高盛CEO David Solomon:AI是否已经见顶?| TomorrowX Summit
- Solomon认为,AI股抛售是漫长投资周期中的预期重估。 Baker指出,许多AI股票跌幅达到50%或以上;Solomon认为,当前估值倍数比6个月前更有吸引力,但市场正讨论AI基础设施投资将达到6万亿美元、8万亿美元或10万亿美元,同时重新校准增长预期。他预计,未来10年、15年或20年,应用层将创造最多价值,而企业采用技术需要时间。
- 高盛的AI机遇,在于把运营效率转化为增长空间。 高盛在45,000名员工中有13,000名工程师;Solomon猜测,“5年后工程师人数可能会更少”,但公司仍需要更多卓越的资深工程人才。自动化开户、KYC和反洗钱流程,应能腾出资源,重新投入人才、增长和业务版图。
- 高盛财富业务的扩张仍取决于增员。 其约45,000名财富客户的可投资资产约为2,500万美元或以上,他们需要投资组合建议、信托与遗产规划服务,也希望尽早获得SpaceX和Anthropic等公司的投资机会。Solomon称这是一门“以人为本、由技术增强的业务”,人员配置制约着业务扩张速度。
- Solomon长期看好美国,同时接受利率走高和显著的政治波动。 他不做未来3至6个月的预测,并认为多极化世界更加脆弱;但他认为,10年期美债收益率达到5%也可以与经济活力并存。Baker认为,转向中间路线的政党或许能更长时间执政;Solomon认同,凝聚中间选民可能很有力量,并表示未来25年会“大举押注”美国。
- Solomon支持Nvidia,但明确表示不对正在发酵的消息作判断。 他说自己不了解事实,呼吁在Jensen作证、事实逐渐明朗之际保持耐心。他称赞Colette Kress和运营团队;Baker称她是“一位非常称职的临时CEO”。
- 高盛面向小企业的AI计划,旨在通过教育、研究和融资渠道推动采用。 Solomon介绍了一项5亿美元承诺,合作方包括Google、Anthropic、OpenAI和Meta;该计划在此之前已通过教育项目服务约20,000家企业。若能做好,创业成本下降或将在未来10年扩大创业机会;但过程“不会一帆风顺”。
- Solomon眼中,AI改善普通人生活最鲜明的例子,是更早发现疾病。 他举例称,Mayo Clinic的工具可在胰腺癌确诊前最多3年检出73%的病例,从而让患者更早接受治疗。Baker回应:“没有什么比生命更珍贵。”
1. 高盛的增长战略:再投资与以人为本的财富业务并行
- Solomon将过去8年的变化概括为:推动一家拥有130年历史、以私人合伙制为核心的机构,转向为公众股东再投资。高盛通过2.3万亿美元资产负债表调动约1,150亿美元资本,每年在技术上投入超过60亿美元。过去8年,营收大致翻倍;他还提到,过去5年的盈利增长接近20%。
- 财富平台服务约45,000名客户,其可投资资产约为2,500万美元或以上;业务既提供SpaceX和Anthropic等公司的早期投资机会,也涵盖投资组合策略、信托和遗产规划。增加员工有助于加快增长;这仍是一门“以人为本、由技术增强的业务”。
2. 长期看好美国,不代表经营环境不会更动荡
- Solomon不做未来3至6个月的预测。他的乐观判断与多极化带来的地缘政治脆弱性并存,也面对持续15年以上的超低利率时代落幕;10年期美债收益率达到5%,仍可能伴随“非常有活力的经济”。
- Solomon表示,在过去12次选举中,有11次出现众议院、参议院或总统职位至少一项易手;美国人始终在要求改变。
- Baker提出质疑:赢得浪潮式选举的政党往往转向极端立场,而非中间路线;若中间派胜选,执政时间可能长得多。Solomon将Texas和Kansas的变化解读为“中间派正在移动”,而非永久转向,并表示未来25年会“大举押注”美国。
3. Nvidia面临的不确定性,并不能判定AI周期走向
- 谈到Nvidia,Solomon明确表示自己不了解事实,掌握的情况也非常有限。他呼吁耐心等待事实浮现和Jensen作证,并称赞Colette Kress及造芯片的团队。他认为Nvidia对美国的技术领导力和安全至关重要,但不知道事态将如何发展。Baker称Kress是“一位非常称职的临时CEO”。
- Baker问,在许多AI股票跌幅达到50%或以上后,AI是否已经见顶。Solomon认为,漫长周期才刚刚开始,当前估值倍数比6个月前更有吸引力。他提到,市场讨论的基础设施投资规模包括6万亿美元、8万亿美元和10万亿美元,这些数字反映了市场在重新校准预期;但他也表示,这不意味着市场的预期就完全准确。芯片和技术成本同样重要。
- Solomon认为,更大的价值池将在未来10年、15年或20年间从应用层涌现。他以Apple为例:Apple在2000年制造电脑;后来,手机平台催生了Uber等概念,时间已是10年之后。
- 技术能力的提升快于采用速度:“技术的发展总是快于企业和个人。”成本下降仍需要组织作出调整。
4. AI正在改变高盛的人员结构和节省资金的用途
- 高盛在45,000名员工中有13,000名工程师,并选出历来占比更高的一批工程师进入合伙人阶层。Solomon猜测:“5年后我们的工程师人数可能会更少”,但会有更多“能力超强、资历深厚的工程师”。
- 除了为员工提供工具,让他们做得更多、更好、更快,管理层也在重建运营流程:最初确定6项,随后又增加2项,包括开户、KYC和反洗钱。目标不止于提高利润率:把效率提升带来的资源重新投入人才、增长和业务版图,能扩大公司承接更多业务、服务更多客户的能力。
5. AI带来的广泛收益,取决于实际可及性和采用进度
- Solomon举出的医疗案例来自Mayo Clinic:AI可在胰腺癌达到可确诊阶段前最多3年检出73%的病例,让许多患者有机会更早接受治疗。Baker回应:“没有什么比生命更珍贵。”
- 高盛宣布投入5亿美元,为曾参加10,000 Small Businesses项目的企业及其他企业创立AI研究院。过去20年,该项目已通过为期6个月的商业教育项目服务约20,000家企业。计划有3大支柱:工具教育、研究和融资渠道;Google、Anthropic、OpenAI和Meta将参与合作。
- Solomon表示,小企业每年贡献美国新增就业岗位的65%,而AI降低了创业成本。他将此称为“加码版”的互联网式普及;若能做好,未来10年可能产生巨大影响。过程“不会一帆风顺”。
完整逐字稿
本集逐字稿暂无译文,当前显示原文。
We probably won’t have as many engineers 5 years from now.
We may have seen the peak.
We’re still in the early innings of a long cycle, with AI tools able to detect 73% of cases up to 3 years in advance.
Nothing’s more valuable than life.
I’d bet heavily on that over the next 25 years.
1. Goldman Sachs and Atreides
First, David, I want to say thank you to Goldman Sachs. Goldman Sachs has been an important partner of iConnections since the beginning. You were one of our 2 prime brokers when we launched. That was very meaningful to me. I want to say thank you for the partnership. I also want to say, Jack Sebastian, I believe he’s in the audience.
Is Jack here?
Yes, Jack is here. Oh, wow.
I think Jack had a lot to do with that. I’ve known him for nearly 30 years, so I’m very grateful to Goldman Sachs as a firm and to Jack personally, and to you for the friendship and partnership we’ve developed over the last few years.
Well, thank you. You’re a great partner, and thank you for having me here. I’m thrilled to be here. This is a super cool event.
Yeah.
2. How Goldman has evolved since 2018
Thank you for coming. So, David, you’ve been CEO since 2018. Let’s talk about how you’ve evolved and grown the firm. Where are we?
The firm’s in a really good place. Given everything that’s going on in the world, the firm is very correlated to growth in the world and very correlated to growth in the United States. My big macro view is that I have no idea what’s going to happen in the next 3 to 6 months. But when I look out over the next 10 years, I’m incredibly optimistic about how technology is driving growth, productivity, and opportunity for society in very powerful ways. The firm is very correlated to that.
The firm has always been a client-centric business. It’s been around for 157 years. Originally, it was all people, with a tiny bit of capital and a tiny bit of technology. Now the firm has a lot of really powerful people, a lot of capital—we’ve got about $115 billion of capital that we leverage into a $2.3 trillion balance sheet—and a lot of technology. We spend over $6 billion a year on technology. We put those together to serve clients.
3. Goldman's ultra high net worth wealth business
The journey over the last 8 years has been about finalizing the move from being a private partnership for 130 years to really being a public company. We can serve our clients and do really well for shareholders. I’m a big believer that no public company can really succeed or perform well on a relative basis if it doesn’t grow, if it doesn’t invest in growth. Partnerships are different because they really pay out to the partners every year. We’ve really started to see the dividends from the culture of running Goldman Sachs to invest in medium- and long-term growth. We’ve roughly doubled revenue over the last 8 years and grown earnings close to 20% over the last 5 years. I think the firm is very well positioned to serve our clients really well, continue to grow, and therefore benefit our shareholders.
Awesome. Wealth has been a really important part of your strategy. We have John Underwood here, so maybe—
Another hugely important Goldman Sachs guy.
Absolutely. Maybe talk about the evolution of the firm’s wealth platform. We have a lot of individual investors here.
Sure. It’s a very unique wealth business. It’s a focused wealth business: an ultra-high-net-worth wealth business, and a very full-service wealth business for people who have, roughly, $25 million or more in investable assets. They need access to great products. You look at all the cool things going on in the world: we have an incredible platform and we’re at the center, where we can provide our clients with access to things like SpaceX and Anthropic early. They need access to portfolio strategy, but they also need trust and estate work. They need planning. They need help with all sorts of things in a complicated financial life, and we can deliver that in a full-service package.
We only have about 45,000 clients globally, but that business is growing very nicely, given what’s going on in the world. It would grow faster if we had more people working on it, but it takes time to scale. We see it as a big opportunity, particularly here in the U.S., but also all over the world. We’ve been very focused on putting in the people resources. This is the interesting thing: technology is very powerful, and there’s going to be more technology that accelerates our wealth business. But this wealth business is people-first, enhanced by technology. It’s not a technology-replaces-people business, and that’s going to continue, in my opinion, for quite some time.
A recurring-revenue, fee-based business is a beautiful thing for any business.
We like fee-based businesses.
4. David Solomon's state of the world
So, David, one of my favorite things about getting to know you over the last several years has been that every time I see you, there’s a “state of the world from David,” and I always find it extremely valuable. We’ve had the midterms. What’s your state of the world right now that’s relevant to investors?
The world is dynamic and complicated, but if you’re an investor, you have to decide what kind of investor you are. I was watching CNBC today. There was a trader and portfolio manager on, and he was talking about what was going to happen in the next 3 months. I can’t look at the world that way. I have a deep-seated belief in the United States and our power to grow our economy, innovate with technology, and build great companies. The markets are a way to participate in that. If you’re going to invest with a 5-, 10-, 20-, or 25-year perspective, I think the world sets up now in almost as exciting a way as it has for quite some time. I’ve been at this for almost 45 years.
5. Why a 5% ten-year doesn't worry him
At the same time, in the short term, there’s heightened geopolitical uncertainty. We’ve moved back into a multipolar world, which makes the world more fragile, and you have to be prepared for the volatility that comes with that. We’re coming off a period of more than 15 years since the financial crisis of artificially low interest rates. I’m not concerned about a 5% 10-year Treasury because I’m 65 years old, and for most of my life, except for the last 15 years, the 10-year Treasury has been higher than 5%. We can have a very vibrant economy with that, but it’s a different operating environment.
6. Elections and the political middle
There’s more volatility and more geopolitical uncertainty, but more technology innovation and, ultimately, productivity and growth around that. We have to be prepared for a political environment that’s going to be very back and forth. Every 2 years, we have a significant election in this country, and in 11 out of the last 12 elections, there’s been a change in one or all 3 of the House, the Senate, or the presidency. Americans aren’t happy with the way that process is working, and they’re saying, “Let’s keep changing. Let’s keep changing. Let’s try to make it different.” As private enterprise, we have a responsibility to try to contribute to that. I think we can make real progress, but we’re going to have to work collectively, public and private sectors, to do that.
It is sad to me for America that each time a party wins a big wave election, they often seem to take that, from my lifetime, as a mandate for their most extreme views instead of a mandate for change to move toward the center. It really feels like the first party that wins and moves to the center might be in power for a really long time.
Well, from your lips to God’s ears. I’d just say this was an interesting election in that context because, when you look at Texas or Kansas, I don’t think there’s been a permanent political shift in Texas or Kansas. It’s the middle moving. I think you’re right that if somebody can figure out how to galvanize the middle and do some constructive things across the aisle, that’s going to be a very powerful thing for that leadership, no matter what side of the aisle it’s on.
I’m going to continue to be optimistic. There are things that I like and things that I don’t like, but I’m an optimist at heart. I have a deep belief in this country: we figure it out, we make mistakes, we evolve, and we get it more right than wrong. Sometimes there are bumps, but we get it more right than wrong. I’d bet heavily on that over the next 25 years.
7. Jensen Huang and NVIDIA
Love it. You and me both. I think we’re going to talk a little bit about AI. I’m curious—you’ve shared some thoughts on Jensen. It’s hard to talk about AI without talking about Jensen. Any thoughts you’d care to share?
Sure. I’d share a couple of things. First of all, I said this on TV last Monday, right when this story broke. I’ve known Jensen for quite some time. The firm has been close to Jensen and to the company.
I have enormous respect for Jensen as a businessman and as a leader. NVIDIA is an incredible company. It’s a great American company. It’s important strategically for the United States, from a security perspective, and as part of America’s leadership in technology broadly.
I saw your letter—Brad’s letter—and I really agree with it. This is a time to go slowly and thoughtfully. One thing I know is that I don’t know all the facts around this. I know very few of them. I’m watching the news and all the discussion, but I actually don’t really know. I think the board and the world should move slowly on what’s best for this company.
As the facts come out, I know Jensen is going to testify. It’s great that he’s going to testify. The facts will allow the right decisions to be made and position the company really well. I’d also say that I know Colette Kress, and I think very highly of her. She’s a very capable executive. I know a lot of other important leaders and executives at NVIDIA, and they’re really talented and really good. They woke up this morning and went to work, making chips and the best technology at what they do. That’s continuing, and it will continue no matter what happens with all this. I have no idea how this will all play out, but in my opinion, this will continue to be a great, important American company.
8. Have we seen the peak in AI?
I agree with all that. Colette is an exceptional CFO and a very capable interim CEO. On AI overall, many AI equities are down 50% or more from their recent highs. It feels like the market is saying we may have seen the peak. What do you think if we look out over the next 2, 3, or 4 years? Have we seen the peak in AI investment here in America and in the world? What do you think about the AI cycle overall?
“Peak” is an interesting word. I think we’re still in the early innings of a long cycle. I’m not smart enough to know. People talk about $6 trillion, $8 trillion, or $10 trillion of investment in AI infrastructure in the U.S. Whenever there’s a technology acceleration, people imagine a growth curve, and then, as more information comes to fruition, they recalibrate that growth curve. Right now, we’re going through a recalibration, but that doesn’t mean people have it exactly right. I think we’ll see significant additional investment and growth. The multiples look more attractive to me than they did 6 months ago.
9. Why the value is in the application layer
A lot of this is going to depend on how chips develop and how the cost of the technology evolves. I think the big value creation over the next 10, 15, or 20 years is going to be in the application layer, not the infrastructure and the initial models. The models and infrastructure are super important, and the hyperscalers have been super important. But if you step back and think about a 10- or 20-year view, the application layer created around this is going to spawn businesses, opportunities, and things we can’t even imagine.
It’s not exactly the same thing, but it’s useful to think about. If you go back to 2000 and the internet acceleration, Apple was a company that made laptop and desktop computers. Nobody had imagined, in that context, the platform of that phone. Uber as a concept and all the things like Uber were imagined 10 years later, when you had a platform that could create that application layer. We’re going to see all that in spades over time. We’re early in this.
The cost of technology generally goes down. Technology always moves faster than enterprises and people, so it’s going to take a while for people and enterprises to adapt. This is going to be a really exciting and interesting ride. I’d call what we’re going through right now a short-term recalibration of the growth curve.
10. Engineers and Goldman's newest partner class
Awesome. You referenced enterprise adoption. I read, or I was told, that this year’s newest Goldman partner class had the highest percentage ever of engineers and technologists. Can you talk about what that means for Goldman and how Goldman is using AI internally, including any tangible benefits?
It’s true that even though we’re a public company, we’ve maintained this partnership culture. We elect partners of the firm who want a one-firm incentive arrangement, and we elected a bigger percentage of engineers than we have before.
One interesting thing is that the firm operates with 13,000 engineers today, in a 45,000-person firm. My guess is we probably won’t have as many engineers 5 years from now, but we will have more super-powerful, very senior engineers. As this AI journey has gone on, we’re finding a need for extraordinary talent, as opposed to base coding talent, because that’s changing with the power of AI.
I think the way to think about AI for Goldman Sachs—serving our clients and running our firm—is twofold. We’re a professional services firm. Fundamentally, we have people who connect with people and use the tools available to them—their own intellect, capital, and technology—to help our clients achieve their goals. That’s our business. These are super-smart, super-motivated people. You give them these tools, and they figure out how to use them to do more, do things better and faster, and make themselves smarter. We’re really good at getting tools to our people and having them use those tools.
11. Reinvesting AI efficiency in people
The more interesting thing in this journey is that we run a huge operation, and this technology is giving us an opportunity for the first time to remake and automate operating processes in a way that generates powerful efficiencies. The first reaction is, “Oh, wow, that can expand margins.” But the real opportunity isn’t margin expansion. It’s taking that efficiency and reinvesting it in people, growth, footprint, and the firm’s capacity to do more with more people and have a bigger footprint. That’s pretty exciting.
We have a very top-down approach to remaking processes. We picked 6 initial processes to remake. One example is client onboarding, KYC, and anti-money-laundering. Those processes are much more automated now, and we’ve added 2 more processes. Basically, we’re taking processes that were highly dependent on people operationally and automating them, making them more efficient. That creates more capacity to invest in and grow the firm, and that’s pretty exciting.
12. AI and pancreatic cancer
Super exciting. Maybe that’s a good segue into the 3 questions we’re going to ask everybody on stage. From your perspective, David, how is AI making the lives of ordinary Americans better today?
I think there are lots of examples, but I’ll give one thing I saw recently that was amazing. Think about pancreatic cancer. It’s a horrible medical condition to be diagnosed with. The Mayo Clinic recently used AI tools to detect 73% of cases up to 3 years in advance, before they were diagnosed.
Wow.
That puts you in a position to treat and address many of those cases. If we could make progress on pancreatic cancer, the benefit to an individual who gets that diagnosis and can see it up to 3 years before it would be diagnosable is very powerful. When I talk about science and medicine, I think the power to help individuals is going to be extraordinary.
13. 10,000 Small Businesses and a $500M AI institute
Nothing’s more valuable than life. What is Goldman Sachs doing to make sure AI is good for America?
One thing Goldman Sachs has invested significantly in over the last 20 years is a platform we call 10,000 Small Businesses. We think small businesses are enormously important to U.S. economic vitality. In fact, year to year, 65% of new jobs in the United States come from small businesses—from entrepreneurs building all different kinds of businesses in the economy.
We’ve had a program and brought about 20,000 businesses through it over the last 20 years, providing a broad business education. We bring them into cohorts. They spend about 6 months coming to the firm and other locations to get a business education on all sorts of things they need to accelerate their growth. We have all sorts of data showing a strong correlation between participating in the program, revenue growth, job growth, and success, and we continue to invest in it.
A few weeks ago, we announced a commitment of $500 million to launch an AI institute for the 20,000 businesses that have been through the program and other businesses. It would do 3 primary things for small businesses, primarily in the United States. First, it would provide AI education on the tools and how to use them to accelerate a business. Second, it would provide research and information to help small businesses understand how AI is being used and how it affects business growth and development. Third, it would accelerate access to capital to allow people to do this.
And so we made that announcement. We’re putting that in place, and we’ve got partnerships with Google, Anthropic, OpenAI, and Meta, where they’re participating with us to drive that small-business investment.
14. AI and the American dream
I actually think that’s awesome. I guess that ties into the third question: How is AI going to help revitalize the American dream for ordinary Americans who feel a little left behind?
Democratization of access. We can talk about small businesses: The cost to start a small business, to have an idea, to go be an entrepreneur, just came down. Access to information to be smarter, access to participate. I just think it’s a democratization process.
We talked about that. You and I would remember 25 years ago talking about this with the internet, that it was democratizing access. This is doing it on steroids, and I think it’s going to be very powerful over the next decade if we get it right. It won’t be a straight line. Things will go wrong, but I’m confident we’ll figure it out. We’ll get that right.
Love it. Thank you, David. Appreciate it.
Appreciate you. Thank you so much.