Nvidia, Alphabet's Google and hot startup OpenAI are turning to "synthetic data" factories amid demand for massive amounts of data needed to train deep learning artificial intelligence algorithms. At this week's Consumer Electronics Show, the chipmaker's chief executive touted synthetic data's capabilities, which could be a plus for Nvidia stock.
The move toward synthetic data comes amid reports that AI companies are running out of real world data to train powerful AI models.
"Synthetic data offers a vital solution for addressing scarce or sensitive data requirements. This trend is accelerating as major AI companies exhaust available internet data for training," said data scientist Ben Lorica in a 2025 outlook report.
"Teams can already leverage foundation models to generate synthetic data for specific use cases, while larger organizations may combine synthetic data with their proprietary datasets," added Lorica, who edits the Gradient Flow AI newsletter. "Look for improved synthetic data generation tools emerging from major AI labs, making this technology more accessible to practitioners."
At CES 2025, Nvidia CEO Jensen Huang touted artificial intelligence's future role in automotive and robotics applications. Synthetic data will help, he says.
Nvidia Stock: 'Data Factory'
"Nvidia is collecting and organizing traditional data and using them to create synthetic data," said economist Ed Yardeni in a report on CES developments. "Both traditional and synthetic data will be used to train AI agents and robots in Nvidia's data factory."
Yardeni added: "(The chipmaker) developed the Nvidia Cosmos by having it watch 20 million hours of video about nature, humans, and anything to do with the physical world. Based on those real scenarios, it can also create synthetic data to create even more scenarios. It can then use its real and synthetic data to train robots that need to navigate in the world, whether working in a warehouse or driving an autonomous vehicle."
Google's cloud computing unit also is making a big push into synthetic data for enterprise applications. Further, OpenAI's newest foundation models capable of enhanced reasoning skills use synthetic data generation techniques.
Heading into 2025, one big debate is whether AI models have started to plateau amid difficulty associated with sourcing high-quality, human-made training data. Also, tech behemoths like Google and Meta Platforms own in-house proprietary data from YouTube, Maps, Instagram and Facebook that can be used to build bigger models.
Meanwhile, Nvidia stock has gained 4% in 2025. Nvidia stock jumped 171% in 2024 after surging 239% in 2023.
Follow Reinhardt Krause on Twitter @reinhardtk_tech for updates on artificial intelligence, cybersecurity and cloud computing.
