Tech companies · Science & technology
NVIDIA エヌビディア
えぬびでぃあ
NVIDIA is an American company that designs graphics processing units (GPUs), originally for video games. Because GPUs are well suited to training artificial intelligence, NVIDIA became the leading supplier of AI chips and rose into the ranks of the world's most valuable companies from 2024.
- Headquarters
- Santa Clara, California, US
- Founders
- Jensen Huang, Chris Malachowsky, Curtis Priem
- Main products
- GeForce GPUs, data-centre AI chips, CUDA platform
History
Jensen Huang, a Taiwanese-born engineer, founded NVIDIA with two colleagues in 1993; legend has it the idea was hatched in a Denny's restaurant. In 1999 NVIDIA launched the GeForce 256, marketing it as the first 'GPU', and the GeForce line became a staple for PC gamers.
From gaming to AI
A GPU has thousands of cores computing in parallel, ideal for the matrix maths of neural networks. NVIDIA released the CUDA programming platform in 2006, and in 2012 AlexNet, trained on NVIDIA GPUs, won the ImageNet contest. When generative AI exploded from 2022, demand for data-centre chips such as the H100 soared, multiplying NVIDIA's revenue and market value.
In Vietnam
Jensen Huang visited Vietnam in 2023 and 2024, saying he wanted it to become NVIDIA's 'second home'. In 2024 NVIDIA acquired VinBrain, a healthcare AI start-up from Vingroup, and signed an agreement with the government on an AI research and development centre. FPT and Viettel are building 'AI factories' using NVIDIA GPUs.
In Japan
In Japan, SoftBank, telecom companies and research institutes are building AI supercomputers using NVIDIA GPUs. The Nintendo Switch runs on NVIDIA's Tegra chip, and Huang has often recalled the company's early ties with Japanese game makers such as Sega.
Vocabulary corner
Japanese: ジーピーユー (GPU), 画像処理 gazō shori (image processing), グラボ / グラフィックボード (graphics card), 並列計算 heiretsu keisan (parallel computing), 時価総額 jika sōgaku (market capitalisation), ファブレス (fabless). Vietnamese: card đồ hoạ, chip AI, vốn hoá.