Last Modified: Oct 31, 2025
These Product-Specific Terms for AI products (“AI Product-Specific Terms”) are an exhibit to the NVIDIA Software License Agreement at https://www.nvidia.com/en-us/agreements/enterprise-software/nvidia-software-license-agreement/ (the “Agreement”) and are incorporated as an integral part of the Agreement.
Scope. The Agreement and these AI Product-Specific Terms govern the use of NVIDIA AI products including:
1.1 "Community Products" means Free SDKs, the Software in the NVIDIA NGC AI catalog at https://ngc.nvidia.com/ (as updated from time to time) and publicly-available versions of NVIDIA Base Command Manager software and NVIDIA AI Workbench software, and
1.2 "DGX Systems" means NVIDIA-branded DGX systems, except for DGX Spark which may be NVIDIA or OEM branded systems.
1.3 Enterprise Products means the Software in NVIDIA AI Enterprise, NVIDIA Base Command Manager software, NVIDIA AI Workbench software, NVIDIA Aerial CUDA-Accelerated RAN, NVIDIA Run:ai, NVIDIA DGX Software (as defined) and NVIDIA Mission Control (as defined).
1.4 “NVIDIA AI Enterprise” means NVIDIA AI software in the NVIDIA NGC AI Enterprise catalog at https://ngc.nvidia.com/nvaie-supported, including NVIDIA NIMs and NVIDIA NeMo (except as otherwise expressly permitted under these Product Specific Terms), as updated from time to time.
1.5 “NVIDIA DGX Software” means NVIDIA-provided operating system and software specific for DGX Systems as indicated by NVIDIA in documentation, including the following if delivered: NVIDIA Base Command Manager, NVIDIA Magnum IO, NVIDIA Unified Fabric Manager, and other software and materials.
1.6 “NVIDIA Enterprise” means one Part Number for access to NVIDIA AI Enterprise and Omniverse Enterprise.
1.7 “NVIDIA Mission Control” means the NVIDIA provided software tool for management and monitoring NVIDIA GPU systems and includes features like cluster management, health monitoring and optimization.
1.8 “Research Projects” means the NVIDIA licensed source code released as part of the research projects listed at https://docs.nvidia.com/research-projects.pdf and documentation, and expressly exclude for the purpose of the Agreement any models, checkpoints, JSON file datasets, or third party licensed content available from the same repositories.
Grant. Section 1.1 of the Agreement is amended and restated to read as follows:
1.1 Subject to the terms of the Agreement, Customer’s Order Form and license parameters of an Enterprise Product Part Number, and payment of applicable fees, NVIDIA grants Customer a limited, worldwide, non-exclusive, non-transferable, non-sublicensable (except as expressly provided in the Agreement) license to do the following for the duration of the license:
1.1.1 install, use, reproduce, create Derivative Samples and Derivative Models, and configure (if applicable) the Software,
1.1.2 offer as a service the Software, Derivative Samples and Derivative Models as part of a Customer Product, and
1.1.3 sublicense and distribute the Software, Derivative Samples and Derivative Models as part of a Customer Product.
Limited Grant. Section 1.2 of the Agreement is amended and restated to read as follows:
1.2 Limited Grants. The grants above are further limited as follows:
1.2.1 If the Customer Product is an application, it must be a Compatible Application.
1.2.2 NVIDIA DGX Software are only licensed for use in DGX Systems, and firmware licensed in or for specific DGX Systems may not be used in other NVIDIA or non-NVIDIA products.
1.2.3 Base Command Manager when included in NVIDIA AI Enterprise only requires licensing for compute nodes under management, not management nodes.
1.2.4 NVIDIA AI Workbench is licensed for use under Section 1.1.1 and Section 1.1.3, not licensed for use under Section 1.1.2.
1.2.5 Customer may create a Custom TTS Application for their own use but cannot offer it as a service to third parties as described in Section 1.1.2 or distribute as part of a Compatible Application as described in Section 1.1.3. Customer must have sufficient rights and licenses for content used with a Custom TTS Application to generate new content. “Custom TTS Application” means a custom text-to-speech application that enables the creation of synthetic voice output that resembles an input voice.
1.2.6 “not for resale” or “NFR” (typically a license to a distributor or an authorized reseller/distributor) licenses to Enterprise Products are licensed for limited use under Section 1.1.1 solely for internal evaluation or to demonstrate the Enterprise Product to others.
1.2.7 trial licenses to Enterprise Products are licensed for limited use under Section 1.1 solely for the trial period.
1.2.8 the developer program for AI products provides access to Enterprise Products under licenses for personal use under Section 1.1.1 above for Software that is part of this program (as updated by NVIDIA from time to time) solely for internal evaluation, development or test non-production purposes. Software offered as part of the developer program is not for use, distribution or deployment in production. NVIDIA may admit persons into the developer program for AI products, as well as change or discontinue the program and its offerings, at NVIDIA’s sole discretion.
1.2.9 Research Projects licenses to Enterprise Product Subscription for AI products may exercise only the grants in Section 1.1.1 and Section 1.1.2 above for Research Projects during the term of the Subscription solely for internal purposes, and Support, Maintenance and Updates are not provided for Software available as part of Research Projects. Customer will retrieve the Software from the Research Project repositories for use in accordance with the terms of the Agreement. Subject to the rights of NVIDIA and its suppliers in Software Offerings, Customer holds all rights, title and interest in and to Customer’s output of Research Projects including their respective Intellectual Property Rights.
1.2.10 NVIDIA inference microservices (NVIDIA NIMs) designated by NVIDIA for use with NVIDIA RTX, GeForce RTX GPUs or NVIDIA Grace-Blackwell client systems as Free SDKs may be used without the purchase of a Subscription or Services if such NIMs are: (a) used on a PC or workstation with NVIDIA RTX or NVIDIA GeForce RTX GPUs and (b) not used in a commercial kiosk, server or other system used to service multiple users. For other uses of these NIMs, you may purchase an NVIDIA AI Enterprise Subscription or obtain a licensed under an NVIDIA program, as available.
License Types. Section 1.4 (“License Types”) of the Agreement is amended to add the following:
NVIDIA may offer and sell, with different prices and terms, Subscriptions for use of Enterprise Products in a designated platform. For example, NVIDIA may indicate that a subset of the Software is supported in a certain platform (e.g., a Subscription for the IGX platform or for RTX PRO Server) and offer and sell a dedicated Enterprise Product Subscription for use of the applicable Enterprise Product only on the identified platform. If you obtained a dedicated Enterprise Product Subscription at a different price solely for use in a particular platform and you use in other platforms, NVIDIA or an authorized reseller/distributor may invoice you to pay a higher value Enterprise Product Subscription.
A new subsection 1.7 is added to the Agreement as follows:
1.7 Requirements. The requirements in this section apply to Software, Derivative Samples and Derivative Models to exercise the license grants in Sections 1.1.2 and 1.1.3:
1.7.1 NVIDIA Works Notice. The following notice must be included in source files of Derivative Samples and Derivative Models distributed: “This software contains source code provided by NVIDIA Corporation.”
1.7.2 Use Report. Customer must report upon NVIDIA’s email request, no more than monthly, the Software in use by Customer Personnel and Customer End Users Customer enabled, quantity, start and end dates, and any other reasonably requested information for NVIDIA to determine the fees due.
1.7.3 License Administration. Customer will be responsible for the administration of licenses and services to Customer’s End Users during their use period.
Limitation. Section 8 of the Agreement is amended and restated to read as follows:
The following limitations and restrictions apply to the Software, Derivative Samples and Derivative Models, and Customer is responsible for the consequences of non-conformance with these limitations:
8.1 Customer will use the Software exclusively for authorized purposes, consistent with the Agreement’s terms and all applicable laws, regulations and the rights of others.
8.2 Product-Specific Terms may indicate that NVIDIA proprietary Software, Derivative Samples and Derivative Models are licensed only to run on NVIDIA Platforms.
8.3 Customer may not combine the use of paid and unpaid Software, Derivative Samples and Derivative Models in a way that avoids incurring fees or exceeding use limits or quotas.
8.4 Customer may not reverse engineer, decompile, disassemble Software components provided in binary form, nor attempt in any other manner to obtain source code of such Software components.
8.5 Except as expressly granted in the Agreement or Product-Specific Terms, Customer may not copy, sell, resell, rent, sublicense, transfer, assign, timeshare, distribute, modify, or create derivative works of any portion of the Software, including, without limitation, in any publicly accessible software repositories.
8.6 Customer may not indicate that a product or service developed with the Software is sponsored or endorsed by NVIDIA unless expressly authorized in writing by NVIDIA.
8.7 Customer may not bypass, disable, or circumvent any technical limitation, encryption, security, digital rights management or authentication mechanism contained in the Software.
8.8 Customer may not use the Software components governed by the Agreement in any manner that would cause components to become subject to an OSS License or other shareware license.
8.9 Customer may not distribute or disclose to third parties the results of benchmarking, competitive analysis, regression or performance data relating to the Software without the prior written permission from NVIDIA, except as described at https://docs.nvidia.com/nvidia-containers-benchmarking.pdf.
8.10 Customer may not replace any NVIDIA software components in the Software that are governed by the Agreement with other software that implements NVIDIA application programming interfaces (APIs).
8.11 Customer may not reverse engineer, decompile or disassemble any portion of the output generated using an NVIDIA proprietary software development kit (e.g., NVIDIA CUDA toolkit), including their development tools and compilers.
8.12 Customer may not use the Software or NVIDIA Confidential Information for the purpose of (i) developing competing products or technologies or assisting a third party in such activities, or (ii)identifying or supporting an assertion or potential assertion of any intellectual property rights against NVIDIA (including patent, copyright, or trade secret).
8.13 Customer acknowledges that the Software as delivered under the Agreement is not tested or certified by NVIDIA for use in any Critical Application. Beyond NVIDIA delivering the Software in accordance with the Agreement, NVIDIA will not be liable to Customer or any third party, in whole or in part, for any claims or damages arising from such uses. Customer is solely responsible for ensuring that systems and applications developed or deployed with the Software include sufficient safety and redundancy features and comply with all applicable legal and regulatory standards and requirements.
8.14 Customer may not use the Software to infringe any third party’s Intellectual Property Rights.
8.15 The following proprietary Software components are licensed only to run on systems with NVIDIA Platforms: NVIDIA software development kits (such as CUDA toolkit, TensorRT, cuDNN) and NVIDIA drivers.
8.16 Customer may not use the Software or any of its components for the purpose of emotion recognition. Any technology included in the Software may only be used as fully integrated in the Software and consistent with all applicable documentation.
11.1 Collection Purposes.
11.1.1 NIM Containers. Customer acknowledges that NIM container releases that collect data, collect it for the following purposes: (a) properly configure and optimize products for use with Software; and (b) improve NVIDIA products and services. NIM containers that collect telemetry will have environment variables to enable or disable data collection. When enabled with telemetry mode, telemetry information collected includes: (i) microservice settings, (ii) usage data; and (iii) hardware environment. Customer can disable telemetry by setting the environment variable NIM_TELEMETRY_MODE=0 and enable by setting NIM_TELEMETRY_MODE=1. For more details on NIM container telemetry, see the listing page for each NIM container in the NVIDIA NGC AI Enterprise catalog at https://ngc.nvidia.com/nvaie-supported.
11.1.2 TAO Toolkit. Customer acknowledges that TAO Toolkit releases that collect data, collect it for the following purposes: (a) properly configure and optimize products for use with Software; (b) deliver content or service through the Software; (c) improve NVIDIA products and services. When enabled, telemetry information collected includes: (i) installed application versions used with Software; and (ii) performance, invocation and usage data. For NVIDIA AI Enterprise Products, telemetry collection is disabled by default. In releases where telemetry is enabled by default, Customer can disable telemetry collection by setting the TELEMETRY_OPT_OUT environment variable or configuring telemetry opt-out parameters in deployment configurations.
11.1.3 Generally. NVIDIA may require certain personal information such as name, email address and entitlement information to deliver Software or provide Services to Customer.
(v. October 14, 2025)