Nvidia CEO Jensen Huang has claimed that artificial general intelligence (AGI) has already been launched with OpenAI’s latest model, GPT-6 Astra, marking a significant milestone in the evolving field of AI.
On Sunday, the Nvidia CEO congratulated the ChatGPT manufacturer for launching AGI via a post on X (formerly Twitter), OpenAI’s new AI system and the Nvidia hardware used to train it. However, OpenAI has not independently confirmed that its latest model has achieved AGI.
He wrote, “GPT-6 Astra, trained on ~100K+ NVIDIA Grace Blackwell NVLink72. From ChatGPT to o1 to Astra in 4 years.”
“AGI has arrived. Congratulations @OpenAI team. 400K GPUs coming online next,” he added.
What is AGI?
AGI generally refers to a cutting-edge AI-powered system, which is capable of performing a plenty of intellectual tasks with precision and perfection at a level comparable to humans.
As per Huang, OpenAI used around 100,000 Nvidia AI computing systems to train the recently launched GPT-Astra model. His initial post reportedly cited 300,000 systems before being deleted and replaced with the lower figure. Nvidia has yet to explain the change.
The claim comes amid increasing demand for Nvidia's AI-powered chips to integrate in computing infrastructure.
Additionally, Huang underscored the significance of computing power to the AI industry, stating, “Now, compute is revenue.”
Despite Huang’s comments, OpenAI executives have stopped short of publicly declaring that AGI has arrived. The company’s president, Greg Brockman, has also avoided making such a definitive claim.
The relationship between Nvidia and OpenAI has attracted additional attention after receiving reports that a previously announced $100 billion partnership between the companies was never finalized.
It is pertinent to mention Huang stated another 400,000 Nvidia chips are likely to become available in the near future, possibly highlighting the computing resources needed to develop the increasingly cutting-edge AI systems.
