Nvidia announced on November 12, 2026 that Jensen Huang has resigned as president, chief executive officer and member of the board, effective immediately, and that CFO Colette Kress is serving as interim CEO while the board searches for a permanent chief executive. For IT managers and AI buyers, the key official line is that the board expects "the company's operations, customer commitments and product roadmap to continue without interruption." Nvidia has said nothing product-specific beyond that, so the practical work is checking your own orders, support contracts and licences, and watching the November 17 earnings call.
This article sticks to what Nvidia and its regulatory filing state, labels everything else as reported by a named outlet, and focuses on what the change means for teams that buy GPUs, AI software and GPU cloud capacity. This is not financial advice, and nothing here is a view on Nvidia's stock.
Who this is for
This guide is for IT managers, infrastructure and platform teams, procurement leads and AI engineering managers who:
- have Nvidia GPU servers on order or in delivery through an OEM or reseller;
- run workloads on Nvidia GPUs in AWS, Microsoft Azure, Google Cloud, Oracle Cloud or a GPU cloud provider;
- hold NVIDIA AI Enterprise licences or support contracts, or depend on CUDA-based software;
- are planning 2027 AI capacity and want to know whether anything has changed.
At the end you will have a short list of what Nvidia has officially said, what remains unconfirmed, and a checklist of contract and planning items worth reviewing this week.
What Nvidia officially announced
Two primary documents cover the change: Nvidia's press release dated November 12, 2026, and a Form 8-K current report Nvidia filed with the US Securities and Exchange Commission (SEC) the same day.
The press release
The press release, headlined "NVIDIA Announces CEO Transition," states:
- Jensen Huang has resigned as president, CEO and a member of the board, effective immediately.
- Colette Kress, executive vice president and CFO, has agreed to serve as interim CEO. She continues as executive vice president and CFO until a permanent successor is named. Kress has been CFO since September 2013.
- The board is initiating a search for a permanent CEO and "will consider internal and external candidates."
- The board appointed a long-serving director as its first-ever chair of the board.
- Third-quarter fiscal 2027 results will be released on Tuesday, November 17, at approximately 1:20 p.m. PT, "as previously scheduled," followed by the earnings call at 2 p.m. PT.
On continuity, the new board chair said: "The Board has full confidence in Colette and the leadership team, and we expect the company's operations, customer commitments and product roadmap to continue without interruption." Kress said the company's focus "remains on serving" its customers, partners, developers and shareholders.
The release does not give a reason for the resignation.
The SEC filing
The Form 8-K (Item 5.02) adds one detail on timing: it says that on November 11, 2026, the board accepted Huang's resignation as president, CEO and director, effective immediately, and appointed Kress as interim CEO "while the Board conducts a search for a new Chief Executive Officer." The public announcement followed on November 12. The filing is signed by Kress as interim CEO, executive vice president and CFO, and attaches the press release as Exhibit 99.1.
What the board has said since
CNBC reported on November 13 that the board told it: "We are in the process of selecting a search firm for the Nvidia CEO role," "We have not formally engaged any potential candidates," and "Any speculation regarding individual candidates, whether internal or external, is premature." Treat any names you see in the press as speculation until Nvidia announces an appointment.
What has been reported, and what is unconfirmed
These items come from news reporting, not from Nvidia. They are included only so you can tell confirmed facts from open questions.
- Board review. Reporting by CNBC and BNN Bloomberg cites a board investigation announced before the resignation. Nvidia's announcement does not mention it, and no findings have been reported.
- Stock move. Reported by CNBC and BNN Bloomberg: Nvidia shares closed down 4.3% on November 12 after falling sharply at the open. Reported by CNBC: the shares fell 13% for the week ending November 13. This is not financial advice.
- Investor letter. Reported by CNBC on November 14: two large investment firms, Altimeter Capital and Atreides Management, urged the board to "pursue a slow, measured, and deliberate process regarding the future leadership of Nvidia" and to weigh all options, including "the potential return of Jensen Huang." CNBC reported that the letter did not call for his immediate reinstatement. Nvidia had not publicly responded to it in that report.
- Possible congressional testimony. Reported by CNBC: members of Congress have asked Huang to testify, and a House committee chair said Huang had agreed to appear, with the date still being negotiated. Nvidia has not confirmed this, and no hearing date has been officially set.
None of these reports describes a change to Nvidia products, pricing, shipments or support.
What it means for your Nvidia products and roadmaps
The short answer: Nvidia's official position is continuity, stated in general terms. The release does not mention specific product lines, so the sections below state what is publicly documented for each product and what to check yourself.
| Area | What Nvidia has said | What to check |
|---|---|---|
| Data center GPUs (Blackwell, Rubin) | Board expects the product roadmap to continue without interruption; no product-specific statement | Delivery dates and allocation with your OEM or reseller; the November 17 call |
| CUDA and the developer stack | No statement in the release | Toolkit version you depend on; licence terms you accepted |
| DGX Cloud | No statement in the release | Who your contract is with, and its term |
| NVIDIA AI Enterprise | No statement in the release | Licence type, term, support level and renewal date |
| Customer orders and support | Board expects customer commitments to continue without interruption | Signed order terms and support entitlements |
Blackwell and Rubin GPUs
Nvidia said in its January 5, 2026 announcement that "NVIDIA Rubin is in full production, and Rubin-based products will be available from partners the second half of 2026," and named AWS, Google Cloud, Microsoft and Oracle Cloud Infrastructure among the first cloud providers to deploy Vera Rubin-based instances. Nothing in the November 12 announcement changes or restates that schedule. If your 2026 or 2027 plan depends on a Rubin delivery or a Blackwell expansion, the dates that matter are the ones in your own purchase order or cloud capacity agreement, not press coverage.
CUDA
CUDA is distributed under the NVIDIA CUDA Toolkit End User License Agreement. The agreement states that amendments must be "in writing and signed by representatives of both parties," and section 1.6 lists the grounds on which Nvidia may terminate it, such as an uncured breach. It also says the form and content of the SDK may change without prior notice. A leadership change does not alter the terms you accepted; what to track is the toolkit version your applications and drivers depend on, as you would for any release cycle.
DGX Cloud
Nvidia's DGX Cloud page now describes it as "NVIDIA's internal cloud environment for building and operating AI at scale" that "runs across CSPs and NVIDIA Cloud Partners," naming AWS, Google Cloud, Microsoft Azure and Oracle Cloud Infrastructure. If you use DGX Cloud capacity, check whether your agreement is with Nvidia directly or with a cloud provider, and note its term and renewal date.
NVIDIA AI Enterprise licences
The NVIDIA AI Enterprise licensing guide is the document to read before any renewal:
- Licensing is per GPU: "NVIDIA AI Enterprise software is licensed on a per-GPU basis," with one licence for each GPU on a multi-GPU board.
- Subscriptions include Business Standard support for the length of the subscription and must then be renewed.
- Perpetual licences allow use "indefinitely" and include five years of Business Standard support, renewable annually afterwards.
- Cloud marketplace use on AWS, Google Cloud, Azure and Oracle Cloud is priced per GPU per hour.
- Some GPUs, such as H100 PCIe/NVL, include a five-year subscription whose start date is fixed at the GPU's ship date to the OEM plus 90 days. Hopper-based DGX systems include the software; for Blackwell-based DGX systems, the guide says licences must be purchased separately.
Business Standard support, per the same guide, allows cases to be filed 24x7, with an initial response within 4 hours and coverage during local business hours.
A practical checklist for this week
None of these steps assumes anything will go wrong. They are routine checks that a leadership change at a key supplier is a good reason to do now.
1. Check open orders and delivery dates
- List every open Nvidia-based purchase: servers from OEMs (Dell Technologies, HPE, Lenovo, Supermicro and others), GPU cards, networking and DGX systems.
- Confirm the promised delivery date, allocation status and any change-order deadlines with your reseller or OEM account team in writing.
- For cloud capacity, note reserved or committed GPU capacity, its term and its cancellation terms. The EpiFive explainer on EC2 pricing, Savings Plans and Reserved Instances covers how commitment terms work on AWS.
2. Check support contracts
- Find the support level (Business Standard or Business Critical) for each NVIDIA AI Enterprise entitlement and the support end date.
- Make sure your support portal contacts are current, so cases are not delayed by an out-of-date admin list.
- For OEM-supported servers, check whether hardware support runs through the OEM or Nvidia.
3. Check software licence terms
- Record licence type (subscription, perpetual, marketplace or included with hardware), term, start date and renewal date for each NVIDIA AI Enterprise entitlement.
- For included subscriptions, remember the start date is fixed by the GPU ship date, not by when you activated it.
- Keep a copy of the CUDA Toolkit EULA version your teams accepted alongside your other third-party licence records.
4. Watch the November 17 earnings call
The results are due at about 1:20 p.m. PT and the call at 2 p.m. PT on November 17, 2026, per Nvidia. Listen for statements on data center supply, product timing and customer commitments, and note them in your vendor risk file. Statements made on an earnings call are management commentary, not contract terms.
5. Keep a documented multi-vendor option
Most AI buyers already keep a second source in their plans. If yours is not written down, this is a reasonable time to do it. The main alternatives, described from the vendors' own pages:
- AMD Instinct. AMD's current MI350 Series includes the MI350X and MI355X accelerators, which AMD lists with 288 GB of HBM3E memory. They run on AMD's ROCm software stack, which AMD says supports frameworks including PyTorch, JAX and vLLM.
- AWS Trainium. AWS designs its own Trainium AI chips; its current Trainium page leads with Trainium3 and the AWS Neuron SDK.
- Google Cloud TPUs. Google's Tensor Processing Units run code compiled by XLA and support PyTorch and JAX workloads.
- Microsoft Maia 200. Microsoft announced its Maia 200 inference accelerator in January 2026 and says it serves models for Microsoft Foundry and Microsoft 365 Copilot, so most Azure customers meet it through those services rather than as a GPU they rent.
Each of these uses its own software stack, so CUDA-specific code, custom kernels and some libraries will not move over unchanged. A realistic multi-vendor plan starts with an inventory of which workloads are framework-level (often portable after testing) and which depend on CUDA directly (usually a larger effort). The EpiFive guide to production LLMOps architecture shows where model serving sits in a typical stack, which helps with that inventory.
What to watch next
- November 17, 2026: Nvidia's third-quarter fiscal 2027 results and earnings call (confirmed by Nvidia).
- Permanent CEO: the board says it is selecting a search firm and has not formally engaged candidates (board statement to CNBC). No date has been announced.
- Investor pressure: the investor letter reported by CNBC asks for a deliberate process. Whether the board responds is unconfirmed.
- Congressional hearings: reported by CNBC, not confirmed by Nvidia, with no official date.
- Any product or pricing notice: if Nvidia changes a roadmap, support policy or licence term, it will appear in Nvidia's newsroom, documentation or your account team's written notice, which are the sources to rely on.
What we checked
- Opened Nvidia's November 12, 2026 press release on nvidianews.nvidia.com and the same text filed as Exhibit 99.1 on SEC EDGAR on 2026-11-15; the quotes above match both.
- Read Nvidia's Form 8-K (Item 5.02) on SEC EDGAR on 2026-11-15; it gives November 11, 2026 as the date the board accepted the resignation, and the press release is dated November 12.
- Confirmed the earnings date and times (November 17, about 1:20 p.m. PT results, 2 p.m. PT call) in the press release.
- Cross-checked the November 12 closing move (down 4.3%) in CNBC's live updates and in BNN Bloomberg, and the 13% weekly fall in CNBC's November 13 report.
- Read CNBC's November 13 report for the board's statement on the CEO search and its November 14 report on the investor letter.
- Checked the NVIDIA AI Enterprise licensing guide, the CUDA Toolkit EULA and the DGX Cloud page on nvidia.com on 2026-11-15 for the licence and support details above.
- Checked AMD's MI350 Series pages, the AWS Trainium page, Google Cloud's TPU introduction and Microsoft's Maia 200 announcement for the alternatives section.
- Not checked: any customer contract, price list or allocation data; Nvidia has published none of these, and the earnings call had not taken place at the time of writing.
Last tested: 2026-11-15
References
- NVIDIA Newsroom, "NVIDIA Announces CEO Transition" (November 12, 2026): https://nvidianews.nvidia.com/news/nvidia-announces-ceo-transition
- NVIDIA Form 8-K filed November 12, 2026 (SEC EDGAR): https://www.sec.gov/Archives/edgar/data/1045810/000104581026000082/nvda-20261111.htm
- Exhibit 99.1, press release (SEC EDGAR): https://www.sec.gov/Archives/edgar/data/1045810/000104581026000082/nvidiaannouncesceotransiti.htm
- CNBC, Nvidia CEO Jensen Huang resigns: live updates: https://www.cnbc.com/2026/11/12/nvidia-jensen-huang-ceo-live-updates.html
- CNBC, Nvidia board on the CEO search (November 13, 2026): https://www.cnbc.com/2026/11/13/nvidia-huang-ceo-replacement-search.html
- CNBC, market report on Nvidia's weekly share decline (November 13, 2026), cnbc.com
- CNBC, investor letter to the Nvidia board (November 14, 2026): https://www.cnbc.com/2026/11/14/nvidia-investors-jensen-huang-return-board-letter.html
- BNN Bloomberg, Nvidia shares on November 12, 2026: https://www.bnnbloomberg.ca/markets/2026/11/12/nvidia-shares-plunge-13-at-open-after-ceo-jensen-huang-resigns/
- NVIDIA Newsroom, Rubin platform announcement (January 5, 2026): https://nvidianews.nvidia.com/news/rubin-platform-ai-supercomputer
- NVIDIA AI Enterprise licensing guide: https://docs.nvidia.com/ai-enterprise/planning-resource/licensing-guide/latest/licensing.html
- NVIDIA CUDA Toolkit EULA: https://docs.nvidia.com/cuda/eula/index.html
- NVIDIA DGX Cloud: https://www.nvidia.com/en-us/data-center/dgx-cloud/
- AMD Instinct MI350 Series: https://www.amd.com/en/products/accelerators/instinct/mi350.html
- AWS Trainium: https://aws.amazon.com/ai/machine-learning/trainium/
- Google Cloud TPU introduction: https://docs.cloud.google.com/tpu/docs/intro-to-tpu
- Microsoft, Maia 200 announcement (January 26, 2026): https://blogs.microsoft.com/blog/2026/01/26/maia-200-the-ai-accelerator-built-for-inference/
Comments
Questions and tips are welcome. Comments are checked by the EpiFive team before they appear; please do not share passwords, phone numbers or other private details.