
Key Points
- Nvidia’s near-term earnings are intact despite the CEO exit, so we reiterate BUY with a higher target price of USD237, implying 52% upside.
- Revenue rose 103% y/y to USD115.8bn, with Vera Rubin contributing nearly USD20bn in its first quarter of shipments.
- Management guided 4Q revenue to USD130bn and raised FY28 growth to about 75%, though we keep our forecast at 45%.
- The open CEO search and intensifying regulatory scrutiny, including an SEC investigation, weigh on the valuation multiple rather than on near-term earnings.
- At 11.8x FY28E EPS, much of this uncertainty is already priced in, so we see the risk-reward as favourable.
On 17 November 2026, in its first earnings report under interim CEO Colette Kress, Nvidia delivered another beat on both the top and bottom lines. Revenue rose 103% y/y and 20% q/q to USD115.8bn, 7.3% above guidance, led by the fastest product ramp in the company's history, with Vera Rubin contributing nearly USD20bn in its first quarter of shipments. Management guided 4Q revenue to USD130bn, above consensus of about USD125bn, and raised its FY28 growth outlook to about 75% from 70%.
Following Jensen Huang's resignation, the quarter shows that Nvidia's near-term business carried through the leadership change, with customers, suppliers and the product roadmap all holding, keeping earnings through FY28 on track. That said, the results do not address the longer-term questions over who will permanently lead Nvidia and how far regulatory scrutiny will extend, which has intensified with the SEC opening a civil investigation and the Treasury Secretary demanding answers from the board. As such, we raise our earnings forecasts on the stronger quarter and faster buybacks, and reiterate our BUY rating with a higher TP of USD237 (up from USD226).
Figure 1: Nvidia's share price.
Another beat and raise, with growth balanced across both customer groups
Data centre revenue rose 110% y/y and 21% q/q to USD107.6bn. Hyperscale revenue reaccelerated to 23% q/q from 13% in 2Q, reaching USD60.1bn as Rubin supply increased, while AI Clouds, Industrial and Enterprise (ACIE) grew 18% q/q to USD47.5bn, nearly triple its level a year ago, as Rubin shipped to leading neoclouds and frontier labs began building their own data centres. Encouragingly, hyperscalers brought Rubin into production in about a month, half the time Grace Blackwell took, which we believe removes the risk of a revenue gap between the two generations and reduces the risk of customers diversifying away from Nvidia’s GPUs. Meanwhile, edge computing grew 14% q/q to USD8.2bn on Blackwell workstations and physical AI applications., while networking revenue exceeded USD21bn, up about 20% q/q, reflecting rising networking attach as rack-scale systems ramp.
Non-GAAP gross margin came in at 74.9%, above the 74% guide on a stronger Blackwell Ultra mix, and edged down only 0.1ppt q/q as Rubin began to ramp. Similarly, operating expenses rose 7% q/q to USD8.8bn, slightly below guidance, lifting operating margin to 67.3% from 63.3% a year ago. As a result, non-GAAP EPS more than doubled y/y to USD2.68.
Free cash flow rose to USD43.8bn from USD21.3bn in 2Q, partly helped by the timing of tax payments, while receivable days rose further to 64 from 60. However, we see the rise in receivables as largely timing-related, with limited credit risk. First, it reflects Rubin shipments concentrated late in the quarter, which should be collected as the ramp matures. Second, the extended payment terms apply only to investment-grade customers, which largely excludes the financing-backed neoclouds at the centre of market concerns. Meanwhile, inventory rose to USD39.7bn from USD31.6bn to support the Rubin ramp, as demand continues to outpace supply. Notably, even with this working capital build, free cash flow comfortably covered a record USD28.1bn returned to shareholders, and Nvidia plans to use its remaining USD227.3bn buyback authorisation through FY28, providing meaningful support for the share price while leadership questions remain open.
Figure 2:
Nvidia's 3Q FY27 earnings summary.
Source: Company, Claude, iFAST compilations.
Data as of 26 October 2026.
Guidance raised again, with FY28 growth now at about 75%
Management guided 4Q revenue to USD130bn, plus or minus 3%, implying about 12% q/q growth including an extra week, with no data centre compute revenue from China assumed. More importantly, FY28 growth was raised to about 75% from 70%, off a higher FY27 base, as demand continues to grow and partners add supply, while CPU and enterprise revenue are both expected to more than double in FY28.
However, following the recent resignation of former CEO Jensen Huang, Nvidia's near-term growth depends more on how much it can deliver than on demand, so we maintain our FY28 revenue growth forecast at 45% and treat management's outlook as upside, for three reasons. First, new US export controls were released in the week of the results, and management is still assessing their impact. Second, supply is now allocated according to whether customers have their land, power, capital and offtake in place, which could slow deliveries to smaller customers. Lastly, memory remains tight, leaving little room for slippage in the Rubin ramp.
Figure 3:
Management guidance.
Source: Company, Claude, iFAST compilations. Data as of 26 October 2026.
Gross margin bottoms in 4Q before improving on new pricing in FY28
Moving forward, management expects gross margin to bottom at 71-72% in 4Q before improving to 72-73% in FY28. We believe the 4Q trough is largely a timing effect, as Rubin, whose bill of materials is heavily weighted towards memory, makes up a larger share of the mix just as higher-cost memory flows through from inventory, while many of these systems still ship against orders taken at earlier prices. Nevertheless, the dip does little to slow earnings growth, as 4Q guidance still implies about USD93bn of gross profit, up around 7% q/q from USD86.7bn in 3Q.
We continue to see the company updating its disclosure on its commitments. As of end-October, Nvidia has strengthened its supply position, with supply commitments rising to USD390bn as of end-October from USD279bn, mainly multi-year memory procurement, securing the supply needed for the Rubin ramp and supporting cost visibility into FY28. As such, we expect margins to recover in FY28 as new pricing takes effect. That said, FY28 will also see the launch of Rubin Ultra, and as this quarter’s slight dip during the early Rubin ramp shows, new generations tend to weigh on margins as they ramp. Hence, our FY28 gross margin forecast remains below management’s range.
Figure 4: Nvidia's gross margin and management
guidance.
Competition is narrowing the gap, making roadmap execution more critical
In 3Q, Rubin contributed nearly USD20bn of revenue in the early stage of its ramp, delivering up to 10x more tokens per watt than GB300 in MLPerf inference tests, while management reaffirmed the launch of Rubin Ultra next year and Feynman in 2028. Notably, OpenAI and Meta have adopted the Vera CPU even as they diversify their accelerator supply, suggesting Nvidia remains able to win business from customers spreading their compute across suppliers. As such, we believe Nvidia continues to execute well on its roadmap and retains a clear technology lead.
That said, we note that competitors are gradually closing the performance gap, with AMD’s MI455X estimated to deliver higher inference performance per chip than Rubin, while frontier labs have been diversifying their compute, with Nvidia accounting for less than a third of their announced capacity for deployment from 2026. More importantly, AI is lowering the cost of switching away from CUDA, with Anthropic using its own model to bring up one of its models on AMD GPUs in about three days, against six to 12 months previously. Hence, we believe Nvidia’s product roadmap has become more critical than in previous cycles, as potentially lower switching costs allow customers to move more readily even on a narrower performance gap. Moving forward, we see the launch of Rubin Ultra next year as the next test of whether Nvidia can sustain its lead, with timely delivery and its performance as the key factors to watch.
Figure 5:
Anthropic cuts the time to run models on AMD GPUs from months to days.
Source: Anthropic, iFAST compilations.
The results reaffirm strong AI demand across the supply chain
For the broader industry, the results reaffirm that AI demand remains strong. Management cited rising GPU rental prices across generations, with Hopper re-contracted at up to twice its original price and B300 rates up more than 60% since launch, while Nvidia's compute remains fully utilised at every hyperscaler and neocloud it serves. In our view, this shows that older systems keep earning long after newer ones arrive, supporting customers' returns and their willingness to keep investing in each new generation.
These supply commitments, alongside new capacity agreements across substrates and advanced packaging, also extend earnings visibility for memory, foundry and packaging suppliers. However, these commitments raise the stakes if data centre builds are delayed or financing tightens, as Nvidia and its suppliers would be left carrying capacity built ahead of deployment.
Nvidia turns more selective on financing, lowering its exposure to weaker-credit customers
Notably, Nvidia has become more disciplined in how it supports neoclouds and frontier labs, with its neocloud agreements falling to USD32bn from USD36bn and now taking effect only once neoclouds have secured their own financing and deployed the related capacity, which we see positively. Management also said it is rigorously evaluating future investments, commitments and guarantees, and will prioritise land, power and shell arrangements at sites where demand is visible. This reduces Nvidia's exposure to customers whose balance sheets are weaker than those of the hyperscalers, which fund their own build-outs and still accounted for 56% of data centre revenue in 3Q.
That said, Nvidia's broader commitments are still growing. Maximum guarantee exposure rose to USD118bn from USD108.5bn, although each guarantee only takes effect as its related data centre is completed and no payments have been made to date, while cloud and data centre lease commitments rose to USD102bn from USD74bn. Nvidia also recognised a USD1.8bn impairment on its private investments. We see the risk from these commitments in two places: the guarantees would only lead to losses if counterparties fail to meet their obligations, while the leases commit Nvidia to paying for capacity regardless of demand. With compute fully utilised across its customers, we see both risks as low for now, and none of these change our earnings forecasts.
Figure 6:
Nvidia's commitments and guarantees.
Source: Nvidia 3Q FY27 CFO commentary, iFAST compilations. Data as of 26 October 2026.
The board does not intend to reinstate Huang, leaving the CEO search open
The elephant in the room remains unresolved. On the call, the board stated that it unanimously accepted Jensen Huang's resignation and does not intend to reinstate him as CEO. It has retained a search firm and is considering both internal and external candidates, noting that the next CEO will need to navigate both advanced silicon technology and national security dynamics, and has set up a national security and export control oversight committee to strengthen its engagement with the US government.
Pressure on the board is coming from both directions. Huang has publicly said he would return if the board decided it was in the company’s best interest, and some major shareholders have urged a measured process that leaves that possibility open, while the Treasury Secretary has called on the board to appoint an outsider. If Huang were to return, we would see it as positive for Nvidia, given his role in setting the roadmap and in its customer and supplier relationships despite Washington currently favouring an outsider, in which his return would likely bring closer regulatory scrutiny. That said, the board’s emphasis on national security, together with its new oversight committee, suggests Washington’s views will weigh heavily on the outcome. As such, with congressional scrutiny ongoing, we expect leadership uncertainty to persist until a permanent CEO is named.
Regulatory scrutiny intensifies, keeping the overhang in place
Alongside the results, regulatory pressure on Nvidia has increased. The SEC has opened a civil investigation that includes whether Nvidia’s disclosures misled investors, while the Treasury Secretary has questioned whether sales through Southeast Asian intermediaries inflated its revenue, asking the board and its outside counsel to respond within ten days. On top of that, Huang will appear before three House committees in closed-door sessions on 1 December, followed by Nvidia’s general counsel on 2 December, while the Senate Banking Committee chair has asked Huang to testify voluntarily on 3 December.
We see the financial penalty from a disclosure case as manageable. The greater risk lies in what the board’s response to Treasury concludes about Nvidia’s sales through Singapore and other third countries, as any finding of diversion would raise questions over the quality of past growth and add pressure for stricter controls on those routes. Notably, we see the general counsel’s session as the more relevant one for the company, as it centres on Nvidia’s own compliance processes, while Huang’s appearance mainly concerns him personally. Hence, we expect the regulatory overhang to remain until these reviews are complete, which could pressure valuations even as earnings continue to grow.
Stronger 3Q results and buybacks lift our FY29E EPS
The stronger-than-expected 3Q results and 4Q revenue guidance, together with a faster pace of buybacks, lift our FY27E starting point. We raise our FY27E revenue by 4.4% to USD423.7bn and FY27E EPS by 5.5% to USD9.55, and now assume the share count falls 3% in FY28 as Nvidia uses its buyback authorisation. With our FY28 growth assumption kept at 45%, which remains below management’s outlook, and our margin and leadership-related assumptions unchanged, these revisions raise our FY29E EPS by 4.9% to USD15.80.
Reiterate BUY with a higher TP of USD237
In short, the quarter shows that Nvidia’s business has so far been unaffected by the turmoil at the top. Demand continues to run ahead of supply, the Rubin ramp is on track, and with supply secured and guidance raised again, earnings through FY28 are well underpinned. In fact, management kept the call firmly on business as usual, while the question of who will lead Nvidia next remains open. As such, the debate has moved beyond FY28, to the choice of the next CEO, the outcome of the regulatory reviews, and whether the roadmap can stay ahead of competitors. We believe these questions bear more on the multiple than on near-term earnings, and at 11.8x our FY28E EPS, we see much of this uncertainty already reflected in the share price. Applying our fair P/E of 15x to our FY29E EPS of USD15.80, we derive a TP of USD237, implying 52% upside, and reiterate our BUY rating.
Moving forward, we see three events as key:
- The outcome of the House closed-door briefing on export controls;
- Further detail on the new export rules; and
- Progress on the CEO search.
Table 1: Nvidia's valuation table.
|
|
FY26A |
FY27E |
FY28E |
FY29E |
|
Revenue (in USD millions) |
215,938 |
423,680 |
614,336 |
734,132 |
|
Growth %, y/y |
65.5% |
96.2% |
45.0% |
19.5% |
|
EPS |
4.36 |
9.55 |
13.23 |
15.80 |
|
Growth %, y/y |
48.6% |
119.0% |
38.4% |
19.5% |
|
P/E at current price |
35.6 |
16.3 |
11.8 |
9.8 |
|
Fair P/E (x) |
|
15 |
||
|
Upside potential |
52% |
|||
|
Target price |
USD237 |
|||
|
Source: Bloomberg Finance L.P., iFAST compilations. Data as of 17 November 2026. |
||||
Downside Risks:
1. Tighter export controls and regulatory findings. New rules on sales through third countries, or restrictions on remote access to US chips in overseas data centres, would cut demand from cloud providers in Southeast Asia and beyond. Adverse findings from the board’s review or the SEC investigation would call the quality of past revenue into question and invite further restrictions.
2. A prolonged leadership transition. Without a permanent CEO, decisions on large customer and supply agreements, many built on Huang’s relationships, may slow. The greater risk lies beyond Feynman in 2028, when the roadmap will be set without the leader who drove Nvidia’s annual product cadence.
3. Competition closing the gap. With AMD and custom ASICs narrowing the performance gap and AI cutting the cost of porting away from CUDA, any slip in the upcoming product roadmap could erode Nvidia’s market share.
4. Slower AI infrastructure spending. A pullback in hyperscaler capital expenditure, triggered by weaker AI monetisation or higher funding costs, or delays to data centre builds from power and permitting constraints, would push out orders.
Disclaimers:
This research report was prepared with the assistance of artificial intelligence (AI) tools. iFAST Financial Pte Ltd does not rely exclusively on AI for content generation; the content of this report – including all investment theses, ratings, price targets and conclusions – has been independently reviewed and verified by the research analyst(s) to ensure accuracy and professional integrity.
For specific disclosure, at the time of publication of this report, IFPL (via its connected and associated entities) and the analyst who produced this report hold a NIL position in the abovementioned securities.

