Altman Predicts AI Superintelligence by 2030
- •Sam Altman predicts AI will reach superintelligence and surpass human capabilities by 2030.
- •Anthropic CEO Dario Amodei expects AI to outperform humans in most tasks by 2027.
- •OpenAI, Oracle, and SoftBank are building an 800-acre data center in Texas to support Stargate.
Sam Altman, CEO of OpenAI, projects that artificial intelligence will surpass human intelligence by 2030. In an interview with the German newspaper Die Welt in late 2025, Altman stated he would be surprised if models by the end of the decade lacked extraordinary capabilities exceeding those of humans. He further estimated that AI could eventually handle 30% to 40% of economic tasks. Other industry leaders offer more aggressive timelines: Anthropic CEO Dario Amodei predicts AI will outperform humans in nearly every area by 2027, and Elon Musk has suggested that AI and robotics will eventually make money irrelevant.
Altman emphasized that despite rapid technical advancement, human-centric qualities regarding interpersonal utility will remain vital. He noted that ChatGPT, currently the world's fifth most popular website, faces operational constraints due to high energy requirements and substantial data center infrastructure needs. While OpenAI has developed models more advanced than GPT-5, public deployment remains limited by existing infrastructure capacity.
To address these demands, OpenAI is collaborating with Oracle and SoftBank on an 800-acre data center complex in Abilene, Texas. This site serves as a flagship location for the Stargate infrastructure project, which recently received federal approval. According to Altman, even this massive scale will not be enough to fully meet current demand. He anticipates that within a couple of years, AI will likely achieve scientific breakthroughs unattainable by humans, a milestone he characterizes as the emergence of superintelligence.