
Nvidia’s board announced on 12 November that CEO Jensen Huang has resigned with immediate effect following a board investigation, with CFO Colette Kress appointed interim CEO while the board searches for a permanent successor. The shares fell about -10% in premarket trading on the news, adding to a loss in market value of more than USD600bn this week. We believe the sell-off reflects investors pricing in the loss of the founder who had set Nvidia’s technology direction, closed its largest customer deals and represented it in Washington for 33 years, without a designated successor in place.
Huang’s departure also removes the industry’s most influential voice for looser export controls in Washington. That said, the near-term earnings impact is limited, as Nvidia has already excluded China from its guidance and its FY27-28 revenue rests on Rubin orders and supply already committed, although tighter controls on sales routed through Southeast Asia remain a risk to be monitored. Where Huang mattered more is further out, as Nvidia’s new deals, pricing and next product cycles depended on him personally. As such, we lower our terminal growth rate, as we see revenue growth beyond FY28 more exposed to the leadership transition. Nevertheless, we see limited risk of lasting damage to the franchise, which rests on Nvidia’s supply-chain scale, full-stack platform and developer ecosystem, sustained by a deep engineering bench. Hence, we maintain our BUY rating with a lower TP of USD226 (from USD375), as we cut our fair P/E to 15x from 24x on a weaker long-term growth outlook and a higher regulatory and governance overhang.
Figure 1:
Nvidia’s share price.
Interim leadership provides continuity, but the CEO search will take time
Alongside Huang’s resignation, the board appointed CFO Colette Kress as interim CEO, named its first board chair, separating oversight from management for the first time, and began searching for a permanent CEO among internal and external candidates. As much of Nvidia’s FY27-28 revenue is already committed through customer orders and secured supply, we see limited impact from the change at the top in the near term, and our base case for FY27-28 remains unchanged.
That said, the decisions that drive growth beyond FY28 had rested on Huang personally, and Kress, whose career has been entirely in finance, is unlikely to fill that role. Beyond setting the architecture roadmap and leading Nvidia’s largest sovereign and AI-lab deals, Huang’s long-standing relationships with the heads of its Taiwanese, Korean and Japanese suppliers have secured foundry, memory and packaging capacity in short supply on favourable terms. Moreover, with around 60 executives reporting directly to Huang and no designated second-in-command, we expect leadership uncertainty to persist until a permanent CEO is named, as his successor will need to rebuild a management structure that has been centred on one person.
Meanwhile, we believe Jensen Huang’s resignation would weaken Nvidia’s position in the next rounds of supply negotiations and make customer deals harder to win in the meantime. Hence, we trim our growth and margin assumptions from FY29.
Nvidia loses its most influential voice in Washington
Over the past few years, Huang has been the industry’s most influential voice for looser controls on chip shipments to China. His departure removes that voice just as lawmakers move to tighten controls, with the AI OVERWATCH Act, Chip Security Act and MATCH Act already included in the Senate NDAA manager’s amendment. As such, we see limited prospects for further H200 approvals to China, and a higher risk of chip-tracking requirements on exports to all destinations, which would raise compliance costs well beyond China.
That said, the direct earnings impact is limited. China’s share of data centre revenue has already fallen to below 1% in 2Q FY27 from mid-single digits in 4Q FY24, with H200 the only contribution, and Nvidia has excluded China from its guidance. Hence, tighter controls would mainly cost Nvidia the optionality of further H200 approvals, unless the restrictions extend to overseas customers already contributing to revenue.
Figure 2: China revenue exposure
for the Data Centre segment.
Third-country controls put Nvidia’s hidden China exposure at risk
The leadership change has also brought Nvidia’s regulatory exposure into focus. Megaspeed, a Singapore-based cloud company that bought an estimated USD4.8bn of Nvidia chips, is under US criminal investigation for allegedly diverting chips into China, which it denies, while reports that Nvidia’s general counsel contacted Commerce Department officials on the case have drawn scrutiny from lawmakers.
In the near term, we see two possible outcomes:
1. A fine, which we see as manageable. Seagate’s USD300m settlement in 2023 over shipments to Huawei, the largest standalone penalty in the Commerce Department’s history, was set at more than twice its estimated profit on those sales. A similar penalty on USD4.8bn of sales would amount to less than 2% of our FY28E net profit.
2. Stricter controls on sales through third countries. Former US export control officials are already calling for licence requirements on non-trusted buyers in Southeast Asia and a ban on Chinese companies renting remote access to restricted chips.
We view stricter third-country controls as the bigger threat, as the key regulatory risk lies in how much of Nvidia’s demand ultimately comes from Chinese users buying or renting compute through Southeast Asia, which its disclosures do not show. Before Nvidia moved to reporting revenue by customer headquarters in 3Q FY26, Singapore-billed revenue had risen to 22% of total revenue in 2Q FY26 from 18% in FY25, although Nvidia stated that almost all of it was for US-based customers. As such, we expect the scrutiny following Huang’s departure to accelerate controls on sales through third countries, which could lead to a drop in revenue that Nvidia’s reported numbers do not currently signal.
That said, our FY28 forecast of 45% revenue growth sits about USD100bn below management’s framework of about 70%, a buffer that should absorb a meaningful loss of China-linked demand without cutting our numbers. Nevertheless, any finding of diversion would raise questions over the quality of past growth and weigh on the multiple, which our lower fair P/E partly reflects.
Committed orders shield FY27-28 earnings from the leadership change
Beyond the third-country risk, which our conservative forecast already buffers,, we see limited risk to Nvidia’s earnings over the next two years, as its FY27-28 revenue rests on commitments already made on both sides, although new legislation or an export ban could still prevent some of these orders from being delivered. Every major hyperscaler, AI cloud and system OEM has placed purchase orders for Vera Rubin, which entered full production in August, while Nvidia more than doubled its supply commitments q/q to USD279bn in 2Q FY27 to secure the memory needed to deliver them. As these orders are tied to customers’ own data centre build-outs, we expect them to carry through the leadership change, keeping management’s FY28 framework of about 70% revenue growth well supported.
However, we maintain our more conservative assumption of 45% revenue growth in FY28, treating management’s framework as upside, as memory supply remains tight and leaves little room for slippage in the Rubin ramp. For the same reason, we assume an FY28 gross margin of 70%, below the 75% delivered in 1H FY27, as rising memory costs take a larger share of the Rubin system bill of materials.
Further out, new deals and pricing will test the new leadership
Huang personally led Nvidia’s sovereign and AI-lab agreements, and anchored its pricing with hyperscalers that increasingly have alternatives of their own. Based on Bloomberg Intelligence estimates, custom ASICs are expected to more than double their share of AI accelerator revenue to 20% in 2027 from 9% in 2025, giving hyperscalers more leverage with each generation of in-house silicon. So far, Nvidia has held its pricing by staying a generation ahead and selling the full system. Without Huang, we expect new agreements to take longer to close and hyperscalers to push harder on price, particularly from FY29, when the Feynman architecture is priced.
Figure 3: AI accelerator total
addressable market, GPUs and custom ASICs.
Apple shows a franchise can outlast its founder when the bench is in place
The larger concern for investors is whether Nvidia can sustain its pace of innovation without Jensen Huang. Looking back at 2011, where Apple faced the same concern after Steve Jobs stepped down in August, its shares fell -42.8% from September 2012 to April 2013, as investors questioned its ability to innovate in a post-Jobs era amid intense competition from Samsung, and its forward P/E fell below 10x, pricing Apple as if its best years were behind it.
That said, under Tim Cook, Apple instead became the first US company to reach USD1trn in market value in 2018, as growth shifted from new product categories to monetising its installed base through services, supply-chain scale and in-house Apple Silicon. We believe the franchise held because Jobs left behind an ecosystem that retained customers, a roadmap already in motion and a successor already tested. (We note that Apple may not be the perfect analogy, as it was not trading at a discount when Jobs stepped down, whereas Nvidia was already trading at 1 standard deviation below its 5-year average forward P/E before the recent sell-off.)
In contrast, Intel shows the cost of a finance-led succession at a technology inflection. When CEO Brian Krzanich resigned in June 2018 over a breach of company policy, the board appointed CFO Bob Swan as interim CEO and confirmed him in the role in January 2019. Under Swan, Intel announced a further delay to its 7nm process in July 2020 and lost share to AMD, before the board replaced him in early 2021 with Pat Gelsinger, seeking a leader with a stronger technology background.
We see Nvidia as much closer to Apple in the strength of its franchise. Assuming no unexpected financial exposure, we see limited risk of lasting damage, as its supply-chain scale, full-stack platform and CUDA ecosystem are supported by a deep engineering bench, with around 31,000 of its 42,000 employees in R&D, and its roadmap through Feynman in 2028 already set. However, unlike Apple, Nvidia has no tested successor, and its lead depends on getting each annual architecture right as custom ASICs gain share. As such, we see the board’s choice of permanent CEO as the key variable. A technologist from within Nvidia’s engineering ranks would narrow the leadership discount, while confirming a finance-led CEO would, in our view, raise the risk of an Intel-style slowdown in innovation.
Figure 4: Apple’s share price during
the 2011 CEO transition.
We trim growth, margin and terminal assumptions from FY29
With Huang’s departure, we have revised our assumptions from FY29 onwards. We lower our revenue growth forecasts, as new sovereign and AI-lab agreements are likely to take longer to close, and trim our gross margin on a weaker negotiating position against hyperscalers with in-house silicon. We also raise our operating expense assumptions for higher governance and compliance costs, and lower our terminal growth rate to 2.75% from 3.0% for the greater risk to Nvidia’s innovation cadence beyond its current roadmap. Overall, these revisions lower our FY29E EPS by 3.7% to USD15.07.
Lower fair P/E to 15x; maintain BUY with a TP of USD226
We lower our fair P/E to 15x from 24x, as the leadership change makes Nvidia’s earnings beyond FY28 less certain, and that longer-term growth is what supported its premium multiple. First, without a permanent successor, there is less certainty that Nvidia can sustain its annual innovation cadence beyond its current roadmap through Feynman in 2028. Second, as custom ASICs gain share, a weaker negotiating position puts Nvidia’s pricing power at greater risk from FY29. Lastly, Nvidia now carries a higher regulatory and governance overhang, from tighter export controls and possible chip-tracking requirements to the shareholder lawsuit, which we expect to weigh on the multiple until a permanent CEO is in place. Applying 15x to our FY29E EPS of USD15.07, we derive a TP of USD226.
That said, we maintain our BUY rating, with 42% upside, as earnings through FY28 remain underpinned by committed orders and secured supply. With the shares trading at 12.7x FY28E EPS after a loss of more than USD600bn in market value, we believe the share price already reflects much of the risk of a disorderly transition.
Moving forward, we see two events as key. The first is the 3Q FY27 results on 17 November, the first under the interim leadership, where we expect the numbers to hold up given orders already committed. The focus will instead be on Nvidia’s path without Huang, including its commitment to the roadmap and any clarity on the CEO search. The second is the closed-door congressional briefing on 18 November, which should give a clearer sense of how far Washington intends to tighten export controls, particularly on sales through third countries, where we see the greater risk. With FY27-28 earnings underpinned by committed orders and a franchise built on assets that do not depend on one person, we would view any further weakness around these events as an opportunity to accumulate.
Table 1: Nvidia’s valuation table.
|
FY26A |
FY27E |
FY28E |
FY29E |
|
|
Revenue (in USD millions) |
215,938.00 |
405,921.20 |
588,585.74 |
703,359.96 |
|
Growth %, y/y |
65.50% |
87.98% |
45.00% |
19.50% |
|
Adjusted EPS |
4.36 |
9.06 |
12.53 |
15.07 |
|
Growth %, y/y |
48.64% |
107.69% |
38.32% |
20.24% |
|
P/E at current price |
36.45 |
17.55 |
12.69 |
10.55 |
|
Fair P/E (x) |
15 |
|||
|
Upside Potential |
42% |
|||
|
Target Price |
$226.00 |
|||
|
Source: Bloomberg Finance L.P., iFAST compilations. Data as of 12 November 2026. |
||||
Implications for AI infrastructure
Beyond Nvidia, the leadership uncertainty extends to how AI infrastructure outside the hyperscalers is funded, as Huang has been the architect of Nvidia’s financing push. Under his leadership, Nvidia invested nearly USD50bn in frontier AI labs, agreed to buy CoreWeave’s unsold capacity through 2032 and took on USD108.5bn of maximum guarantee exposure, mostly on OpenAI-linked data centre leases, while its partnerships with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR aim to mobilise more than USD500bn of third-party capital.
On top of that, although signed commitments should hold, we expect new financing to slow while leadership is unsettled, as lenders and partners reassess who stands behind the next deal, with frontier labs and neoclouds that rely on this support the most exposed.
That said, we see AI demand continuing beyond the leadership change, as hyperscalers, which fund their own build-out, accounted for 55% of Nvidia’s data centre revenue in 2Q FY27 and have already placed orders for Rubin. Moreover, partners in the financing platform underwrite each project independently, and as Nvidia systems are widely adopted and transferable across operators, we believe their confidence rests largely on the value of the hardware, which a change of CEO does not alter.
On top of that, while we expect AI demand to hold, the leadership transition could give other fabless players an opening to gain share. AMD and Broadcom are already capturing a growing share of AI compute spending, and customers looking to diversify their suppliers may use this period of uncertainty to direct more of their next-generation orders to alternatives, particularly beyond the Rubin orders already committed.
Table 2: Nvidia’s balance-sheet support for AI demand.
|
Item |
USD bn |
Nature |
|
Maximum guarantee exposure |
108.5 |
Includes up to 105 on PORTS-Pike (OpenAI-linked) data centre leases |
|
Non-marketable equity investments |
99 |
Stakes in AI labs, neoclouds and suppliers, including nearly 50 in frontier AI labs |
|
Future equity commitments |
25 |
Committed, not yet funded |
|
AI cloud agreements |
36 |
Nvidia commits to capacity; AI clouds resell at higher rates with revenue share |
|
Data centre leases for third parties |
20 |
About 15-year leases starting FY28-29, to be reassigned |
|
Third-party financing platform (target) |
500+ |
Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, KKR |
|
Source: Company filings, Morgan Stanley, iFAST compilations. |
||

