, if we have more stimulus, all of that is, I think, pretty clearly reflected in market pricing today. What about the knock-on effect that it would have with the mortgage market?

I think we'll have less of a recovery of home sales than we would have in the absence of these policies. But, you know, that's relative. And if you're comparing to 2024, that was the weakest year since the mid-90s for existing home sales. I think the direction, even with mortgage rates staying high, is probably a bit higher this year. It's just it's going to be a more measured recovery for existing home sales than we otherwise would have seen. New home sales, I think, look

better because there's home builders are offering incentives to home buyers and the kind of the absolute number of existing homes that are listed right now is still pretty low. And that is directing home buyers, the ones who can afford to be buying in a market like this, to have a look at new construction. I mentioned a moment ago that consumer spending in the fourth quarter of last year was very robust at 4.2% growth.

How do you evaluate the American consumer right now? Are things beginning to become less robust in a way, perhaps?

That's a funny question. So on average, the American consumer is in great shape, benefiting from the big increases in the stock market, big increases in home equity for homeowners, and that wealth effect is funding robust spending on discretionary categories, on new vehicles. New vehicle sales also helped along by

Both fears of higher prices because of tariffs and then also the idea that maybe those incentives for EV sales could go away. But in general, the average American consumer is performing well, but that average is really being held up by high-income, high-wealth, affluent households.

And if you look at the typical consumer, the median consumer, they are much more affected by the increase in rents, by the increase in the cost of food and other essentials. And that side of the American consumer is struggling to an extent still. I think it's been directionally a little bit better over the last six to 12 months.

But if you look at high frequency measures of food insecurity, for example, it does look like you see more stress right now than you saw before the pandemic hit and transformed the U.S. economy. Bill, we'll leave it there. Thank you so much for joining us today. Bill Adams there. He is the chief economist at Comerica Bank. Joining us here on the Daybreak Asia podcast.

Thanks for listening to today's episode of the Bloomberg Daybreak Asia Edition podcast. Each weekday, we look at the stories shaping markets, finance, and geopolitics in the Asia Pacific. You can find us on Apple, Spotify, the Bloomberg Podcast YouTube channel, or anywhere else you listen. Join us again tomorrow for insight on the market moves from Hong Kong to Singapore and Australia. I'm Doug Krisner, and this is Bloomberg.

Some people like A and others like B. At BMW, we prefer X. Like the dynamic X3 meant for ultimate exploration.

The X5 built to conquer even the most difficult paths or the pinnacle of comfort and luxury, the X7. And since every X range vehicle is packed with performance and versatility, you'll always get the best of X. The BMW X range. Your next X venture starts here. Hurry into the BMW President's Day sales event and receive exceptional offers through February 28th. This podcast is supported by BetterHelp, offering licensed therapists you can connect with via video, phone,

or chat. Here's BetterHelp Head of Clinical Operations Hesu Jo discussing who can benefit from therapy. I think a lot of people think that you're supposed to be going to therapy once you're like having panic attacks every day. But before you get to that point, I think once you start even noticing that you feel a little bit off and you can't maintain this harmony that you once had in relationships, that could be a sign that maybe you want to go talk to somebody.

There's always a benefit in talking to someone because we can all benefit from improved insight about ourselves and who we are and how we behave with other people. So if you're human, that's like a good indicator that you could benefit from talking to somebody. Find out if therapy is right for you. Visit BetterHelp.com today. That's BetterHelp.com.