Also read | Nvidia’s partnership with Reliance, Tatas to boost India’s AI infrastructureHuang said that computer hardware technology today has reached its limit in terms of how many transistors can be crammed into a chip, and by extension how powerful processors can be.“We can no longer afford to do nothing in software and expect to improve computing experience," he said. “We started our company to improve software."Nvidia makes high-end graphics processing units (GPU) that coupled with the firm’s CUDA software handle multiple simpler tasks simultaneously rather than handling complex calculations that conventional processors do.
Each year, the computation power of these GPUs grows by a factor of 4, Huang said, outpacing Moore’s Law.These GPUs, which first revolutionized the gaming industry, are now touted to become the backbone of the AI revolution. Nvidia’s latest Hopper GPU and upcoming Blackwell GPU are the building blocks of large servers that handle the computation for AI.To be sure, just a month before OpenAI’s ChatGPT, the first mainstream generative AI application based on Nvidia’s chips, was launched, Nvidia was worth $275 billion.
In 24 months since then, its market cap has soared over 12x. The unprecedented surge catapulted the chipmaker ahead of Elon Musk’s Tesla to the second most valuable in the world.
Read more on livemint.com