Company
ActiveNvidia
Santa Clara · Since 1993
Still independent. A graphics-chip company became the main supplier of the processors used to train large models.
A later timeline chapter projects a role for this organization. That chapter is a projection. This page is about the institution as it exists, or as it existed.
Jensen Huang, Chris Malachowsky, and Curtis Priem founded Nvidia in 1993 to make chips that drew graphics for games. The processors turned out to be good at the particular arithmetic neural networks require. A gaming company became the supplier the AI labs could not easily replace.
Nvidia designs the chips. Others, principally TSMC, manufacture them. The employment is engineers in Santa Clara and a manufacturing workforce in Taiwan that is not on Nvidia's payroll and without which the company has nothing to sell.
The pick and shovel
Cuda, Nvidia's software layer, is why researchers write for these chips and not for a rival's. That software employs developers, and it locks in the hardware sale. Data-center operators then hire the people who rack the machines, cool them, and keep them fed with electricity. Nvidia's own headcount is modest next to the power stations and the foundries. Its position in the 2030 chapter is as the supplier whose output sets the pace of everyone else's hiring.
Still a chip company
Nvidia remains independent and, for a period, extraordinarily valuable. Valuation is not a history. The history is the shift from a graphics board in a personal computer to a machine sold to companies that train models. The 2030 chapter's use of Nvidia is a projection about how far that shift goes. The company is already here.
The work it organized
- Chip designer
- Cuda developer
- Foundry technician
- Data-center technician
On the timeline
Related
- IntelActive