On Monday (April 8) local time, Tesla CEO Elon Musk predicted that by next year or 2026, artificial intelligence (AI) may be smarter than the smartest humans.

Musk was interviewed by Nicolai Tangen, CEO of the Norwegian Sovereign Wealth Fund, on the X (formerly Twitter) voice space on Monday.

When asked about the development timeline of General Artificial Intelligence (AGI), Musk said, “If you define AGI as smarter than the smartest humans, I think it will probably be next year (realized), that is, within two years.”

Over the past 18 months, the development of AI has made a series of breakthroughs, including more powerful chatbots and video generation tools. These have all pushed AI forward faster than expected, and people are beginning to imagine when AGI will be realized.

Demis Hassabis, co-founder of Google's AI company DeepMind, predicted earlier this year that AGI is expected to be realized in 2030.

As chips to train AI models are in short supply, the pace of AI development has slowed down. Musk revealed that the lack of advanced chips hindered the training of the second version of the Grok model.

Grok is a chatbot launched by Musk's AI company XAI in November last year. Musk founded XAI to compete with Open AI.

According to Musk, XAI is training the 2.0 version of Grok. This version of the model is expected to be more powerful than Open AI's GPT-4, and the relevant training work is expected to be completed by May.

Musk said that training the Grok 2 model requires about 20,000 Nvidia H100 GPUs, while the Grok 3 model and higher will require around 100,000 Nvidia H100 GPUs.

However, Musk added that although the shortage of chips has been a major constraint on the development of AI so far, power supply will be critical in the next year or two.

Editor/Jeffrey