Jensen Huang is NVIDIA’s co-founder, president, and chief executive, having led the company since its founding in 1993. He transformed a graphics-chip business into an accelerated computing platform spanning processors, developer software, networking, and the physical infrastructure behind modern artificial intelligence.
Born in Taiwan, Huang moved to the United States as a child and earned electrical-engineering degrees from Oregon State University and Stanford. He worked at AMD and LSI Logic before founding NVIDIA with Chris Malachowsky and Curtis Priem. The company initially targeted gaming and multimedia; its GeForce 256, introduced in 1999, helped establish the modern graphics processing unit.
NVIDIA’s decisive expansion beyond graphics came with CUDA, introduced in 2006, which enabled developers to program its processors for general-purpose parallel computing. Scientific applications followed, and in 2012 Alex Krizhevsky and collaborators used NVIDIA GPUs to train AlexNet, an influential deep-learning breakthrough. Huang’s contribution was building the hardware and software platform on which such work could run—not creating AlexNet itself.
- Different workloads demand different infrastructure. In recorded questions for AI Engineer World’s Fair 2025, Huang contrasted long-running, memory-intensive reasoning agents with always-available speech and vision assistants. Their conflicting prefill and decode tradeoffs, throughput requirements, and latency constraints make flexible infrastructure a central engineering challenge.
- Domain-specific agents require deliberate engineering. Huang identifies tools, planning, short- and long-term memory, and specialized workflows as components developers must combine when adapting increasingly capable general-purpose models to particular applications.
- AI factories are physical infrastructure. Huang describes AI factories as facilities that convert energy and data into computational output. His infrastructure strategy encompasses chips, networking, electricity, buildings, and long-lived sites capable of supporting successive generations of computing equipment.
