NVIDIA was founded in 1993 to make 3D graphics chips for PC games. It launched the GeForce 256, which it called the first GPU, in 1999, and CUDA in 2006. After…
Jensen Huang, Chris Malachowsky and Curtis Priem set out to build 3D graphics chips for PC games. Huang is still CEO 33 years later.
The NV1 bet on a rendering method Microsoft's Direct3D did not use. It sold poorly. The RIVA 128 embraced Direct3D and sold about a million units in four months.
NVIDIA listed on Nasdaq at $12 a share, then launched the GeForce 256, which it marketed as the world's first GPU.
CUDA let developers program NVIDIA GPUs in a C-like language for any parallel maths, not just graphics. It runs only on NVIDIA chips.
A neural network trained on two GeForce GTX 580 gaming cards cut image-recognition errors to 15.3 percent, against 26.2 percent for the runner-up.
The DGX-1 was an AI computer in a box; the first one went to OpenAI. Volta's V100 added Tensor Cores for neural-network maths.
The $6.9 billion deal gave NVIDIA the network that links thousands of GPUs. Networking earned $31.4 billion in fiscal 2026.
The H100 had 80 billion transistors and a Transformer Engine. After ChatGPT, demand exploded and NVIDIA became the first chipmaker worth $1 trillion.
Two dies joined into one 208-billion-transistor GPU, sold in racks such as the GB200 NVL72 with 72 GPUs acting as one.
A $589 billion one-day loss after DeepSeek, new China export limits, then the first $4 trillion and $5 trillion market values.
A platform of CPU, GPU and networking chips designed together. The NVL72 rack links 72 Rubin GPUs and 36 Vera CPUs.
$193.7 billion from Data Center against $16.0 billion from Gaming in fiscal 2026. In the July 2026 quarter Data Center alone was $89.0 billion.
Read the complete story, FAQs and key facts.