AI News

Goldman Sachs CEO responds to Jensen Huang controversy: The AI ecosystem does not revolve around one person

HuxiuUpdated Nov 11 13:40

AI summary

Goldman Sachs Chairman and CEO David Solomon said at Goldman Sachs' 25th Annual Alternative Investments Summit that the AI ecosystem does not revolve around any one person, and that Nvidia will remain an important company, but that the facts of the Jensen Huang-related matter await investigation by the board and regulators. He also said the market still has strong investment demand for AI infrastructure, and that the current change is a repricing of risk rather than a reversal of demand. The article also mentions that as AI shifts from training to applications, inference costs are drawing attention; AMD, Google TPU, Amazon Trainium and Broadcom custom chips are competing for the market; and IDC data show that in the first half of 2026, Nvidia's share of China data center AI chips fell to 34%, while Huawei's reached 28%.

Why read this

AI score 56

The Goldman Sachs CEO responds after the Jensen Huang controversy on whether the AI ecosystem depends on a single person, and discusses the repricing of AI investment and the chip competitive landscape.

Topics

Tags

Body · AI translation

[[**Jensen Huang admits to marital infidelity and apologizes; Nvidia board launches investigation; market worries his personal controversy could spill over into corporate governance; Goldman Sachs CEO says the AI ecosystem does not revolve around one person, and Nvidia will remain important.**

**Key Points**

**1. Jensen Huang scandal triggers governance scrutiny**

On November 8, Jensen Huang admitted to marital infidelity and apologized; the Nvidia board launched an investigation, and the market worries about management distraction, excessive concentration of leadership, and corporate governance risks.

**2. Goldman Sachs CEO says the AI ecosystem does not depend on a single person**

Solomon said he does not believe the AI ecosystem revolves around one person, and Nvidia will remain an important American company; but the facts are not yet clear, and the board and regulatory investigations are still pending.

**3. AI investment is entering a repricing phase rather than a demand reversal**

**4. Inference demand opens opportunities for non-Nvidia chips**

As AI moves from training to application, enterprises are paying more attention to invocation costs; AMD, Google TPU, Amazon Trainium, and Broadcom custom chips are competing for the market, while Nvidia's software ecosystem remains a barrier.

**5. Changes in China share and the test of returns will determine the long-term landscape**

IDC said that in the first half of 2026, Nvidia's share of China's data center AI chips fell to 34%, while Huawei reached 28%; the long term depends on whether AI application revenue can support massive computing investment. 2026-11-11 11:43 Jensen Huang is mired in controversy, and the Goldman Sachs CEO's latest response: "The AI ecosystem does not revolve around one person" Cold Eye Quick View

Jensen Huang has recently become embroiled in an infidelity scandal. On November 8, Jensen Huang admitted to marital infidelity and apologized. This also triggered widespread attention across the chip market. Regarding his personal relationships, as well as the potential intelligence and chip export issues mentioned in media reports, the Nvidia board launched an investigation.

Market concerns quickly extended from Jensen Huang personally to Nvidia's corporate governance. As founder and long-time leader, Jensen Huang is deeply involved in technology roadmaps, customer cooperation, and external communications. If the incident continues to escalate, will Nvidia face management distraction, excessive concentration of leadership, or even a reassessment of risk by the capital market? For the entire AI industry, Nvidia is not only a core computing power supplier but also intertwined with many companies through investments, financing support, and other transactions.

Against this backdrop, on November 9, Goldman Sachs Chairman and CEO David Solomon was interviewed by CNBC reporter Leslie Picker at Goldman Sachs' 25th Annual Alternative Investments Summit. The conversation began with the repricing of risks in the AI market and extended to data center construction, Nvidia and Jensen Huang, the Anthropic IPO, the private capital market, and Goldman Sachs' asset management business.

The following is a translated transcript:

Leslie Picker: When we sat down in June, you described the market at that time, especially AI financing and the capital market, as having "more greed than fear." Prices have fluctuated sharply over the past few weeks. Has that relationship now reversed?

David Solomon: Investors are always assessing the environment around them. I remember saying at the time that there would certainly be constant adjustments in this process. I specifically mentioned that the growth and investment in U.S. AI infrastructure cannot be a straight line. As people further understand the facts, the political environment, market conditions, demand, and how the application layer develops and what signals emerge, the market will reprice, with ups and downs. What we are seeing now is part of that. But I remain deeply convinced that AI can improve productivity across the entire economy in the long run. The foundation supporting that judgment has not changed. Investors constantly reassessing is precisely a strength of the market.

Leslie Picker: This is the first time we have sat down since the midterm elections. During the election period, there were many anti-AI voices, and some advocated pausing data center construction. Will this change the cost of financing or the overall calculation of continued investment in data centers? Could it pose a major risk to the development path you just mentioned?

David Solomon: There was indeed a lot of related discussion before the election. Looking ahead, it is necessary to distinguish political noise from the policies and actions that will ultimately be implemented. First, not every state in the country opposes data centers. Some states have strong opposition, while other places understand that infrastructure investment can bring economic benefits and understand its importance to the economy.

Looking back about 50 years, the United States made some policy decisions out of concern about nuclear power. Standing today, there are probably some decisions we would wish had been handled differently. The computing infrastructure being discussed today will provide computing power for American enterprises, and as the economy develops, it will become increasingly important. We must consider how to build it in the right places, in the right way, and with the necessary safeguards. But of course we do not want this infrastructure to move outside the United States, because computing demand objectively exists: if the United States does not build it, projects may naturally flow elsewhere.

Things are still developing and are not as black-and-white as the political narrative presents. What happens next depends largely on how states and local communities make decisions, and many politicians have already begun discussing it. I remain optimistic that U.S. data centers and computing infrastructure will receive substantial investment. This is what we need. If these constructions are truly blocked, looking back in ten or twenty years, we may not think that was a wise strategic choice made by the United States.

Leslie Picker: What happened today really illustrates the geopolitical and competitive environment in which AI operates. I have to ask about Nvidia's central position in the entire AI system. Through investments, backstop arrangements, and securitization deals, Nvidia is increasingly intertwined with the financing of the ecosystem, including deals in which Goldman Sachs is involved. Given the news and questions surrounding Jensen Huang over the past 24 hours, to what extent do you think this ecosystem revolves around one individual?

David Solomon: I don't think this ecosystem revolves around one individual. I know Jensen Huang, and I admire him. I think Nvidia is an outstanding American company. No matter how this ultimately develops, it will remain an outstanding and important American company. I am also following the news, but I don't have the facts or inside information. The board is investigating, and regulators will review it as well. As more information emerges, we will know what actually happened, and investors will be better able to judge how the market views the company. But no matter what happens, Nvidia will remain a very large and important company in this ecosystem and in the United States.

Leslie Picker: So, as this continues to develop, you are not concerned about the risk of management distraction at Nvidia, or the risk of excessive concentration?

David Solomon: When news like this comes up, I don't speculate about what might happen next, because I don't have enough information. But this is a large company, and an important one. In my view, that will continue. Of course, we must keep watching and figure out what actually happened and what in fact did not happen.

Leslie Picker: Anthropic is also a very large and very important company. I know it is a Goldman Sachs client. It has been reported that Goldman Sachs is leading its IPO, and that, as required by the U.S. Securities and Exchange Commission, the roadshow could begin as early as tomorrow. There are also reports that it is scaling back the offering size and lowering its target valuation. I know you can't discuss specifics, but how would you describe the current market's receptiveness to AI and AI financing?

David Solomon: There is still very strong investment demand for AI infrastructure and the technologies that support AI development. As we discussed at the outset, the market's assessment of risk is now slightly different from a few months ago, but that has not changed fundamental interest in the technology and infrastructure. The market is still open, and deals are still getting done. As for how this IPO will develop if it proceeds, we will see when we get there. I can't comment further on specific deals, but overall the market environment remains quite favorable.

All judgments are relative. People will compare now with two months ago, but today's market is still good. These are important infrastructure assets, and there will still be a lot of capital flowing in. Look at Goldman Sachs' business activity over the past few weeks; we have been very active in the capital markets. Based on what we are seeing today, I expect that to continue over the next two months.

Leslie Picker: We are here at Goldman Sachs' Alternative Investments Summit, which is now in its 25th year. When it comes to the unease and concerns around AI in the broader world, do you see different behavior in the public and private markets?

David Solomon: Capital flows between the public and private markets are quite efficient. One advantage of the private markets is that not all assets have to be revalued every day. So when there is volatility and repricing, the public markets react much faster than the private markets.

This is the 25th Alternative Investments Summit. Goldman Sachs has a very large alternatives platform, with more than $700 billion in alternative assets under management and a broad client base. I think capital raising and formation in the private markets are still at an early stage. Compared with the public equity market, the fixed income market, and the government bond market, the private capital market is still relatively small. Some people mistake private equity for the entire private market, but private debt, real estate, and other areas are also huge, and capital is still being raised and formed. This trend will continue, but it won't be a straight line; there will certainly be ups and downs along the way.

The summit brings together hundreds of investors who play an important role in allocating private capital. What we are seeing is still strong growth. Goldman Sachs' alternatives fundraising has been very strong this year, and our credit business is also strong. There is a lot of noise around private credit, but institutional investors are also seeing wider spreads. Have default rates risen somewhat? Yes, but they remain below historical averages. So interest in private investing still has good momentum. I think this is a long-term trend and will not stop anytime soon.

Leslie Picker: Over the past few years, Goldman Sachs has made several acquisitions in asset management, including NEOS, LCN, Innovator Capital Management, and Industry Ventures. On your most recent earnings call, you mentioned that there are still some gaps across the platform. Where exactly are they? Will continued acquisitions be the solution?

David Solomon: Our asset and wealth management platform is already quite comprehensive. We manage more than $4 trillion in assets, have an excellent ultra-high-net-worth client business, and strong capabilities in developing asset management investment products. The purpose of these acquisitions is to accelerate growth in certain strategic areas. Innovator and NEOS are mainly focused on actively managed ETFs. Previously we were not even in the top 50 in this area; now, depending on the measure, we have reached sixth to eighth place. Given the development of active ETFs and the positioning of the Goldman Sachs Asset Management brand, this is a very good strategic move...

Over the past few years, Nvidia has become almost the most direct beneficiary of the AI boom. Large model companies need training compute, cloud computing giants need to build data centers, and startups need to rent GPUs. With chip performance, the CUDA software ecosystem, and a full set of server and networking systems, Nvidia has built advantages that competitors find difficult to replicate quickly. To some extent, investors betting on Nvidia are also betting that global demand for AI compute will continue to grow.

But the competitive logic of the chip market has already begun to change.

As AI moves from model training to large-scale application, more and more companies are no longer concerned only with whether they can obtain the most powerful GPUs, but with how much each model invocation actually costs. Training requires highly flexible general-purpose compute, while inference may be better suited to chips optimized for specific tasks. This gives other vendors new opportunities.

AMD is competing for the AI accelerator market, Google continues to expand the application scope of its in-house TPUs, and Amazon is also advancing Trainium. At the same time, chip design companies such as Broadcom are helping large technology companies develop custom AI chips. On November 9, 2026, Google launched two new TPU products, further expanding its chip lineup for enterprise customers. For these tech giants, in-house chips mean not only lower costs but also not having to leave future compute supply entirely to a single external vendor.

This change does not mean Nvidia's advantage is about to disappear. Its mature software ecosystem, developer habits, and ability to deploy large-scale clusters remain important barriers. But the market no longer has only one choice. In the future, different AI tasks may be handled by chips with different architectures, and chip competition will shift from simply comparing computing performance to comparing the cost, energy consumption, software compatibility, and delivery capability of entire systems.

More noteworthy is that behind the chip competition there is also a highly concentrated global supply chain. Whether it is Nvidia's GPUs or tech giants' custom chips, none can do without advanced processes, advanced packaging, and high-bandwidth memory. The number of chip design companies can increase, but the companies truly capable of undertaking large-scale advanced manufacturing tasks remain limited. Even if more competitors emerge in the AI chip market in the future, the industrial chain may not necessarily become sufficiently diversified as a result.

The Chinese market offers another perspective. U.S. export controls are changing the global AI chip competitive landscape and also pushing Chinese domestic chip companies to accelerate substitution. In October, Reuters cited IDC data reporting that in the first half of 2026, Nvidia's share of China's data center AI chip market fell to 34%, while Huawei reached 28%. This does not mean the two companies are already evenly matched in the global market, but it shows that the same AI compute competition may lead to completely different outcomes in different regions.

If companies can continue to earn returns from AI applications, chip demand may gradually expand from concentrated purchases by a few large model companies to a broader range of industries. But if application revenue growth cannot keep up with infrastructure investment and data center construction slows, the impact will not be limited to Nvidia; it may also affect chip design, wafer manufacturing, memory, packaging, servers, and other links.

This is also another layer of thinking that Goldman Sachs CEO's interview leaves for the market. A huge industry certainly should not revolve around a single person, but it also should not pin all growth expectations on a few companies continuously increasing capital expenditure.

Source material:

Has published 33 articles on Huxiu. This content may not be reproduced without permission. For authorization, please contact hezuo@huxiu.com. Everyone who is changing and wants to change the world is on the Huxiu APP[ ]

Source

Huxiu