The Nvidia crisis and financing risks are at their peak! Barclays, however, is still optimistic about the long-term market for AI trading

The Zhitong Finance App learned that Julian Lafague, chief market strategist at Barclays Private Bank, said that although the NVDA.US (NVDA.US) crisis and financing risks are testing the resilience of artificial intelligence (AI) transactions, investors should continue to participate in the AI boom.

Lafarge introduced Barclays's upcoming 2027 outlook report called “Until the Music Stops” (Until the Music Stops) in advance. He said that the agency still believes that there are good reasons to continue holding relevant investments. In an interview, Lafarge said, “For us, this is not a bubble, but a technological revolution that will affect every industry. But we need to be aware that this process can't be done overnight.”

The Nvidia storm didn't break the AI deal

Nvidia's stock price fell more than 7% before the US stock market on Thursday. Three days after posting an article admitting her affair on November 8, Hwang In-hoon was revealed to have a second extramarital relationship. The woman was Nvidia's vice president who reported her work directly to her. Meanwhile, according to media reports, citing people familiar with the matter, Nvidia's board of directors is intensively preparing to dismiss Wong In-hoon. The board of directors will be held on November 12 as soon as possible to decide where to stay, and has cooperated with external headhunters to find a successor and consider appointing an interim CEO to transition first.

Some commentators believe that Nvidia may be able to get through this turmoil, but the process will be extremely difficult, and it is quite likely that Hwang In-hoon will be removed from the position of CEO. According to data from the forecast market Polymarket, the probability of the market betting that Hwang In-hoon will leave his job or be removed has soared. Currently, the probability of Hwang In-hoon leaving his job before November 30, 2026 has risen to 86%, 91% before December 31, 2026, further rising to 92% before March 31, 2027, and 93% before December 31, 2027.

As Nvidia is about to release financial results for the third fiscal quarter next week, “key person risk” worries investors. Fiona Sincota, a senior analyst at Jiasheng Group, said that until the board of directors investigation results are announced, relevant uncertainty may continue to weigh on Nvidia's stock price, especially in the context of Nvidia's upcoming earnings report later this month.

The risk of Nvidia's current governance crisis to the entire AI industry is that Nvidia has become the core node of the global AI industry's capital, technology, and supply chain network. In addition to supplying chips to AI companies around the world, Nvidia has also participated in the upstream and downstream development of the industrial chain through investment. Its investments include companies such as OpenAI, Anthropic, XAI, CoreWeave, and Nebius. In August of this year, Nvidia also signed preliminary agreements with organizations such as Apollo Global Management, BlackRock, Brookfield, Goldman Sachs, and KKR to raise more than 500 billion US dollars in capital in the future. Hwang In-hoon said earlier that Nvidia may support up to a quarter of this funding.

This means that once Nvidia has a serious leadership crisis, the impact may not only be limited to the company itself, but may also affect many companies that rely on its chips, investments, and financial support. Well-known technology stock analyst Dan Ives believes that if Huang Renxun leaves the CEO position, Nvidia's stock price may drop 25% to 30%, and the corresponding stock price may even fall to around $122. Meanwhile, Jay Goldberg, an analyst who has been cautious about Nvidia for a long time and is one of the few Wall Street analysts to give a “sell” rating, said that if current pressure eventually causes Wong In-hoon to leave office, it could be a serious mistake for Nvidia and all parties involved. As the core target in the AI industry chain, if Nvidia's stock price falls sharply, it will undoubtedly spread to the wider AI sector, once again overshadowing the prospects for AI transactions.

Referring to Nvidia, Lafarge said the company has always been known as the “central bank for AI transactions,” but he believes that even if the company may change its CEO, that doesn't mean all deals have to be re-examined and renegotiated.

Lafarge acknowledged that national security considerations may prompt the parties concerned to take stricter measures to restrict the export of Nvidia or other American-made chips overseas, which may have an impact on the entire industry. However, he added that the direction of development in the next two to three years is very clear. Competition and innovation in the US and other regions will further intensify, drive the continuous emergence of new technologies, and ultimately weaken the impact of any short-term disturbances.

Lafarge compared this situation faced by Nvidia's leadership to the leadership transitions between Microsoft (MSFT.US), Alphabet (GOOGL.US), and Amazon (AMZN.US). He believes that the transition of these companies from charismatic leaders to successors has not diminished their corporate value.

Financing risks are rising due to concerns about AI returns and uncertainty about the Federal Reserve's policies

Lafarge said that Barclays has identified 12 factors that may disrupt the market, with interest rates and the financing environment posing the most significant risk. Still, he said, “We want to continue to hold our investments and maintain our investment exposure until there are signs that music starts to slow down.”

Lafarge believes that investors need to recognize that technological progress does not mean that all companies can get the same returns, and that the competitive landscape and profitability within the AI industry chain may be clearly divided. Looking at broader AI investment transactions, Lafarge believes that the market needs to readjust expectations about the speed at which computing power production capacity will be launched. He pointed out that the trillions of dollars of funds already promised cannot be converted into actual delivered computing power infrastructure in a short period of time.

In fact, AI infrastructure construction itself also has a long delivery cycle. Even if the construction of a data center is completed, it does not mean that it can immediately be put into commercial operation. Grid access, power supply, equipment installation, system commissioning, and customer deployment may all cause delays. Rating agency Moody's recently pointed out that there is a clear time gap between data center construction and actual stable cash flow, and insufficient power supply and transmission infrastructure is becoming an important constraint on the expansion of AI infrastructure.

At the same time, as the scale of investment in AI infrastructure continues to expand, the focus of market attention is gradually shifting from “whether demand for computing power actually exists” to “whether huge capital expenditure can obtain sufficient returns.” Although demand for computing power from cloud computing companies and AI model developers is still strong, data centers, GPUs, network equipment, and power facilities all require huge investment, and the commercialization process of AI applications on the enterprise side and consumer side is not completely synchronized. For investors, the long-term development prospects of AI technology and whether related companies can achieve return on investment within a reasonable period of time are two different issues. However, due to concerns about whether huge AI investments can be reasonably rewarded, investors are asking related companies to provide more premiums in terms of financing.

In addition to this, uncertainty about the Fed's policy outlook has also increased financing risks. AI data center construction has the characteristics of intensive capital, long investment cycle, and huge upfront investment. Companies not only need to pay high costs for GPUs, servers, and network equipment, but also bear the cost of land, electricity, cooling systems, and long-term financing. Therefore, changes in interest rate levels and credit market conditions will directly affect the project's capital cost, payback cycle, and final return.

After the release of the higher-than-expected CPI data for October this week, the market has increased its bet that the Federal Reserve will further tighten monetary policy in the future. Higher interest rates have a double impact. On the one hand, enterprises need to pay more interest, and the cost of debt financing to build data centers and computing power facilities will increase; on the other hand, rising interest rates will increase the discount rate of future cash flow and reduce the valuation of unprofitable AI businesses and related infrastructure assets. For enterprises that rely on long-term computing power leasing contracts, fixed price service agreements, and debt financing to achieve expansion, if financing costs rise and customer payment prices do not increase simultaneously, project profit margins may narrow significantly.

In addition, rising financing costs may also have a ripple effect: some emerging cloud service providers may not be able to obtain refinancing at an acceptable interest rate and be forced to delay data center construction or cut GPU procurement; while large cloud service providers are more capable of financing, they may also re-evaluate the pace of capital expenditure and prioritize capital investment in projects with higher returns and a clearer commercialization path. Once investment in the industry as a whole slows down, expectations for order growth from chip vendors, server manufacturers, data center operators, and power equipment companies may also be affected. All of these factors may cast a shadow over the future of AI trading.

However, large tech giants still have strong operating cash flow, high profit margins, and relatively stable balance sheets, which can withstand the impact of rising financing costs to a certain extent. J.P. Morgan believes that the demand for AI computing power is still huge, and as intelligent workflows and inference tasks continue to expand, the demand for computing power is expected to continue to grow in the future. The tightening of financing conditions may further change the pace of investment in the industry and the competitive landscape of enterprises, rather than directly eliminating the long-term demand for AI infrastructure construction.