ChipMaker Nvidia relies on the production of AI generative developers with a synthetic data company Gretel for more than $ 320 million, according to A. Wired report on Wednesday.
The move comes when generative firms are struggling to find enough training data and improve their models, increasing the need to generate data.
According to the report, Gretel employees will be folded into NVIDIA. Gretel, which produces synthetic or simulated training data for AI models, will strengthen NVIDIA's AI developers' contributions.
A spokesman for Nvidia declined to comment on the report.
Why are synthetic data important
Training for generative AI models, such as Openai's chatgpt, a large model of languages, requires a lot of data. Real-world data can cause problems for AI-namely developers, it can be noisy, and not enough.
AI companies are opposed to the freely -available training limit, leading to conflicts on whether they can use copyrighted content. Hundreds of actors, writers and directors submit an open letter to the Trump administration's Cabinet and Technological Policy to increase its concerns of copyright -protected data. Currently, Openii are calling on the government to Allow greater access to copyrighted material To train models of AI, or otherwise US companies will be left behind China.
Watch this: Watch GTC on NVIDIA GTC 2025: All the highlights in 16 minutes
Synthetic data also has a value in protecting private information. Gretel says her synthetic data It can be used for training models and tools without exposing sensitive or personal information – for example, health care data that does not identify individual people and potentially violate privacy laws.
There are concern about using such data In training models. The overwhelming side of information that is not rooted in reality can increase the likelihood that the model will make things wrong. If the problem worsens bad enough, it can cause a problem known as the model collapse, when the model becomes so incorrect that it becomes useless.