-+ 0.00%
-+ 0.00%
-+ 0.00%

Nvidia (NVDA.US) on the list tonight: Damo says the business is still strong. What is really tight in the market is the interim CEO's debut and guidance prospects

Zhitongcaijing·11/17/2026 02:17:06
Listen to the news

The Zhitong Finance App learned that Nvidia (NVDA.US) will announce the results for the third fiscal quarter of FY2027 on Wednesday morning Beijing time. The judgment given by Morgan Stanley in its earnings outlook is straightforward: the focus of this earnings report is not on this season's figures — the performance itself has been fully priced — but on Wong In-hoon's first management appearance after leaving office, and a “probably not even give” guidance for the next year. Damo maintains an “overweight” rating, with a target price of $267.

This season's numbers: the part where there is no suspense

Damo expects F3Q27 revenue of $108.1 billion, slightly lower than the market consensus of $108.9 billion; earnings per share of $2.47, which is in line with the consensus. Analysts wrote in the report that channel research showed that “business is still strong” and that even if some projects in Southeast Asia slow down, “they are unable to meet demand.”

What is really worth hearing about acting CEO Colette Kress (Colette Kress) on the call: the attitude towards supporting and investing in the AI ecosystem, whether Nvidia's innovation engine can continue, and the risk assessment of further tightening export controls. Hwang In-hoon did not usually make an opening statement before, but was very active in the Q&A session — Damo suggests that this “general enthusiasm for AI's transformative impact” may be marginally diminished. The report also directed readers to its November 13 topic “NVIDIA: Where the Way Forward,” arguing that the management handover itself is more worth developing than a single quarter's profit and loss.

Disagreements outside of this season: Big Mobile is about $70 billion less than the consensus

Extend the timeline, and differences immediately become apparent. Damo expects revenue of US$623 billion and earnings per share for the 2028 fiscal year. The market consensus is that the difference between US$694.8 billion and US$15.75 — about US$100 billion, is the result of “last week's figure cut”: Damo's estimate for next year's revenue growth rate is 52%, while the company's guidance is about 70% and customer demand is increasing 100%. However, analysts believe that this figure has been “fully risked,” and the research is still pointing towards continued optimism; the real suspense is that the acting CEO is in place and there are many questions, and they are “uncertain whether the company will give guidance next year.”

Focus 1: Demand — the gap between 100% and 70%

Nvidia previously directed revenue growth of around 70% for FY2028, based on a nearly 100% increase in customer demand. However, after the midterm elections, concerns about land, power, and computer room (LPS) restrictions heated up significantly: SoftBank, Nvidia's Ports-Pike project partner, has written off $3.7 billion on related power equipment. Damo's judgment is that the midterm elections threaten 2028 and 2029 more than 2027 — data centers under construction will not be affected by new bans, and the ecosystem will have time to bypass bottlenecks; although the hyperscale cloud vendors it has come into contact with acknowledge the challenges, they are “confident that these challenges can be overcome.”

Financing is the other side of the same issue. Huang Renxun has always supported the region's new cloud and used Nvidia balance sheets to help customers lock in land power resources. Whether these promises will continue after management changes will be the focus. Damo, for example: CoreWeave, Nebius, and IREN's three new cloud companies have a total capital expenditure of about $100 billion in 2027, but the total market capitalization has evaporated by $38 billion in the past month — if the phrase “100% increase in demand” relies in part on financing, then financing will not come yet. In particular, financing for small long-tail customers in the ACIE sector requires continuous monitoring.

Damo's bottom line is: demand from hyperscale cloud vendors is enough to support its forecast. Next year's shipment is about “ten GW higher”, and SpaceX alone plans to deploy 10GW. The report also specifically reminds us of a more fundamental variable: investors' willingness to invest in AI is “a greater construction trajectory risk than any factor Nvidia can control” — in other words, what you should really listen to tonight is whether changes in the financing environment are included in management's wording.

Focus 2: Supply and Guidance - Reiteration is Good

Damo does not expect changes in its annual growth guidelines, but out of prudence, it has lowered its revenue forecast by about 10% due to China's export control risks. On the supply side, memory in particular is “still very tight”, and there are no signs of loosening — “if we had to talk about direction, what we are seeing is a tightening”.

An intriguing detail: Some investors are questioning why Nvidia has been slow to repeat the guidelines after the CEO change. Damo's interpretation is instead a positive sign — according to convention, financial reports are often accompanied by increases, and early forecasting may be “more appropriate to leave the financial reports to be discussed on the same day.”

Focus 3: Rubin climbing the hill - mass shipment in the first quarter

This is Rubin's first high-volume shipping quarter. A quarter ago, Rubin was expected to account for approximately 20% of data center revenue (approximately $20 billion). The climb wasn't all smooth sailing: key testing partner Jingyuan Electronics lowered its 2026 guidance for the second consecutive quarter, partly due to Rubin climbing slower than expected. However, Damo doesn't think this will impact Nvidia's revenue recognition: first, shortening the testing period may explain some of the gaps; second, its January quarter-on-month estimate “should still be conservative,” and Rubin's larger increase will occur in the January and April quarters.

Latest evidence from the demand side

In the attachment to the report, Damo summarized the Q3 earnings report's statement on AI capital expenditure, which can be used as a backdrop for tonight's conference call: Amazon said it “installs an AI accelerator approximately every 14 seconds,” and the recycling cycle is still shortening—its capital expenditure in 2026 will be significantly more in 2027, and “has a clear vision of future returns”; Google said that AI investment “brought about the highest query growth rate in five years” and believes that each dollar of capital expenditure “can support more revenue over time”; Oracle said “the demand for AI is real and huge. The construction capacity and financing are all tasks; the tasks may be difficult, but they can all be solved”; Meta raised the lower limit of the capital expenditure range from 35 billion to 37 billion US dollars; Microsoft said “I would be surprised if the computing power to go online is less than 13 GW by the end of next year”; SpaceX said “the possibility is not zero” to launch more than 20 GW of self-developed computing power next year, and plans to deploy 440,000 Vera Rubin to form the strongest training cluster at the end of the second quarter; on the new cloud side, Nebius said “the customer promised us a longer term, higher price, and more upfront payment,” while CoreWeave said that existing production capacity “sold out to queuing buyers.”

Valuation: Three scenarios and implied probability of options

Damo's three-tier scenario: a bull market of $336 (about 21 times earnings per share in 2027), triggering conditions include supply easing to exceed expectations, split agreements bringing in billions of dollars in net profit revenue, and GPU AI PCs opening up the consumer market; benchmark $267 (about 20 times $13.37, discounting the entire semiconductor sector and a small premium on the market — Damo explained that high market share and high gross profit left the valuation ratio “lacking leverage for near-end expansion”); bear market of $130 ( About 10 times), corresponding to two major debates that are negative at the same time: supply is catching up with demand faster than expected, or competitors entering the market, and the share of customer-developed chips being diverted. According to the implicit volatility of futures, the probability of breaking through the bull market price, above the benchmark price, and falling below the bear market price within a year is about 0.9%, 11.2%, and 24.2%, respectively — the options market's pricing for the downside situation is clearly heavier than the upward trend.

Seller consensus distribution: 77% increase, 22% neutral, 2% reduction. In the risk-return theme ratings given by Damo, the three categories of the new data era, pricing power, and long-term growth were all positive.