Who is Jensen Huang?
Jensen Huang, also known as Jen-Hsun Huang, is NVIDIA’s cofounder, president and chief executive officer. He has served as chief executive and a board member since the company was established in 1993. [1]
He was born in Taiwan in 1963 and immigrated to the United States at nine. His career combines electronic engineering, semiconductor design and business leadership. [7]
Education and early employment
He earned his electrical engineering BSEE at Oregon State University in 1984, where he met his future wife, Lori Mills. [8]
He completed a master’s degree in electrical engineering at Stanford in 1992. This education complemented professional experience focused on tools and technologies for designing integrated circuits. [9]
He worked as a microprocessor designer at AMD from 1983 to 1985 and then at LSI Logic until 1993. There he led Coreware, a unit focused on system-on-chip technology. [10]
Founding NVIDIA and early challenges
He founded NVIDIA on 5 April 1993 with Chris Malachowsky and Curtis Priem. Its initial aim was to bring advanced three-dimensional graphics to games and multimedia applications, a market where visual performance became a way to differentiate personal computers. [4]
During a 2013 visit to Oregon State, Huang recalled an early crisis: falling memory prices and an influx of competitors undermined the value of the design the company was developing. The episode forced a strategic rethink while it was still a young business. [8]
From graphics to accelerated computing
The 1999 GeForce 256 launch was a milestone: it integrated transformation, lighting and rendering functions into the graphics processor. [7]
CUDA, introduced in 2006, opened those GPUs to programming tasks beyond drawing images. Its tools enabled C and mathematical libraries to distribute calculations across numerous threads. A GPU could work alongside a CPU and accelerate parallel parts of a program, expanding the market into science and engineering. [11]
In a 2016 article, Huang explained his commitment to combining parallel processors, programming tools and researcher access to those capabilities. He identified AlexNet’s GPU training in 2012 as a turning point. NVIDIA’s contribution was to provide infrastructure on which scientific teams developed their networks and methods. [12]
Systems for training and running artificial intelligence
At GTC 2016 NVIDIA introduced DGX-1 as an integrated deep learning system. It combined eight Tesla P100 accelerators, NVLink interconnects and software including cuDNN and DIGITS. The offering brought together components researchers needed to coordinate when training networks, extending the company’s products from individual chips to complete systems. [13]
In January 2026 the company introduced its Rubin platform, combining CPUs, GPUs and several jointly designed networking components. Huang placed it within a regular renewal of its AI systems. The approach addresses not only computation but also data movement, memory and communication between processors. [16]
Mellanox, Arm and regulatory challenges
The Mellanox acquisition closed in April 2020. It added high-performance networking technologies to NVIDIA’s computing capabilities. Huang described the transaction as a step towards an offering spanning processors, communications and software for data centres, where connecting equipment matters as much as accelerating individual chips. [14]
The proposed Arm acquisition followed a different course. NVIDIA and SoftBank cancelled it in February 2022 because regulatory obstacles prevented completion. Huang said NVIDIA would continue working with Arm as a licensee. The transaction did not make NVIDIA the owner of Arm. [15]
In its fiscal 2026 annual report, NVIDIA identified advanced-processor export controls as a risk to sales and supply. Its international expansion also depends on authorisations and trade restrictions. [10]
Omniverse, simulation and three-dimensional environments
Omniverse connects three-dimensional data through OpenUSD and simulates scenes, sensors and physical behaviour. Its tools enable robot testing, synthetic data generation and facility representations. [2]
Its asset-readiness checks distinguish a visual scene from a usable simulation model, which requires appropriate geometry, physical properties and sensor behaviour. [2]
His view of AI as infrastructure
In March 2026 he described AI through five layers: energy, chips, infrastructure, models and applications. He argued that expansion requires physical investment and skilled workers as well as software. This business thesis about the sector’s development connects data centres and energy requirements with new products and services. [17]
Supporting education and research
Jensen and Lori Huang pledged $30 million for the Stanford engineering centre bearing his name. The contribution was announced as support for a space bringing together teaching, research and collaboration across disciplines. [19]
The Jen-Hsun Huang Engineering Center was formally dedicated on 5 October 2010. Stanford presents it as a home for creativity, collaboration, lifelong learning and entrepreneurship within its science and engineering buildings. [20]
In 2022 Oregon State announced a $50 million gift from the couple for the Jen-Hsun and Lori Huang Collaborative Innovation Complex. The project was designed for research in AI, materials and robotics, with fields including climate, oceans and water resources. The contribution forms part of a wider university collaboration. [18]
Awards and professional recognition
The Semiconductor Industry Association awarded him its 2021 Robert N. Noyce Award, recognising leadership and contributions to the semiconductor industry. The association highlighted his work on accelerated computing platforms and NVIDIA’s development. [21]
The Computer History Museum named him a Fellow in 2024 for leadership in graphics, accelerated computing and artificial intelligence. [7]
In 2026 he was recognised with the IEEE Medal of Honor. The award acknowledged his accelerated-computing career; Huang connected the recognition to the collective work of NVIDIA’s engineers and researchers and the community that has developed its ecosystem. [22]
He also received the US National Medal of Science in 2026. [23]
BMW: testing the factory before changing it
At GTC 2023, Huang and Milan Nedeljković presented a demonstration of BMW’s virtual factory in Omniverse. It showed how to decide where to place a robot in a constrained space using a model shared by planning and logistics teams. [5]
The project documentation explicitly includes worker safety and ergonomics among factory-reconfiguration challenges. BMW uses virtual environments to coordinate planning and study interactions between people, equipment and robots before transferring changes to the physical space. [6]
The digital twin’s preventive connection
The OSH value lies in anticipating design conflicts: changing a machine also changes access, routes, intervention space and manual tasks. Reviewing these relationships in a common model allows discussion before implementing the modification.
Huang publicly presented this collaboration; preventive planning belongs to the teams designing and validating each installation. The twin must represent actual workplace conditions, including maintenance operations and abnormal situations. A convincing simulation does not by itself establish installation safety or quantify a reduction in accidents.
