Latest / Bloomberg Talks / Goldman Sachs CEO David Solomon Talks ‘Very Robust’ Demand for AI Deals
Transcript
- 0:02Speaker 1: Bloomberg Audio Studios. Podcasts. Radio. News.
- 0:08Speaker 2: Joining us here on set is the Goldman Sachs chairman
- 0:11Speaker 2: and CEO, David Solomon.
- 0:12Speaker 1: David, good morning and very nice to see you.
- 0:14Speaker 3: Good to see you guys, and it's nice to be
- 0:15Speaker 3: back with you again here at the HKMA Summit.
- 0:17Speaker 1: Yeah, it's been busy.
- 0:18Speaker 2: You guys have been extremely busy, you and your team,
- 0:20Speaker 2: underwriting all these deals.
- 0:22Speaker 1: Give us a sense of momentum.
- 0:24Speaker 2: What can we glean from what we've heard into the
- 0:27Speaker 2: next couple of months?
- 0:28Speaker 1: Sure.
- 0:28Speaker 3: You know, certainly activity levels have been very, very high.
- 0:31Speaker 3: And you guys highlighted that last week. We had a very,
- 0:34Speaker 3: very busy week. You know, a combination of a handful
- 0:38Speaker 3: of very significant financings coming together at the same time,
- 0:41Speaker 3: both for SpaceX, for Meta, obviously announcements around the.
- 0:45Speaker 1: Upcoming Anthropic IPO.
- 0:47Speaker 3: You know, I'd say the capital markets remain very active
- 0:49Speaker 3: while there's been a little bit more volatility in the
- 0:52Speaker 3: short run and recalibration in certain parts of the market.
- 0:56Speaker 3: I'd say investors are still very focused on the trajectory
- 0:59Speaker 3: of AI demand, compute demand. And, you know, the capital
- 1:04Speaker 3: markets are open and receiving. And so we found a
- 1:07Speaker 3: very robust demand for these transactions.
- 1:10Speaker 4: I think when you were on stage with Ted Pick
- 1:11Speaker 4: from Morgan Stanley, he had a line about 10 to
- 1:13Speaker 4: 15 trillion, right?
- 1:15Speaker 2: Yeah, the spending, I think.
- 1:16Speaker 1: The spending that he expects, maybe.
- 1:18Speaker 4: I mean, how long do you think this sort of
- 1:20Speaker 4: investment cycle can last still?
- 1:23Speaker 3: Yeah, I think we're early in the build-out of the
- 1:27Speaker 3: infrastructure to support this technology. But I'd also be the
- 1:30Speaker 3: first to caution us all that You know, at the moment,
- 1:33Speaker 3: there's an expectation that.
- 1:35Speaker 1: It's going to be constant in a straight line.
- 1:37Speaker 3: And if you look at any other technology super.
- 1:39Speaker 1: Cycle, you know, it's never constant. There are recalibrations, there
- 1:43Speaker 1: are stops and starts.
- 1:44Speaker 3: You know, I think it's very hard to know how
- 1:47Speaker 3: much will ultimately be spent to build this infrastructure all
- 1:50Speaker 3: over the world. I think one of the things people
- 1:52Speaker 3: don't talk about enough is that there are going to
- 1:54Speaker 3: be new technologies and evolutions in technologies that will change
- 1:57Speaker 3: the trajectory of all this. But I think one thing
- 2:00Speaker 3: that's pretty clear is the demand for compute right now
- 2:03Speaker 3: far exceeds the supply. And certainly in the next 12
- 2:05Speaker 3: to 24 months, there's going to be robust financing to
- 2:09Speaker 3: meet some of that demand.
- 2:11Speaker 1: I think it's a super exciting time.
- 2:13Speaker 3: I can't tell you that it'll go without bumps in
- 2:17Speaker 3: the coming years. But I can say when you step back,
- 2:19Speaker 3: And you look kind of five, 10 years from now,
- 2:22Speaker 3: I'm very excited about how this technology can get deployed
- 2:25Speaker 3: in the economy and really drive productivity gains that can
- 2:28Speaker 3: be really, really fruitful. So investors see that they're looking
- 2:31Speaker 3: to participate. I think one of the things investors are
- 2:34Speaker 3: trying to do is run broad, diversified portfolios. Debt investors
- 2:38Speaker 3: are obviously paid to not make mistakes.
- 2:41Speaker 1: Sure.
- 2:41Speaker 3: And so if they think the market's getting a little
- 2:43Speaker 3: bit more challenging, they ask for higher returns because debt
- 2:46Speaker 3: investors don't get to make mistakes. Equity investors take more
- 2:49Speaker 3: risk because they look for upside. And I think there's
- 2:52Speaker 3: still a lot of upside available as this technology gets
- 2:55Speaker 3: built and deployed and.
- 2:56Speaker 1: The application layer gets built over the next decade or two.
- 2:59Speaker 2: The size of the deals has grown substantially, exponentially. Is
- 3:04Speaker 2: there a theoretical ceiling or do you see the size
- 3:07Speaker 2: of the deals being able to just continue growing? to
- 3:09Speaker 2: become larger and larger from someone who puts these things together?
- 3:12Speaker 1: I think the constraint.
- 3:16Speaker 3: Is not the amount of capital in the world and
- 3:19Speaker 3: the size of these deals on a relative basis. Yes,
- 3:22Speaker 3: there's no question these deals are bigger and the capitalization
- 3:27Speaker 3: of these companies is larger than what we've seen in
- 3:31Speaker 3: other tech super cycles, but we also operate in a
- 3:34Speaker 3: market that's much, much larger where equity market cap has
- 3:37Speaker 3: grown very substantially. Public equity market cap in the world
- 3:40Speaker 3: is about $ 160 trillion. And so there's a lot of
- 3:43Speaker 3: capital out there to support this growth. I think sometimes
- 3:47Speaker 3: investors or the dialogue around this gets confused between risk
- 3:53Speaker 3: reward and what's the appropriate pricing for certain kinds of
- 3:56Speaker 3: risk and how that pricing parameter can change in the
- 3:59Speaker 3: short run as the market moves around versus whether capital
- 4:02Speaker 3: is available. But there's no question, you know, these numbers
- 4:04Speaker 3: are big. And also capital, capital is competitive to some degree.
- 4:09Speaker 1: Rates are higher. Governments around the world are issuing more debt.
- 4:13Speaker 3: You know, that's a competitive factor in all this, but
- 4:15Speaker 3: I don't think it affects availability.
- 4:17Speaker 1: I think it affects cost.
- 4:19Speaker 3: And if cost goes up too much, it changes the
- 4:22Speaker 3: returns in some of these projects. And that will change
- 4:25Speaker 3: the way some of this happens.
- 4:26Speaker 1: And the way it's financed.
- 4:27Speaker 4: And here in Asia, I mean, you guys have had
- 4:29Speaker 4: a really record run. I think in terms of how
- 4:32Speaker 4: much this region has contributed to your total revenue. I
- 4:34Speaker 4: think it's making up 20% now, which has never really
- 4:36Speaker 4: happened in the history of the firm.
- 4:39Speaker 1: It's at the high end of the range.
- 4:40Speaker 3: I mean, if you look over a long period of time,
- 4:42Speaker 3: our business here has been kind of 15% to 20%
- 4:44Speaker 3: of our global business.
- 4:46Speaker 1: It moves up and down.
- 4:47Speaker 4: Is that like a cyclical high, or do you think
- 4:49Speaker 4: this is a structural shift in how Google makes its
- 4:51Speaker 4: money now?
- 4:51Speaker 3: I think Asia and the economies in Asia, particularly the
- 4:54Speaker 3: Chinese economy, are very, very large economies. They're going to
- 4:57Speaker 3: be an important part of the global ecosystem. Goldman Sachs
- 4:59Speaker 3: is a big global firm. You know, we operate globally. And,
- 5:03Speaker 3: you know, we are, you know, to some degree, you know,
- 5:06Speaker 3: correlated to equity market cap growth and GDP growth around
- 5:09Speaker 3: the world.
- 5:10Speaker 1: And this part of the world is always going to
- 5:11Speaker 1: play an important role in that. You know, I think
- 5:14Speaker 1: the U.S.
- 5:14Speaker 3: Is in a position of real preeminence in terms of
- 5:17Speaker 3: its economic position and its innovation ecosystem. But China is
- 5:21Speaker 3: super important and other economies over here are super important, too.
- 5:24Speaker 3: And so Goldman Sachs will always have a significant footprint
- 5:27Speaker 3: over here.
- 5:28Speaker 1: Right.
- 5:28Speaker 2: Just the outlook for next year, what specific verticals of
- 5:31Speaker 2: your business here in the Asia-Pacific are you leaning into?
- 5:36Speaker 3: Well, the thing that's clear in every part of our
- 5:39Speaker 3: business is this technology super cycle is going to dominate
- 5:42Speaker 3: capital markets activity and financing activity broadly until there's some
- 5:46Speaker 3: shift in the direction of travel. So whether here in
- 5:49Speaker 3: Asia or in the U.S. or even in Europe, it's
- 5:51Speaker 3: a technology ecosystem that's really driving capital markets activity and
- 5:56Speaker 3: financing activity. I'd say when you look at M &
- 5:58Speaker 3: A activity around the world, you know, particularly in the U.S.
- 6:02Speaker 3: and Europe, it's more shaped by a regulatory environment that's
- 6:06Speaker 3: allowing deals at scale. I think everybody recognizes the importance
- 6:10Speaker 3: of scale in almost any industry. And so you're seeing
- 6:13Speaker 3: more broad consolidation across industries, and that's leading to more
- 6:16Speaker 3: M & A activity broadly. And our firm has been
- 6:18Speaker 3: a big beneficiary of being in the middle of that,
- 6:20Speaker 3: given our leading franchise.
- 6:21Speaker 1: In the M & A world.
- 6:23Speaker 4: A lot of the growth that you've been seeing is
- 6:25Speaker 4: undoubtedly from the capital markets sort of business. What about
- 6:28Speaker 4: when it comes to asset wealth management? Is there more
- 6:31Speaker 4: of a pivot now in terms of where you get
- 6:34Speaker 4: your earnings growth now?
- 6:35Speaker 3: Well, I mean, Goldman Sachs laid out eight years ago,
- 6:38Speaker 3: this leadership team laid out a strategy to really grow
- 6:40Speaker 3: Goldman Sachs. We've taken the revenue from kind of $ 34 billion,
- 6:44Speaker 3: $ 35 billion, you know, going back to 2018 when this
- 6:48Speaker 3: leadership team came in. The street consensus for the firm
- 6:51Speaker 3: is somewhere in the low 70s this year. So we've
- 6:54Speaker 3: grown the revenue very materially, grown the earnings very materially
- 6:57Speaker 3: over that period of time. I think that compounded revenue
- 7:00Speaker 3: growth rate is in the nines over that period of time,
- 7:03Speaker 3: and the earnings growth is slightly higher.
- 7:05Speaker 1: And so we've been very focused on growing the firm.
- 7:07Speaker 3: One of the places that we've gotten that growth from,
- 7:10Speaker 3: we've gotten the growth really from two places. One, our
- 7:12Speaker 3: core banking and markets business. We've really been able to
- 7:14Speaker 3: grow market share and take more market share.
- 7:16Speaker 1: In that business.
- 7:17Speaker 3: But secondarily, we've got a very powerful asset wealth management platform.
- 7:21Speaker 3: We supervise over $ 4 trillion of assets, and that platform's growing.
- 7:25Speaker 3: We've stated publicly that we're confident we can grow at
- 7:27Speaker 3: high single digits, but it's actually growing.
- 7:29Speaker 1: Faster than that.
- 7:30Speaker 3: And so the growth in assets around the world, the
- 7:33Speaker 3: growth in wealth around the world, we have a very powerful,
- 7:36Speaker 3: ultra high net worth platform. We supervise over two trillion
- 7:39Speaker 3: dollars of wealth for very wealthy individuals.
- 7:41Speaker 1: Around the world.
- 7:42Speaker 3: Those platforms are growing quickly, and that's supporting the overall
- 7:45Speaker 3: revenue and earnings growth of the firm.
- 7:46Speaker 1: That's been part of our strategy for the last seven
- 7:48Speaker 1: or eight years.
- 7:49Speaker 2: You mentioned in the panel just now that the U.S.
- 7:51Speaker 2: economy remains resilient, and I want to ask you about
- 7:55Speaker 2: the midterms coming up. Not so much a political question,
- 7:57Speaker 2: but do you think, is there anything of market consequence
- 8:00Speaker 2: that you think you should be paying attention to as
- 8:02Speaker 2: you gain outcomes?
- 8:04Speaker 1: I think the.
- 8:06Speaker 3: I'm not a political prognosticator, so the outcome will be
- 8:09Speaker 3: the outcome. But I think the big outcome that we
- 8:13Speaker 3: will immediately switch to in all the dialogue is what's
- 8:17Speaker 3: going to happen in 2028 and the 2028 election, and how.
- 8:20Speaker 1: Will the administration shift.
- 8:22Speaker 3: Of course, this will get a lot of attention, and
- 8:24Speaker 3: of course, if there's a swing in the House or
- 8:27Speaker 3: a swing in the Senate and the House, there'll be
- 8:29Speaker 3: a lot to talk about around that. But from a
- 8:31Speaker 3: substantive policy perspective... The big thing is an administration shift.
- 8:35Speaker 1: And, you know, how will that play out over the
- 8:37Speaker 1: next two years?
- 8:37Speaker 3: I think the discussion is very quickly after the midterms,
- 8:41Speaker 3: no matter what the result is, is going to shift to,
- 8:43Speaker 3: you know, what is the outlook for an administration shift
- 8:46Speaker 3: in 2028 into 2029?
- 8:49Speaker 4: I mean, a lot of the Democrats, what they have
- 8:52Speaker 4: been talking about is a lot of these, the opposition
- 8:54Speaker 4: around this whole AI building when it comes to data
- 8:56Speaker 4: centers and the like. I know you wrote recently an
- 8:59Speaker 4: op-ed about the AI safety.
- 9:01Speaker 1: Issue, right?
- 9:02Speaker 4: And that, you know, we really can't put regulations on
- 9:05Speaker 4: AI like you did with the banks, right?
- 9:06Speaker 1: People are saying you could just supply that.
- 9:08Speaker 4: But maybe Wall Street has a bigger role in really,
- 9:11Speaker 4: I guess, scrutinizing the technology and the like. What role
- 9:14Speaker 4: do you think banks actually have in safeguarding that?
- 9:16Speaker 2: Well, I.
- 9:17Speaker 3: The op-ed that I put forward was to try to
- 9:20Speaker 3: participate in the dialogue that this is a very powerful
- 9:23Speaker 3: technology and a lot.
- 9:24Speaker 1: Of wonderful things that are going to come from it.
- 9:26Speaker 3: I'm super excited about what the technology can do in
- 9:28Speaker 3: areas like health and science and also what it can
- 9:30Speaker 3: do in business and enterprise to increase productivity. But it
- 9:33Speaker 3: also needs a governance structure around it. And I think
- 9:37Speaker 3: that governance structure requires some simple principles that have always
- 9:41Speaker 3: been the principles of responsible business and technology deployment. The
- 9:45Speaker 3: first is government has to set some rules of the road.
- 9:47Speaker 3: The second is the people that make the technology have
- 9:50Speaker 3: to be responsible and thoughtful about what they make and
- 9:52Speaker 3: how they sell it into the market. And the third
- 9:54Speaker 3: is that the enterprises like a Goldman Sachs or other
- 9:57Speaker 3: financial institutions or other industrial companies that take that technology
- 10:01Speaker 3: and deploy it into their enterprises have to be responsible
- 10:03Speaker 3: for the way it's deployed.
- 10:04Speaker 1: Goldman Sachs has been using.
- 10:06Speaker 3: AI trading platforms in our firm for 20 years, and
- 10:10Speaker 3: you never hear anybody at Goldman Sachs say, oh, you know,
- 10:12Speaker 3: the platform went and lost us hundreds of millions of dollars.
- 10:15Speaker 3: You know, there are people that are accountable and responsible
- 10:17Speaker 3: for the deployment of that technology, you know, into our platform.
- 10:20Speaker 3: And so I think there are similarities in that. This
- 10:23Speaker 3: is something that we'll have to get right. You know,
- 10:25Speaker 3: will there be bumps and will there be issues?
- 10:27Speaker 1: Of course.
- 10:28Speaker 3: But I think simple principles that we all exercise and
- 10:32Speaker 3: we think about.
- 10:32Speaker 1: In business are important. There need to be rules of
- 10:34Speaker 1: the road.
- 10:35Speaker 3: The people that make the technology need to take responsibility
- 10:39Speaker 3: for what they're doing. And the people that deploy it
- 10:41Speaker 3: have to take responsibility for how it's deployed. And they
- 10:43Speaker 3: have to have risk and control processes and governance around
- 10:45Speaker 3: it that works. And I think that's a good model.
- 10:48Speaker 3: And by the way, that is the fundamental model of
- 10:51Speaker 3: the financial system, particularly after the financial crisis. It's been
- 10:54Speaker 3: built up over the last 15 years.
- 10:55Speaker 1: 20 years?
- 10:56Speaker 2: When it comes to, I think we can trust the
- 10:59Speaker 2: people who know the technology the most, like the anthropics
- 11:02Speaker 2: of the world. And if they tell us they're slightly
- 11:05Speaker 2: more concerned, and I think the underlying thesis there is
- 11:07Speaker 2: we're giving up some human agency to the technology. You
- 11:10Speaker 2: guys have been using it, as you point out, for
- 11:11Speaker 2: 20 years. Do you get a sense that this The
- 11:16Speaker 2: 2.0 version of this technology is somewhat more independent from
- 11:20Speaker 2: what we are used to.
- 11:21Speaker 1: How do you look at that?
- 11:22Speaker 3: You said something that I think we should be very
- 11:23Speaker 3: thoughtful about.
- 11:24Speaker 1: And that's giving up human agency.
- 11:25Speaker 3: Humans need to be accountable for how technology is deployed
- 11:28Speaker 3: into our economy.
- 11:29Speaker 1: And there has to be human accountability.
- 11:30Speaker 3: And so that doesn't mean that these aren't very powerful
- 11:32Speaker 3: models and they have the capability to do lots of things.
- 11:36Speaker 1: Without our oversight.
- 11:37Speaker 3: But I think that's a principle that needs to be
- 11:39Speaker 3: thought about very carefully and debated very carefully as we
- 11:42Speaker 3: move forward.
- 11:43Speaker 4: How do you assess how your bankers and traders are
- 11:45Speaker 4: using these AI tools?
- 11:48Speaker 3: It's interesting that you raise that. I just want to
- 11:52Speaker 3: highlight something. There was an article on Bloomberg that I
- 11:58Speaker 3: think left an impression that's not correct about how we're
- 12:03Speaker 3: deploying AI into Goldman Sachs. What's actually interesting, the article
- 12:07Speaker 3: implied somehow we're taking it away. It's actually the opposite.
- 12:10Speaker 3: We're growing very quickly and very significantly the tools and the.
- 12:13Speaker 1: Access our employees have.
- 12:15Speaker 3: It wouldn't surprise you we have 45,000 employees and not
- 12:17Speaker 3: every employee can get every single tool every single day.
- 12:21Speaker 1: We have a process about that.
- 12:22Speaker 3: But the acceleration and the availability of tools, including coding tools,
- 12:27Speaker 3: you know, to our employees is accelerating very quickly. But
- 12:29Speaker 3: we're trying to put the right governance around it, the
- 12:31Speaker 3: right controls around it. All the things we were just
- 12:33Speaker 3: talking about, we're trying to do it. So it's actually
- 12:36Speaker 3: accelerating very.
- 12:37Speaker 1: Very quickly.
- 12:38Speaker 3: And I think the article kind of misrepresented the direction
- 12:40Speaker 3: of travel.
- 12:41Speaker 4: We spoke with a DBS CEO, and she says, you know,
- 12:43Speaker 4: they're model agnostic when it comes to whether I use U.S.
- 12:45Speaker 4: or Europe and a Chinese model. I mean, how is
- 12:49Speaker 4: it like in Goldman in terms of what models you use?
- 12:51Speaker 3: Well, in Goldman, we have great partnerships with a number
- 12:53Speaker 3: of the big, you know, the big companies. And... Look,
- 12:56Speaker 3: we're very interested in how open weights access will evolve.
- 13:01Speaker 1: But also, again, we're very cautious from a governance.
- 13:04Speaker 3: Perspective because we're held to a very, very high standard.
- 13:06Speaker 3: And so I think there's going to be a big
- 13:08Speaker 3: role for open weight models in the world. But we
- 13:11Speaker 3: don't have to be first in the way that's involved
- 13:14Speaker 3: or deployed inside Goldman Sachs. But I think that's one
- 13:17Speaker 3: of the exciting things that open weight models, and I
- 13:19Speaker 3: think they will proliferate from lots of different places, will
- 13:21Speaker 3: bring costs down.
- 13:22Speaker 1: And I think one of the things that's also interesting
- 13:24Speaker 1: to think about is that in an enterprise.
- 13:26Speaker 3: Like Goldman Sachs, of course, we need foundation frontier models
- 13:29Speaker 3: for certain things, particularly cyber. But there are a lot
- 13:32Speaker 3: of things where we don't need the frontier. A lot
- 13:34Speaker 3: of things where we don't need the frontier to help
- 13:36Speaker 3: our people be more productive, to recreate operating processes. A
- 13:39Speaker 3: lot of places where we don't need the frontier. But
- 13:41Speaker 3: we have great partnerships with the most important companies, as
- 13:44Speaker 3: you would expect us to.
- 13:46Speaker 1: We think they're doing really, really fascinating things.
- 13:48Speaker 3: And we're significantly increasing the pace of access to these
- 13:53Speaker 3: tools to our people. And I think we're actually on
- 13:55Speaker 3: the forefront of what a big enterprise is doing. But
- 13:57Speaker 3: we do it with governance, caution, and appropriately. And You know,
- 14:02Speaker 3: I'm sure you can find, you know, five people inside
- 14:04Speaker 3: the firm that don't have something they want on a
- 14:06Speaker 3: given day. But the reality of it is we have
- 14:08Speaker 3: 45,000 people. We're going to govern it carefully.
- 14:10Speaker 4: David, it's great to have you back in Hong Kong.
- 14:12Speaker 4: David Solomon there, Goldman Sachs chairman and CEO.