WEBVTT - Bridgewater Co-CIO Greg Jensen Talks AI Boom

0:00:02.440 --> 0:00:06.760
<v Speaker 1>Bloomberg Audio Studios, Podcasts, radio News.

0:00:06.920 --> 0:00:09.920
<v Speaker 2>We're going to talk about Bridgewater, the world's largest hedge fund,

0:00:10.080 --> 0:00:14.200
<v Speaker 2>is on the rebound. The Pure Alpha strategy, very highly watched,

0:00:14.560 --> 0:00:18.320
<v Speaker 2>is up eighteen percent through June after a week twenty

0:00:18.360 --> 0:00:21.320
<v Speaker 2>twenty three. That's all according to people familiar with the matter.

0:00:21.520 --> 0:00:24.439
<v Speaker 2>Bridgewater is also plowing into the high flying world of

0:00:24.560 --> 0:00:27.520
<v Speaker 2>artificial intelligence, joining us now to talk about it more

0:00:27.560 --> 0:00:30.680
<v Speaker 2>as Greg Jensen, the co chief investment officer at Bridgewater

0:00:30.720 --> 0:00:34.959
<v Speaker 2>associates and Greg talking about the performance at Bridgewater has

0:00:35.000 --> 0:00:38.320
<v Speaker 2>clearly been up double digits this year. What is happening

0:00:38.360 --> 0:00:40.560
<v Speaker 2>that's going right? What are the bets that you guys

0:00:40.600 --> 0:00:43.280
<v Speaker 2>made off that you are seeing now play into twenty

0:00:43.280 --> 0:00:43.760
<v Speaker 2>twenty four.

0:00:44.760 --> 0:00:47.880
<v Speaker 1>Well, I think Bridgewater can benefit in multiple ways.

0:00:48.000 --> 0:00:51.320
<v Speaker 3>The basic idea of Pure Alpha is that for over

0:00:51.360 --> 0:00:54.480
<v Speaker 3>forty years, we've built up our understanding of how macro

0:00:54.960 --> 0:01:00.000
<v Speaker 3>developments impact markets, and we continue to improve that process.

0:01:00.120 --> 0:01:01.920
<v Speaker 3>But there's been so many things, as you guys are

0:01:01.920 --> 0:01:05.520
<v Speaker 3>talking about here, so many major macro events going on

0:01:06.080 --> 0:01:09.280
<v Speaker 3>this year that have pressured markets. The fact that the

0:01:09.280 --> 0:01:11.840
<v Speaker 3>FED had to tighten slower than it expected to at

0:01:11.840 --> 0:01:14.280
<v Speaker 3>the beginning of the year, the ripple that had through

0:01:14.319 --> 0:01:14.920
<v Speaker 3>the fact.

0:01:14.680 --> 0:01:16.880
<v Speaker 1>That US growth has been generally resilient.

0:01:16.959 --> 0:01:19.560
<v Speaker 3>All of those pressures add up to opportunities for US

0:01:19.600 --> 0:01:23.320
<v Speaker 3>and particularly differential pressures across countries. After the tightening cycle

0:01:23.400 --> 0:01:26.160
<v Speaker 3>was very global. The impact of that tightening cycle across

0:01:26.200 --> 0:01:28.800
<v Speaker 3>the world has been quite different, creating a tremendous amount

0:01:28.800 --> 0:01:30.800
<v Speaker 3>opportunities for a strategy like perov.

0:01:30.720 --> 0:01:32.000
<v Speaker 4>What do you expect going forward?

0:01:32.040 --> 0:01:34.920
<v Speaker 5>I was watching last night in preparation a panel you

0:01:34.959 --> 0:01:37.720
<v Speaker 5>did at Davos, which you know is back in January

0:01:37.800 --> 0:01:41.080
<v Speaker 5>you were skeptical that we'd get six cuts this year,

0:01:41.120 --> 0:01:43.600
<v Speaker 5>and at that time the market was pricing that in

0:01:44.720 --> 0:01:49.120
<v Speaker 5>Now you know, we're we're only looking at two maybe one,

0:01:49.800 --> 0:01:52.880
<v Speaker 5>and yet the market is up, you know, fifteen percent.

0:01:53.480 --> 0:01:56.120
<v Speaker 4>What's driving that and can it continue going forward?

0:01:56.680 --> 0:01:56.920
<v Speaker 1>Yeah?

0:01:56.920 --> 0:01:58.640
<v Speaker 3>Well, when you talk about the equity market, I think

0:01:58.640 --> 0:02:01.800
<v Speaker 3>the biggest question, right, the thing that reconciles the circle

0:02:01.880 --> 0:02:04.400
<v Speaker 3>between the interest rate market and the equity.

0:02:04.000 --> 0:02:05.680
<v Speaker 1>Market is productivity.

0:02:05.840 --> 0:02:09.560
<v Speaker 3>Is there going to be a major productivity effect of

0:02:09.600 --> 0:02:12.600
<v Speaker 3>the new technologies, machine learning, et cetera coming through because

0:02:12.639 --> 0:02:15.840
<v Speaker 3>you need it really to reconcile the level of stock market.

0:02:15.840 --> 0:02:18.440
<v Speaker 3>The expectation for earnings going forward are quite high. At

0:02:18.440 --> 0:02:20.680
<v Speaker 3>the same time, the expectation the interest rate market is

0:02:20.720 --> 0:02:23.560
<v Speaker 3>that the central Bank will be able to cut rates

0:02:24.000 --> 0:02:25.240
<v Speaker 3>that services.

0:02:24.800 --> 0:02:26.360
<v Speaker 1>Inflation will continue to come down.

0:02:26.400 --> 0:02:29.519
<v Speaker 3>So you need this kind of perfect situation of earning

0:02:29.560 --> 0:02:32.799
<v Speaker 3>staying strong, demand staying strong enough to make those earnings

0:02:32.800 --> 0:02:36.160
<v Speaker 3>strong without keeping inflation above the FEDCE target. How do

0:02:36.200 --> 0:02:39.040
<v Speaker 3>you get there? You've got to get there through higher productivity.

0:02:39.040 --> 0:02:42.040
<v Speaker 3>So that's the big question is how much productivity When

0:02:42.080 --> 0:02:45.079
<v Speaker 3>will it occur? Because to reconcile the markets, you need

0:02:45.120 --> 0:02:48.120
<v Speaker 3>a pretty big jump in productivity to allow higher earnings

0:02:48.480 --> 0:02:49.480
<v Speaker 3>and lower inflation.

0:02:49.880 --> 0:02:51.760
<v Speaker 6>Well, I definitely want to get to machine learning, but

0:02:51.800 --> 0:02:54.480
<v Speaker 6>before we do that, let's wrap in FED risk and

0:02:54.520 --> 0:02:56.960
<v Speaker 6>political risk into one question. Because I was reading a

0:02:57.000 --> 0:02:59.760
<v Speaker 6>note from Goldman saying that the risk of Trump being

0:02:59.800 --> 0:03:02.520
<v Speaker 6>real elected and putting in place some of those tariffs

0:03:02.560 --> 0:03:05.639
<v Speaker 6>that he's talked about, it could lead to five extra

0:03:05.720 --> 0:03:06.520
<v Speaker 6>FED hikes.

0:03:06.720 --> 0:03:08.240
<v Speaker 4>So not talking about cuts there.

0:03:08.440 --> 0:03:09.080
<v Speaker 1>And when you think.

0:03:09.000 --> 0:03:12.080
<v Speaker 6>About all the known unknowns that are out there, especially

0:03:12.120 --> 0:03:15.000
<v Speaker 6>as it relates to politics right now, are you hedging

0:03:15.120 --> 0:03:18.000
<v Speaker 6>for any risk of something like that scenario playing out?

0:03:18.880 --> 0:03:21.080
<v Speaker 3>Yeah, I think you have to be really thoughtful because

0:03:21.080 --> 0:03:25.400
<v Speaker 3>there's so much political risk going on, and the shift

0:03:25.520 --> 0:03:29.280
<v Speaker 3>generally you're seeing across the world is kind of dissatisfaction.

0:03:28.760 --> 0:03:30.400
<v Speaker 1>With the people in power.

0:03:30.880 --> 0:03:35.080
<v Speaker 3>Populism in many cases being continuing to be on the rise,

0:03:35.760 --> 0:03:38.840
<v Speaker 3>and a Trump versus Biden or Trump versus whoever the

0:03:38.880 --> 0:03:43.720
<v Speaker 3>Democratic nomine ends up being is very significant change in policy,

0:03:43.760 --> 0:03:45.720
<v Speaker 3>one of the more significant elections. It will probably happen

0:03:45.760 --> 0:03:47.960
<v Speaker 3>very quickly, just like there was the significant corporate tax

0:03:48.040 --> 0:03:51.280
<v Speaker 3>cut when Trump came in before, he's coming in ready

0:03:51.320 --> 0:03:51.920
<v Speaker 3>with tariffs.

0:03:52.000 --> 0:03:53.880
<v Speaker 1>That's the big first thing, you know.

0:03:53.920 --> 0:03:56.080
<v Speaker 3>The second thing will be the continuation of the tax cuts.

0:03:56.920 --> 0:03:59.839
<v Speaker 3>All of this move towards populism is pointing towards an

0:04:00.080 --> 0:04:06.640
<v Speaker 3>inuation of more government deficits, probably more inflation. Tariffs, at

0:04:06.680 --> 0:04:08.680
<v Speaker 3>least in the short run and probably in the long run.

0:04:08.800 --> 0:04:12.480
<v Speaker 3>Our view would be that they're inflationary and a Trump

0:04:12.480 --> 0:04:16.360
<v Speaker 3>presidency the thing that we kind of most be concerned

0:04:16.360 --> 0:04:18.720
<v Speaker 3>about is the long end of the rate curve.

0:04:18.800 --> 0:04:19.920
<v Speaker 1>Probably bonyots fed.

0:04:20.040 --> 0:04:23.159
<v Speaker 3>They'll try to get it easy fed in and probably

0:04:23.279 --> 0:04:25.440
<v Speaker 3>have the most significant challenges on the long end of

0:04:25.440 --> 0:04:30.120
<v Speaker 3>the curve in terms of bonds. But those policy changes

0:04:30.200 --> 0:04:31.880
<v Speaker 3>are going to be significant. And if you look at

0:04:31.920 --> 0:04:35.960
<v Speaker 3>basically just understanding global macroeconomics today versus twenty five years ago,

0:04:37.760 --> 0:04:43.680
<v Speaker 3>understanding government interaction with markets has become so critical.

0:04:44.720 --> 0:04:47.159
<v Speaker 2>How do you then position if you believe inflation is

0:04:47.160 --> 0:04:49.480
<v Speaker 2>in the future, do you go short the long end

0:04:49.480 --> 0:04:51.440
<v Speaker 2>of the curve? And do you do it significantly?

0:04:53.040 --> 0:04:53.200
<v Speaker 1>Well?

0:04:53.200 --> 0:04:55.560
<v Speaker 3>We try to take a diversified set of positions so

0:04:55.600 --> 0:04:58.599
<v Speaker 3>that no one position can be too dominant in our portfolio.

0:04:58.920 --> 0:05:02.800
<v Speaker 1>The it's tricky, right the long end of the YEO curve.

0:05:03.040 --> 0:05:04.840
<v Speaker 3>A lot's going to depend on politics, right, and a

0:05:04.839 --> 0:05:07.279
<v Speaker 3>lot's going to depend on how unified the government is.

0:05:07.360 --> 0:05:09.240
<v Speaker 3>Will Trump be able to get everything he wants? Or

0:05:09.720 --> 0:05:11.400
<v Speaker 3>if the Democrats are elected, will they be able to

0:05:11.400 --> 0:05:14.240
<v Speaker 3>get everything they want? Dividing government will generally be good

0:05:14.240 --> 0:05:17.320
<v Speaker 3>for bonds, and a unified government one where or the

0:05:17.320 --> 0:05:19.560
<v Speaker 3>other Republican or Democrat would probably be bad.

0:05:19.600 --> 0:05:21.240
<v Speaker 1>Anytime these governments are unified.

0:05:21.320 --> 0:05:25.080
<v Speaker 3>Right now, you're seeing movements towards populism, movements toward bigger

0:05:25.120 --> 0:05:28.240
<v Speaker 3>deficit spending, and we think that's the more likely outcome

0:05:28.279 --> 0:05:30.320
<v Speaker 3>is that you continue down that path. But there's a

0:05:30.360 --> 0:05:32.400
<v Speaker 3>lot of change along the rank along the way. And

0:05:32.440 --> 0:05:34.520
<v Speaker 3>in the short run, what's going to matter is the

0:05:34.680 --> 0:05:37.880
<v Speaker 3>is FED policy and is the FED constrained? We think

0:05:38.000 --> 0:05:44.560
<v Speaker 3>the Fed's itching to ease. They want to essentially have

0:05:44.680 --> 0:05:47.400
<v Speaker 3>the stats come in a way that allows them to and.

0:05:47.400 --> 0:05:48.960
<v Speaker 1>We're close to that they probably do.

0:05:49.279 --> 0:05:52.160
<v Speaker 5>Right, with one hundred billion dollars plus of debt servicing

0:05:52.200 --> 0:05:55.080
<v Speaker 5>costs every month, we need rates to be lower.

0:05:56.040 --> 0:05:58.200
<v Speaker 4>At what point does and I know we want to

0:05:58.200 --> 0:05:59.240
<v Speaker 4>get to AI machine learning.

0:05:59.240 --> 0:06:00.720
<v Speaker 5>I want to get there as well, But at what

0:06:00.760 --> 0:06:05.119
<v Speaker 5>point does this massive debt and the servicing costs really

0:06:05.120 --> 0:06:06.839
<v Speaker 5>start to become a problem.

0:06:06.520 --> 0:06:09.799
<v Speaker 4>For you know, the ten year yield or the strength

0:06:09.800 --> 0:06:10.320
<v Speaker 4>of the dollar.

0:06:11.560 --> 0:06:13.120
<v Speaker 1>Well, and that's what I think we're going to find out.

0:06:13.400 --> 0:06:13.560
<v Speaker 5>Right.

0:06:13.600 --> 0:06:15.280
<v Speaker 3>What you can see right now is there's not much

0:06:15.320 --> 0:06:17.280
<v Speaker 3>risk premium in long duration bonds.

0:06:17.320 --> 0:06:18.520
<v Speaker 1>There hasn't been a problem.

0:06:18.560 --> 0:06:21.440
<v Speaker 3>That doesn't mean there won't be, right, and I think

0:06:21.839 --> 0:06:23.640
<v Speaker 3>the main thing I'd say is the destiny here is

0:06:23.680 --> 0:06:26.359
<v Speaker 3>that governments are going to push until we hit the limits.

0:06:26.440 --> 0:06:26.560
<v Speaker 4>Right.

0:06:26.560 --> 0:06:28.480
<v Speaker 3>You see that in countries like Brazil or whatever, where

0:06:28.520 --> 0:06:31.560
<v Speaker 3>they're at the point where fiscal spending hits the limits.

0:06:31.560 --> 0:06:35.160
<v Speaker 3>You saw that for a brief moment in twenty twenty.

0:06:34.920 --> 0:06:36.240
<v Speaker 1>Two with the UK.

0:06:37.600 --> 0:06:40.320
<v Speaker 3>It Liz Trust moment with Trust moment. But in the

0:06:40.440 --> 0:06:43.280
<v Speaker 3>US you're still probably always away. But I think the

0:06:43.320 --> 0:06:47.000
<v Speaker 3>destiny of it because of the populism and sort of

0:06:47.040 --> 0:06:49.640
<v Speaker 3>the amount of things that need to be fixed in

0:06:49.640 --> 0:06:52.599
<v Speaker 3>the US, the likelihood is we find there that now

0:06:52.680 --> 0:06:54.680
<v Speaker 3>I would guess that's a couple of years away, but

0:06:54.800 --> 0:06:57.279
<v Speaker 3>that we're continue to press in the direction until we

0:06:57.320 --> 0:07:00.719
<v Speaker 3>find the limit. That easing fiscally will continue to be

0:07:00.760 --> 0:07:04.240
<v Speaker 3>the path until that limit is hit.

0:07:04.160 --> 0:07:05.920
<v Speaker 4>Until the bond vigilantes come back.

0:07:06.400 --> 0:07:10.000
<v Speaker 3>Yeah, and it won't be speculators are too small relative

0:07:10.040 --> 0:07:14.160
<v Speaker 3>to what's going on. It's really the FED, the and

0:07:14.720 --> 0:07:16.800
<v Speaker 3>generally the size of that issuance. And so what is

0:07:16.880 --> 0:07:19.480
<v Speaker 3>going to be the limit is inflation. If there's no

0:07:20.040 --> 0:07:22.920
<v Speaker 3>inflation problem, there won't be a problem for bonds. It

0:07:22.920 --> 0:07:25.960
<v Speaker 3>it won't come from anything other than inflation currency problem.

0:07:26.000 --> 0:07:28.200
<v Speaker 2>I want to talk about something people are really excited about.

0:07:28.320 --> 0:07:32.160
<v Speaker 2>That's artificial intelligence, greg By. We recently reported that just

0:07:32.160 --> 0:07:35.680
<v Speaker 2>about a week ago, exactly a week ago, Bridgewater started

0:07:35.720 --> 0:07:39.680
<v Speaker 2>its own AI strategy in a new format here, a

0:07:39.760 --> 0:07:42.760
<v Speaker 2>new fun format. But also you were an early investor

0:07:42.920 --> 0:07:46.360
<v Speaker 2>in open AI in anthropic How do you think about

0:07:46.400 --> 0:07:49.680
<v Speaker 2>investing in AI at Bridgewater? Do you buy into those

0:07:49.760 --> 0:07:52.280
<v Speaker 2>high flying names or are you better off doing it

0:07:52.320 --> 0:07:55.040
<v Speaker 2>in the private names personally given where valuations are today.

0:07:55.920 --> 0:07:58.920
<v Speaker 3>Yeah, well, i'd say when we think about it from

0:07:58.920 --> 0:08:01.240
<v Speaker 3>a pure AFFAM perspective, So what we built up at

0:08:01.280 --> 0:08:04.280
<v Speaker 3>Bridgewater is an expert system based on everything we've learned

0:08:04.280 --> 0:08:08.960
<v Speaker 3>over the forty years, mainly on how economic cycles work,

0:08:09.760 --> 0:08:13.560
<v Speaker 3>and big productivity miracles are pretty separate from economic cycles.

0:08:13.600 --> 0:08:15.080
<v Speaker 3>So when you look at what we're doing in pure AFA,

0:08:15.120 --> 0:08:18.480
<v Speaker 3>we're generally trying to make sure we don't have too

0:08:18.520 --> 0:08:23.400
<v Speaker 3>much exposure to whether the productivity miracle works out or not.

0:08:23.520 --> 0:08:25.440
<v Speaker 3>So when we trade US equities, we trade them kind

0:08:25.440 --> 0:08:28.840
<v Speaker 3>of with AI hedge because the pressures, the normal cyclical

0:08:28.880 --> 0:08:30.680
<v Speaker 3>pressures will play out on the regular companies. And when

0:08:30.680 --> 0:08:32.600
<v Speaker 3>you look at US stocks, it's kind of the regular

0:08:32.640 --> 0:08:35.000
<v Speaker 3>companies that Wou'd say are pretty expensive.

0:08:34.640 --> 0:08:35.880
<v Speaker 1>Relative to the rest of the world.

0:08:36.280 --> 0:08:38.640
<v Speaker 3>Mediocre companies in the US have a US premium in

0:08:38.679 --> 0:08:40.640
<v Speaker 3>them relative to the rest of the world. Now, the

0:08:40.640 --> 0:08:42.720
<v Speaker 3>great companies in the US are the companies hitting the

0:08:42.720 --> 0:08:45.520
<v Speaker 3>cutting edge. That's a different story, and it's very different

0:08:46.200 --> 0:08:48.960
<v Speaker 3>than nineteen ninety nine two thousand. Been at Bridgewater for

0:08:49.000 --> 0:08:51.680
<v Speaker 3>twenty nine years and I lived through ninety nine two thousand.

0:08:51.960 --> 0:08:56.320
<v Speaker 3>That bubble was all about future expectations. Most of what's

0:08:56.360 --> 0:09:01.160
<v Speaker 3>priced into equities today actually say, are based on what's

0:09:01.200 --> 0:09:03.959
<v Speaker 3>happening right now. I mean, what's happening with Nvidia's demand

0:09:04.280 --> 0:09:06.239
<v Speaker 3>just is incredible, and whether.

0:09:05.960 --> 0:09:08.760
<v Speaker 1>It's slightly over priced or slightly underpriced or whatever, it's

0:09:08.800 --> 0:09:09.360
<v Speaker 1>a close call.

0:09:09.480 --> 0:09:12.480
<v Speaker 3>This is not like the ninety nine bubble, and at

0:09:12.559 --> 0:09:15.800
<v Speaker 3>least in my view, because the earnings for those companies

0:09:15.840 --> 0:09:17.680
<v Speaker 3>are actually there in a way they weren't there for

0:09:17.840 --> 0:09:20.480
<v Speaker 3>a company like Cisco in nineteen ninety nine, So there

0:09:20.520 --> 0:09:22.640
<v Speaker 3>is a big difference. Now you still need those earnings

0:09:22.679 --> 0:09:28.560
<v Speaker 3>to be sustainable, which requires a continued growth in those areas,

0:09:29.120 --> 0:09:31.840
<v Speaker 3>but not a not an outlandish one. And that really

0:09:31.840 --> 0:09:34.600
<v Speaker 3>comes to what internally, So there's we need to understand

0:09:34.679 --> 0:09:37.839
<v Speaker 3>AI from a macro perspective, and then we need to

0:09:37.920 --> 0:09:40.520
<v Speaker 3>understand it from a as a user a micro perspective.

0:09:40.600 --> 0:09:43.840
<v Speaker 3>Right that RI Labs, which is twenty five people, we

0:09:43.880 --> 0:09:46.640
<v Speaker 3>set aside a bridge where are fully focused on using

0:09:46.679 --> 0:09:49.000
<v Speaker 3>machine learning for everything you need to do to invest

0:09:49.000 --> 0:09:49.560
<v Speaker 3>in markets.

0:09:50.440 --> 0:09:54.440
<v Speaker 1>That group we're hands on. We see both the problems.

0:09:54.559 --> 0:09:57.400
<v Speaker 3>There are many problems that language models off the shelf

0:09:58.000 --> 0:10:00.000
<v Speaker 3>have many problems with them, but if you can combine

0:10:00.240 --> 0:10:04.880
<v Speaker 3>them with data models, you can create this incredible strength.

0:10:05.080 --> 0:10:08.400
<v Speaker 3>And right now I expect over the next five years

0:10:08.440 --> 0:10:12.679
<v Speaker 3>so much of what human investors and analysts can do

0:10:13.200 --> 0:10:15.720
<v Speaker 3>can be better done with machine intelligence.

0:10:15.720 --> 0:10:17.440
<v Speaker 1>So that's been a journey for me. I've been looking at.

0:10:17.320 --> 0:10:20.680
<v Speaker 3>This for Bridgewater s built an expert system based on

0:10:20.800 --> 0:10:26.240
<v Speaker 3>human intuition being translated into algorithms for AI. Having machine

0:10:26.320 --> 0:10:30.360
<v Speaker 3>learning generate the ideas, generate the algorithms. That's been kind

0:10:30.360 --> 0:10:32.920
<v Speaker 3>of a dream for me, and the technologies really came

0:10:32.960 --> 0:10:35.880
<v Speaker 3>together in twenty twenty two or so that the pieces

0:10:35.920 --> 0:10:37.920
<v Speaker 3>came together such that I thought we were ready to

0:10:37.960 --> 0:10:40.720
<v Speaker 3>build an artificial investor, and we've moved down that track

0:10:40.760 --> 0:10:41.520
<v Speaker 3>at rapid rate.

0:10:41.760 --> 0:10:43.680
<v Speaker 2>Greg, thank you so much for giving us a preview

0:10:43.760 --> 0:10:46.840
<v Speaker 2>on this new strategy over at Bridgewater, and of course

0:10:46.960 --> 0:10:49.120
<v Speaker 2>the view of a very complicated markets of course, that

0:10:49.240 --> 0:10:53.319
<v Speaker 2>is Greg Jensen, CIO at Bridgewater Associates Cocio