AI News

Yiling Strategy Research: Slowing AI Infrastructure Momentum and China Asset Opportunities as the Strong Dollar Peaks

Wall Street InsightsUpdated Nov 16 20:32

AI summary

The top five cloud providers' new lease commitments in 26Q3 fell back from 26Q2, and 26Q1 had already slowed versus 25Q4; the boost from AI infrastructure to US employment and economic growth may slow in the fourth quarter, and the peak of Fed tightening expectations may have passed. The report argues that a weakening dollar will remove the suppression on the non-AI portion of Chinese assets, and recommends focusing on gold, the energy chain, and sectors with strong earnings momentum such as medical services and shipping ports.

Body · AI translation

[Report Introduction] The market's expectations for Fed tightening have risen, premised on the U.S. labor market and economic growth remaining persistently strong. However, judging from the leading indicator of new leasing commitments by major cloud providers, which leads U.S. economic growth and the labor market, the boost from AI infrastructure to U.S. employment and economic growth may both slow in the fourth quarter. Based on the experience of the past year, a strong dollar has clearly suppressed the non-AI portion of Chinese assets; in the future, as the economic vitality of emerging economies improves and the dollar's suppression is lifted, more opportunities will emerge in the non-AI sectors of Chinese assets.

This week, the market's expectations for Fed tightening strengthened somewhat, for two reasons: first, remarks by Fed Governor Waller released a strong hawkish signal, explicitly stating that if U.S. core CPI month-on-month growth is above 0.3%, a 50BP rate hike is an option worth considering; second, the subsequently released October U.S. inflation data exceeded expectations, with core CPI month-on-month growth reaching 0.3%, and according to Bloomberg estimates, year-on-year growth also rose from 1.9% to 2.1%. Structurally, U.S. inflation pressure may come more from the supply side: oil prices remaining high pushed up energy and public transportation prices, and the shortage of electronic hardware transmitted to higher prices for entertainment goods such as audio-visual equipment. Under the combined effect of the above two factors, this week investors have begun to price in a scenario of a 50BP Fed rate hike, while the market believes that through 2027 the Fed will raise rates three more times. It is worth noting that the rise in investors' expectations for Fed tightening is premised on the marked improvement in both the U.S. labor market and economic growth since the second half of 2025, driven by the rapid expansion of AI infrastructure; once this premise changes, the Fed's tightening intensity may weaken.

2 AI's boost to the U.S. economy may slow

Out of consideration for their cash flow and liquidity indicators, U.S. cloud providers, in the process of building data center campuses, are increasingly inclined to move debt off their balance sheets by giving leasing commitments to data center developers. As of the third quarter of 2026, the unstarted leasing commitments of major cloud providers totaled nearly $1.3 trillion, about twice the size of their outstanding corporate bonds. Changes in the scale of new leasing commitments by major cloud providers are an important leading indicator for observing AI's boost to the U.S. economy: changes in the quarterly new leasing commitments of the five major cloud providers lead changes in U.S. construction and manufacturing nonfarm employment by one quarter, lead changes in the boost from information equipment investment to U.S. GDP by one quarter, and lead U.S. Census Bureau data center construction spending by three quarters. It is worth noting that the total new leasing commitments of the five major cloud providers in 26Q3 fell compared with 26Q2, and the new leasing commitments in 26Q1, three quarters earlier, also slowed compared with 25Q4, which may mean that in the fourth quarter, AI infrastructure's boost to U.S. employment, information technology-related investment, and non-residential construction investment may all slow. After Jensen Huang stepped down as NVIDIA CEO this week, there may be a risk that NVIDIA's future support for new cloud providers weakens, thereby dragging down AI infrastructure's boost to the U.S. economy and employment; this would confirm the slowdown trend that began in the third quarter, and the peak of AI investment's boost to the U.S. economy may already be past.

3 The end of the strong dollar will breed hope

AI investment's support for the U.S. economy and employment may be weakening, the premise for dollar tightening expectations is changing, and a weakening dollar is the more likely scenario in the future. Based on the experience of the past year, a strong dollar has clearly suppressed the non-AI portion of Chinese assets. Since 2026, there have been three periods of sustained dollar strengthening: the reversal period after concentrated trading of dollar credit concerns from January to March this year, the period from May to June when U.S.-Iran relations shifted from intense conflict to negotiations, and the period from August to September when a sharp rise in U.S. Treasury yields drove the dollar stronger. In these three periods, the non-AI sectors of A-shares and the Hang Seng Tech Index of Hong Kong stocks both saw clear pullbacks during the sustained rise of the dollar index. It is worth noting that after U.S. Treasury yields peaked in early October, the Citi Emerging Markets Economic Surprise Index has continued to rebound. Although the whole market is still wandering amid uncertainty over the outlook for AI-related industries, we believe investors should not ignore the other side of the matter: more opportunities will emerge: the cycle of U.S. tech—strong dollar—suppressing non-U.S. assets is showing cracks.

4 Before dawn

Current market expectations for dollar tightening may be above their peak; as AI's boost to the U.S. economy weakens, the path toward a weaker dollar and lower capital costs will become clearer. In the future, as the economic vitality of emerging economies improves and the dollar's suppression is lifted, more opportunities will emerge in the non-AI sectors of Chinese assets; investors should remain optimistic and believe that the darkest hour is closest to dawn. Specific allocation recommendations: first, expectations for dollar tightening will gradually recede, and gold is expected to see opportunities; subsequently watch whether the intensity of U.S.-Iran geopolitics picks up, though a medium-term bottom is being established; second, for the energy chain (coal, oil, oil shipping), after the short-term stress test following price declines, it will continue to benefit from global restocking demand for energy; in fact, the energy chain itself is also a resilient asset under fluctuations in capital costs and geopolitical conflicts; third, from an earnings perspective, recommend focusing on: ① continued high prosperity: medical services, shipping ports, and sub-sectors within nonferrous metals and chemicals, etc.; ② turnaround strategy: sub-sectors of textiles and apparel, as well as papermaking, refining and trading, medical devices, express delivery, dairy products, etc.

Risk warning: Domestic economic recovery falls short of expectations; overseas economies decline sharply.

1 The peak of dollar tightening expectations

This week, market expectations for Federal Reserve tightening strengthened somewhat, with the catalysts coming mainly from two aspects. First, Federal Reserve Governor Waller showed a hawkish stance in his speech in London on Tuesday of this week, explicitly stating that "if U.S. core CPI month-on-month growth is above 0.3%, more aggressive monetary policy may be needed, and a 50BP rate hike at the December FOMC meeting is an option worth considering."

Second, the October U.S. CPI data released later that day showed that the United States still faces considerable inflationary pressure, with headline CPI up 0.4% month-on-month and core CPI up 0.3% month-on-month, both exceeding market consensus expectations. Although U.S. CPI data for last October was missing due to the federal government shutdown, referring to Bloomberg's estimates of U.S. price indices, U.S. core CPI year-on-year rose from 1.9% to 2.1% this October, and headline CPI year-on-year also rose above 4%.

Looking at the detailed components, among the main October U.S. CPI components, those with month-on-month growth higher than headline CPI were energy, public transportation, and recreation. This suggests that current U.S. inflationary pressure may come more from the supply side: oil prices remained high in October, pushing up energy prices and public transportation price levels; the rise in recreation component prices may be the result of price pass-through after the earlier shortage of electronic hardware supply.

Under the combined effect of the above two factors, investors' expectations this week that the Federal Reserve will strengthen tightening measures in the future rose significantly: based on investors' expectations for Fed policy implied by federal funds futures, the market has now fully priced in a 25BP Fed rate hike at the December FOMC meeting and has begun pricing in a scenario of a 50BP Fed rate hike; at the same time, the market believes that the Fed will raise rates three more times through 2027.

It is worth noting that the rise in investors' tightening expectations for the Federal Reserve is premised on the marked improvement in the U.S. labor market and economic growth since the second half of 2025, driven by the rapid expansion of AI infrastructure: since the second half of 2025, the rolling 3-month increase in nonfarm payrolls has bottomed out and rebounded, especially with marked improvement in new employment in construction and manufacturing, which directly benefit from AI data center construction; at the same time, driven by AI infrastructure, information technology-related investment has also played an obvious role in driving U.S. economic growth. In 2026Q3, U.S. real GDP quarter-on-quarter annualized rate reached 3.4%, of which information processing equipment investment, intellectual property investment, and nonresidential structures investment together contributed 1.5%. U.S. employment and economic growth remaining at high levels is the premise for the Federal Reserve to continue tightening monetary policy, and once this premise changes, the Fed's monetary policy path will change accordingly.

2 AI's boost to the U.S. economy may slow

The expansion of U.S. AI infrastructure activity is increasingly reflected in cloud vendors' lease commitments. Large cloud vendors such as Google and Amazon have seen rapid growth in capital expenditures related to AI data centers and continued increases in cash flow pressure, so in the process of building data center campuses, cloud vendors are increasingly inclined to achieve off-balance-sheet debt treatment by giving lease commitments to data center developers.

For example, the Hyperion data center construction led by Meta is a typical example of this model: first, Meta and Blue Owl Capital jointly established an SPV entity responsible for project construction; afterward, Meta gave lease commitments to the SPV through contracts and other means, agreeing to continuously lease the data center campus for 20 years after the data center is completed and providing a residual value guarantee; finally, the SPV obtained financing in the credit market through Meta's lease commitments. In this process, because Meta's lease commitments have not yet begun, its on-balance-sheet debt did not increase, while it also obtained future rights to use the data center campus.

For cloud vendors, this model keeps debt off the balance sheet, thereby keeping on-balance-sheet liquidity indicators at a higher level. As a result, cloud vendors' lease commitments for data centers expanded rapidly. Currently, the unstarted lease commitments of major cloud vendors total nearly $1.3 trillion, about twice the size of their outstanding corporate bonds.

Changes in the scale of new lease commitments by large cloud vendors are an important leading indicator for observing AI's boost to the U.S. economy. Since the third quarter of 2025, the scale of new lease commitments by the five major cloud vendors in the current quarter has expanded significantly. During this period, it can be observed that changes in the scale of new lease commitments by the five major cloud vendors in the current quarter are significantly positively correlated with new employment in U.S. construction and manufacturing in the next quarter, as well as with the contribution of U.S. information equipment investment to GDP in the next quarter; for private-sector data center construction spending under the U.S. Census Bureau definition, new lease commitments by cloud vendors lead by roughly three quarters.

It is worth noting that, judging from the recently published third-quarter reports of U.S. cloud vendors, the total new lease commitments of the five major cloud vendors in 26Q3 fell compared with 26Q2, and new lease commitments in 26Q1, three quarters earlier, also slowed compared with 25Q4. This may mean that in the fourth quarter, AI infrastructure's boost to U.S. employment, information technology-related investment, and nonresidential building investment may all slow.

In addition, Jensen Huang's resignation as NVIDIA CEO this week may also have a significant impact on U.S. data center investment activity. Previously, under Huang's leadership, NVIDIA provided substantial financial support and guarantee commitments to emerging cloud providers such as Coreweave and Nebius. In January and March of this year, NVIDIA invested US$2 billion in each of the two companies through equity investments to support their data center construction; in addition, Coreweave previously disclosed in its financial reports that NVIDIA's commitment to purchase its potentially unsold computing power in the future amounted to US$6.3 billion. According to NVIDIA's latest published second-quarter report for fiscal year 2027, NVIDIA's pending cloud service purchase agreements and pending cloud service lease commitments to cloud providers amount to US$56 billion. After Huang steps down as NVIDIA CEO this week, there may be a risk that NVIDIA's future support for emerging cloud providers will weaken, thereby dragging down the boost that AI infrastructure provides to the U.S. economy and employment.

3 The end of a strong dollar will breed hope

U.S. new job creation and economic growth remaining at high levels are the premise for investors' current expectations of dollar tightening, but the supporting role of AI investment in the U.S. economy and employment may be weakening. At present, this premise is changing. The current moment may be when the market's tightening expectations for the Federal Reserve are strongest; in the future, investors' expectations for Fed rate hikes may decline somewhat, and a weakening dollar is the scenario more likely to emerge in the future.

Over the past year, there were three periods in which the dollar continued to strengthen: the reversal period after concentrated trading of dollar credit concerns from January to March this year, the period from May to June when U.S.-Iran relations shifted from intense conflict to negotiations, and the period from August to September when a sharp rise in U.S. Treasury yields drove the dollar higher. In all three periods, a strong dollar clearly suppressed the non-AI portion of Chinese assets; A-share non-AI sectors and the Hong Kong Hang Seng Tech Index both saw notable pullbacks during periods when the dollar index continued to rise. It is worth noting that after U.S. Treasury yields peaked in early October, the Citi Emerging Markets Economic Surprise Index has continued to rebound. After expectations of dollar tightening recede, more opportunities will emerge.

4 Before dawn

The market's current expectations of dollar tightening may be above their peak. As AI's boost to the U.S. economy weakens, the path toward a weaker dollar and lower cost of capital will become clearer. In the future, as the prosperity of emerging economies improves and the dollar's suppression is lifted, more opportunities will emerge in the non-AI areas of Chinese assets. Specific allocation recommendations:

First, expectations of dollar tightening will gradually recede, and gold is expected to see opportunities; subsequently watch whether the intensity of U.S.-Iran geopolitical tensions picks up again, although a medium-term bottom is being established;

Second, for the energy chain (coal, oil, oil shipping), after the short-term stress-test period following price declines, it will continue to benefit from global restocking demand for energy; in fact, the energy chain itself is a resilient asset under fluctuations in the cost of capital and geopolitical conflicts;

Third, from an earnings perspective, we recommend paying attention to: ① continued high prosperity: medical services, shipping ports, and sub-sectors within nonferrous metals and chemicals, etc.; ② turnaround reversal strategy: sub-sectors within textiles and apparel, as well as papermaking, refining and trading, medical devices, express delivery, dairy products, etc.

Risk warnings

Domestic economic recovery falling short of expectations: If subsequent domestic economic data weaken more than expected, it may affect the improvement of the long-term profit trend.

A sharp downturn in overseas economies: If overseas economies decline more than expected, the synchronized recovery in global manufacturing may pause, and demand for physical assets will also slow.

Source of this article: Yiling Strategy Research Risk warnings and disclaimer

Markets carry risks, and investment requires caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial conditions, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their particular circumstances. Investment based on this article is at the user's own risk.

Source

Wall Street Insights