# NVIDIA Deep Learning Institute

> Source: https://aiwiki.ai/wiki/nvidia_deep_learning_institute
> Updated: 2026-08-04
> Fact-checked: 2026-08-04
> Categories: AI Infrastructure, Developer Tools, NVIDIA
> License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/) - attribute to "AI Wiki (aiwiki.ai)"
> Cite as: AI Wiki. "NVIDIA Deep Learning Institute." aiwiki.ai, 4 Aug 2026. https://aiwiki.ai/wiki/nvidia_deep_learning_institute
> From AI Wiki (https://aiwiki.ai), the free encyclopedia of artificial intelligence. Reuse freely with attribution.

**NVIDIA Deep Learning Institute** (DLI) is the training and education arm of [NVIDIA](https://aiwiki.ai/wiki/nvidia), offering hands-on courses, instructor-led workshops, and professional certifications in artificial intelligence, accelerated computing, data science, graphics, and networking. Launched in 2016, the institute had trained more than 600,000 people worldwide as of October 2024 [6]. Its offerings are presented under an "NVIDIA Training" umbrella alongside NVIDIA Academy, a training organization for data center and AI infrastructure professionals whose roots trace to the Mellanox Academy that NVIDIA absorbed through its 2020 acquisition of Mellanox Technologies [9][10][11]. DLI courses give participants access to GPU-accelerated servers in the cloud for hands-on exercises, and completing select courses earns an NVIDIA certificate [1][4]. NVIDIA has also operated a formal certification program of proctored associate-level and professional-level exams, which it expanded broadly from 2024 [7][8].

## Key facts

| Field | Detail |
|---|---|
| Type | Corporate training and education program |
| Parent | [NVIDIA](https://aiwiki.ai/wiki/nvidia) |
| Launched | 2016 [2][5] |
| Formats | Self-paced online courses, live instructor-led workshops, learning paths, free courses, private workshops, certifications [1] |
| Cumulative learners | 7,000+ (November 2016); 250,000+ (October 2020); 400,000+ (March 2023); 600,000+ (October 2024) [2][4][5][6] |
| Certification exams | 12 active exams, 4 associate and 8 professional, priced 125 to 500 US dollars (August 2026) [8] |
| Related programs | NVIDIA Academy, DLI Teaching Kits, University Ambassador Program [11][13] |
| Websites | nvidia.com/training, learn.nvidia.com, academy.nvidia.com [1][11][14] |

## History

### Origins (2016 to 2017)

NVIDIA dates the Deep Learning Institute's launch to 2016 [5]. In a November 11, 2016 company blog post, Will Ramey of NVIDIA wrote that "in just nine months, the institute has trained over 7,000 people worldwide" in applying deep learning to problems such as image recognition, object detection, and natural language processing, and that more than 18,000 hours of hands-on instruction had been delivered at GPU Technology Conference (GTC) events in seven cities "stretching from Amsterdam to Beijing" [2]. The institute worked with Udacity, Microsoft, Coursera, and other educational partners from its first year, and ran onsite workshops at industry conferences such as SC16 [2].

At GTC on May 9, 2017, NVIDIA announced plans to train 100,000 developers that year, which it described as a tenfold increase over 2016 [3]. The company said more than 10,000 developers had already been trained on deep learning using NVIDIA software running on Amazon Web Services EC2 P2 GPU instances, and that it had delivered sold-out public events and onsite training at companies including Adobe, Alibaba, and SAP, at research institutions including the U.S. National Institutes of Health and the Barcelona Supercomputing Center, and at universities including Temasek Polytechnic in Singapore and the Indian Institute of Technology Bombay [3]. "AI is the defining technology of our generation," said Greg Estes, then NVIDIA's vice president of developer programs, announcing an expanded curriculum spanning self-driving cars, healthcare, web services, robotics, video analytics, and financial services [3]. New training labs were co-developed with Amazon Web Services, Facebook, Google, the Mayo Clinic, and Stanford University around the Caffe2, MXNet, and TensorFlow frameworks, coursework was delivered online through AWS and Google's Qwiklabs, and instructors from Hewlett Packard Enterprise, IBM, and Microsoft were certified as DLI training partners [3].

### Remote delivery and growth (2020 to 2024)

By October 2020, DLI had trained more than 250,000 developers globally [4]. When the COVID-19 pandemic halted in-person events, the institute moved its instructor-led workshops into virtual classrooms taught by DLI-certified instructors, and NVIDIA positioned the program as a workforce-upskilling channel for organizations [4]. In March 2023, NVIDIA reported that DLI had trained over 400,000 developers since its 2016 launch, including students from more than 3,000 universities worldwide [5]. At the NVIDIA AI Summit in Washington in October 2024, the company said the institute had "already trained more than 600,000 people worldwide" [6].

## Training catalog and formats

DLI training is organized into self-paced online courses, live instructor-led workshops (public sessions and private deliveries for organizations), curated learning paths, and free introductory courses [1]. Topics span deep learning and generative AI including large language models, data science, accelerated computing with [CUDA](https://aiwiki.ai/wiki/cuda), graphics and simulation, robotics, AI infrastructure, and industry applications such as healthcare and manufacturing [1]. Coursework includes hands-on exercises run on GPU-accelerated servers in the cloud, and completing select courses earns an NVIDIA certificate [1].

GTC conferences remain a major training venue. The GTC 2023 conference offered 29 full-day workshops and 36 hands-on labs, with workshop participants able to earn certificates of competency [5]. At GTC Berlin, scheduled for October 21 to 22, 2026 with workshops on October 20, NVIDIA offers onsite proctored certification exams at no cost to attendees, a value the company puts at 115 to 425 euros per exam [16].

## Certification program

NVIDIA's move from course completion certificates to formal, proctored certification exams began on the infrastructure side of its training business: by March 2023, the NVIDIA Academy site listed two credentials, "NVIDIA Certified Associate - AI in the Data Center" and "NVIDIA Certified Professional - InfiniBand" [10]. At GTC 2024, NVIDIA offered onsite certification at the conference for the first time, running the entry-level AI in the Data Center exam and the intermediate InfiniBand exam, and announced two upcoming associate-level certifications covering generative AI with large language models and multimodal generative AI [7].

As of August 2026, NVIDIA's certification catalog lists 12 exams [8]. Associate-level exams (125 US dollars, one hour) comprise AI Infrastructure and Operations (NCA-AIIO), Accelerated Data Science (NCA-ADS), Generative AI LLM (NCA-GENL), and Generative AI Multimodal (NCA-GENM) [8]. Professional-level exams (two hours, 200 to 500 US dollars) comprise AI Infrastructure (NCP-AII), AI Operations (NCP-AIO), AI Networking (NCP-AIN), AI Rack and Interconnect (NCP-ARI), Accelerated Data Science (NCP-ADS), Generative AI LLMs (NCP-GENL), Agentic AI (NCP-AAI), and OpenUSD Development (NCP-OUSD) [8].

## NVIDIA Academy

NVIDIA Academy is the company's training brand for [data center](https://aiwiki.ai/wiki/data_center) and AI infrastructure professionals. It descends from the training operation of Mellanox Technologies, the high-performance networking company NVIDIA acquired for 7 billion US dollars in a deal completed on April 27, 2020 [9]. As of March 2023, the academy.nvidia.com site, titled "NVIDIA Academy," announced "our new NVIDIA Academy website" while still inviting visitors to register for the free "Mellanox Online Academy," and its catalog centered on networking-lineage topics: [DGX](https://aiwiki.ai/wiki/nvidia_dgx) systems, InfiniBand, Cumulus Linux, RDMA, networking, and AI and data science, with free courses such as InfiniBand essentials and RDMA fundamentals [10].

The academy.nvidia.com domain now redirects to the NVIDIA Academy section of nvidia.com's training site, presented as "Training for AI Infrastructure Professionals" with the tagline "Learn to deploy, run, and optimize AI infrastructure" [11]. Its tracks cover DGX and HGX platform administration, data center deployment, networking, generative AI, high-performance computing, and software tools, alongside AI infrastructure certification [11]. The Academy's catalog page offers self-paced courses, public and private instructor-led workshops, learning paths, and certifications for data center and AI infrastructure roles, and NVIDIA's regional channels actively promote the instructor-led workshops: an August 3, 2026 post by NVIDIA's Asia Pacific account described them as "Designed for infrastructure, platform, and AI operations teams" [12][17].

## Educator and university programs

The DLI Teaching Kits are modular course materials for qualified university educators, covering data science, deep learning, accelerated computing, and robotics, and bundling lecture materials, hands-on exercises, and GPU cloud resources [13]. The first DLI Teaching Kit was developed with Yann LeCun, then director of AI research at Facebook and professor at New York University, covering deep learning theory and application on GPUs using PyTorch; by May 2017 NVIDIA said hundreds of educators were using it, including at the University of Oxford and the University of California, Berkeley [3].

The University Ambassador Program certifies university educators as DLI instructors so they can deliver DLI workshops to students, researchers, and faculty at no cost to participants; NVIDIA says hundreds of universities worldwide have certified DLI University Ambassadors [13].

## Enterprise and public-sector programs

For organizations, DLI sells enterprise packages that bundle self-paced courses and instructor-led workshops with executive briefings and quarterly reporting on employee learning and certifications [14]. Booz Allen Hamilton describes itself as the first federally focused consulting firm certified to deliver the DLI curriculum, Deloitte Consulting has collaborated with NVIDIA on customized curricula, and DLI course content has been developed with industry partners including the Children's Hospital of Los Angeles, the Mayo Clinic, and PwC [14].

In the public sector, NVIDIA announced on September 4, 2025 a commitment valued at 25 million US dollars to K-12 AI education in the United States, made as part of the White House's "Pledge to America's Youth: Investing in AI Education" [15]. Under the program, NVIDIA is partnering with the learning platforms Study Fetch and CK-12 to adapt content from the Deep Learning Institute and NVIDIA Academy for high school classrooms, with a stated aim of reaching 1 million K-12 students within three years [15].

## See also

- [NVIDIA](https://aiwiki.ai/wiki/nvidia)
- [CUDA](https://aiwiki.ai/wiki/cuda)
- [NVIDIA DGX](https://aiwiki.ai/wiki/nvidia_dgx)
- [Data center](https://aiwiki.ai/wiki/data_center)
- [Jensen Huang](https://aiwiki.ai/wiki/jensen_huang)

## References

1. NVIDIA. "Deep Learning Institute (DLI) Training and Certification." Accessed August 5, 2026. https://www.nvidia.com/en-us/training/
2. Will Ramey. "Get Smart: Join the 7,000+ People Trained on Deep Learning This Year." NVIDIA Blog. November 11, 2016. https://web.archive.org/web/2017/https://blogs.nvidia.com/blog/2016/11/11/deep-learning-institute/
3. NVIDIA. "NVIDIA to Train 100,000 Developers on Deep Learning in 2017." May 9, 2017. https://nvidianews.nvidia.com/news/nvidia-to-train-100-000-developers-on-deep-learning-in-2017
4. Katie Young. "Get Trained, Go Deep: How Organizations Can Transform Their Workforce into an AI Powerhouse." NVIDIA Blog. October 1, 2020. https://web.archive.org/web/2021/https://blogs.nvidia.com/blog/2020/10/01/deep-learning-dli-training-organization/
5. Ann Sheridan. "Power-Up Your Skills and Credentials at NVIDIA GTC 2023." NVIDIA Technical Blog. March 13, 2023. https://developer.nvidia.com/blog/power-up-your-skills-and-credentials-at-nvidia-gtc-2023/
6. NVIDIA. "NVIDIA AI Summit Highlights Game-Changing Energy Efficiency and AI-Driven Innovation." NVIDIA Blog. October 8, 2024. https://blogs.nvidia.com/blog/nvidia-ai-summit-washington/
7. NVIDIA. "New Workshops and Certification at NVIDIA GTC 2024." NVIDIA Technical Blog. February 12, 2024. https://developer.nvidia.com/blog/new-workshops-and-certification-at-nvidia-gtc-2024/
8. NVIDIA. "NVIDIA Certification." Accessed August 5, 2026. https://www.nvidia.com/en-us/learn/certification/
9. NVIDIA. "NVIDIA Completes Acquisition of Mellanox, Creating Major Force Driving Next-Gen Data Centers." April 27, 2020. https://nvidianews.nvidia.com/news/nvidia-completes-acquisition-of-mellanox-creating-major-force-driving-next-gen-data-centers
10. NVIDIA Academy website (archived snapshot). March 14, 2023. https://web.archive.org/web/20230314132814/https://academy.nvidia.com/en/
11. NVIDIA. "NVIDIA Academy: Training for AI Infrastructure Professionals." Accessed August 5, 2026. https://www.nvidia.com/en-us/training/academy/
12. NVIDIA. "Find Training: NVIDIA Academy." Accessed August 5, 2026. https://www.nvidia.com/en-us/training/academy/find-training/
13. NVIDIA. "Enhance AI and Data Science Curriculum: NVIDIA Programs for Educators." Accessed August 5, 2026. https://www.nvidia.com/en-us/training/educator-programs/
14. NVIDIA. "Deep Learning Institute Enterprise Solutions." Accessed August 5, 2026. https://learn.nvidia.com/en-us/training/enterprise-solutions
15. NVIDIA. "NVIDIA Pledges AI Education Funding for K-12 Programs." NVIDIA Blog. September 4, 2025. https://blogs.nvidia.com/blog/ai-education-k-12/
16. NVIDIA. "GTC Berlin 2026: Workshops and Training." Accessed August 5, 2026. https://www.nvidia.com/en-eu/gtc/training/
17. NVIDIA Asia Pacific (@NVIDIAAP). X post. August 3, 2026. https://x.com/NVIDIAAP/status/2084157318162243728