The world’s richest man believes artificial intelligence is progressing at a blistering pace and will surpass the capabilities of humans next year. In a remarkable prediction during a recent interview, Elon Musk stated his “guess” that AI models will become smarter than any individual person by the end of next year.

“My guess is that we’ll have AI that is smarter than any one human probably around the end of next year,” Musk declared in a conversation streamed on X/Twitter with Nicolai Tangen, CEO of Norges Bank Investment Management. The billionaire entrepreneur, who helms companies like Tesla, SpaceX, and his AI startup xAI, didn’t mince words when it came to the future supremacy of artificial general intelligence (AGI) – AI systems that can outperform humans across a wide range of tasks.

Musk’s bold proclamation sets a much nearer timeline than his previous projections and those of other AI experts. Just last year, he had forecasted that “full” AGI would arrive by 2029. Demis Hassabis, co-founder of Google‘s pioneering AI research company DeepMind, had estimated AGI could be achieved by 2030. However, Musk now believes that within the next five years, AI’s capabilities will likely exceed the collective intelligence of all humans combined.

The serial entrepreneur’s confidence stems from the remarkable breakthroughs in AI witnessed over the past 18 months. The launch of sophisticated video generation tools and increasingly capable conversational AI assistants like ChatGPT have catapulted the field’s progress far beyond expectations. Musk acknowledged that these developments have pushed the frontier of AI forward at an accelerated rate.

However, this rapid advancement has been hampered by bottlenecks in the supply chain for essential hardware components, particularly the high-performance microchips produced by NVIDIA that are crucial for training and running AI models. “Last year it was chip constrained… people could not get enough NVIDIA chips,” Musk explained. “This year it’s transitioning to a voltage transformer supply. In a year or two [the constraint is] just electricity supply.“

The world’s reliance on AI is quickly outpacing the capacity of existing infrastructure, straining not just the semiconductor industry but also power grids. As AI models become increasingly complex and demanding, ensuring an adequate supply of computing power and electricity will be a significant challenge.

Musk’s enthusiasm for AI development is a stark contrast from his stance just a year ago when he was among the prominent voices calling for a pause in advanced AI research. Citing “profound risks to society and humanity,” he had urged an immediate halt to training any system more powerful than OpenAI‘s GPT-4, then the market-leading AI model.

Now, the tables have turned. Musk revealed that his own AI company, xAI, is currently training a second version of its model called “Grok,” which he believes “should be better than GPT-4” upon its expected completion in May. And that’s just the beginning – he hinted that a new model orders of magnitude more powerful would follow soon after.

The billionaire has shifted significant time and resources to xAI over the past year, attempting to raise billions of dollars from investors in the US, Middle East, and Hong Kong. His goal? To build a formidable AI company that can compete with OpenAI, the firm he co-founded in 2015 but acrimoniously parted ways with in 2018 over disagreements about its direction.

Musk’s relationship with OpenAI has only grown more contentious. In March, he sued the company and its CEO, Sam Altman, alleging breach of contract and claiming they had compromised OpenAI’s stated mission of building AI for the benefit of humanity – an allegation the company has vehemently denied.