# Graphics processing unit: Evolution

Source page: https://www.materialinsights.org/technologies/graphics_processing_unit/tech-evolution/

#### Timeline

The milestones from first idea to today

| Year | Milestone |
|---|---|
| 1994 | Sony first uses the term GPU, for the PlayStation ^[2] |
| 1995 | 3Dlabs brings out its Glint chip set with a geometry processor ^[2] |
| 1999 | Nvidia markets the GeForce 256, a single-chip GPU with 17 million transistors made by TSMC ^[1] |
| 2006 | The GeForce 8800 GTX, the first CUDA-capable graphics card, goes on sale on 8 November ^[3] |
| 2007 | Nvidia releases the CUDA programming system for general computing on GPUs ^[4] |
| 2012 | AlexNet is trained on two Nvidia graphics cards in Toronto ^[4] |
| 2025 | Nvidia's data-centre revenue reaches 115.2 billion US dollars in its fiscal year to January ^[5] |

as of 2026-11-11

#### Lineage

The technologies it came from and the ones that came from it
 Came before it

- [CPU Microprocessor](https://www.materialinsights.org/technologies/microprocessor/) Before the single-chip GPU, 3D accelerators used the central processor for transform and lighting; the GeForce 256 moved that work onto a graphics chip of its own ^[1].
 Builds on

- [IC Integrated circuit](https://www.materialinsights.org/technologies/integrated_circuit/) A GPU is one very large integrated circuit; the GeForce 256 already held 17 million transistors ^[1].
 Enables

- [DL Deep learning](https://www.materialinsights.org/technologies/deep_neural_networks/) AlexNet, the landmark image-recognition network of 2012, was trained on two Nvidia graphics cards ^[4].
- [LLM Large language models](https://www.materialinsights.org/technologies/large_language_models/) Large language models are trained and run in data centres; Nvidia says reasoning AI drives demand for its Blackwell data-centre chips ^[5].
 Works with

- [PCo Personal computer](https://www.materialinsights.org/technologies/personal_computer/) Add-in graphics cards for PCs remain a large market, with 11.6 million shipped in one quarter of 2025 ^[6].

as of 2026-11-11

#### Sources

1. [Nvidia's GeForce 256: The First Fully Integrated GPU](https://www.electronicdesign.com/technologies/embedded/article/21178111/jon-peddie-research-nvidias-geforce-256-the-first-fully-integrated-gpu) Electronic Design (Jon Peddie), 2022
2. [Is it Time to Rename the GPU?](https://www.computer.org/publications/tech-news/chasing-pixels/is-it-time-to-rename-the-gpu) IEEE Computer Society, Chasing Pixels (Jon Peddie), 2018
3. [GeForce 8800 GTX: Celebrating 20 Years Of Gaming And CUDA](https://www.nvidia.com/en-au/geforce/news/8800-gtx-gpu-20-year-celebration-dlss-4-5-path-tracing-steam-demo/) NVIDIA GeForce News (Richard Padilla), 2026
4. [CHM Releases AlexNet Source Code](https://computerhistory.org/blog/chm-releases-alexnet-source-code/) Computer History Museum (Hansen Hsu), 2025
5. [NVIDIA Announces Financial Results for Fourth Quarter and Fiscal 2025](https://nvidianews.nvidia.com/news/nvidia-announces-financial-results-for-fourth-quarter-and-fiscal-2025) NVIDIA Newsroom, 2025
6. [Q2'25 PC graphics add-in board shipments increased 27.0% from last quarter](https://www.jonpeddie.com/news/q225-pc-graphics-add-in-board-shipments-increased-27-0-from-last-quarter/) Jon Peddie Research, 2025

## Other sections

- [Summary](https://www.materialinsights.org/technologies/graphics_processing_unit/index.md)
- [Overview](https://www.materialinsights.org/technologies/graphics_processing_unit/tech-overview/index.md)
- [Science](https://www.materialinsights.org/technologies/graphics_processing_unit/tech-science/index.md)
- [Materials](https://www.materialinsights.org/technologies/graphics_processing_unit/tech-materials/index.md)
- [Energy footprint](https://www.materialinsights.org/technologies/graphics_processing_unit/tech-energy-footprint/index.md)
- [Industries](https://www.materialinsights.org/technologies/graphics_processing_unit/tech-industries/index.md)
