When Will Superintelligence Arrive? What AI Leaders Predict

When will AI become smarter than the smartest people, not just at one task but at almost everything? The honest answer is that no one knows. But the people building AI have made public predictions, and thousands of AI researchers have answered surveys. Here is what they said, when they said it, and why their answers are so far apart.

One point first: most forecasts are about artificial general intelligence (AGI), AI roughly as capable as people across most thinking tasks. Superintelligence goes further: AI far beyond the best human minds in almost every field. For the difference, read Super Intelligence vs Artificial Intelligence, Explained and What Is Superintelligence?

What AI company leaders predict

Sam Altman, OpenAI

In his essay “The Intelligence Age” (September 2024), OpenAI’s chief executive wrote: “It is possible that we will have superintelligence in a few thousand days”. A few thousand days is roughly 5 to 15 years. In January 2025 he went further: “We are now confident we know how to build AGI as we have traditionally understood it,” adding that OpenAI was “beginning to turn our aim beyond that, to superintelligence in the true sense of the word.”

At the India AI Impact Summit in February 2026, Forbes India reported that Altman said the world could be only a couple of years away from early versions of true superintelligence, and that by the end of 2028 more intellectual capacity could sit inside data centers than in human minds.

Dario Amodei, Anthropic

Anthropic’s chief executive writes about “powerful AI”, which he describes as smarter than a Nobel Prize winner across most relevant fields. In October 2024 he wrote that it “could come as early as 2026, though there are also ways it could take much longer.”

In January 2026 he repeated that powerful AI “could be as little as 1–2 years away, although it could also be considerably further out.” In the same essay he argued that if the current pace of improvement continues, “it cannot possibly be more than a few years before AI is better than humans at essentially everything.”

Demis Hassabis, Google DeepMind

The Nobel Prize-winning head of Google DeepMind has been more cautious. In March 2025 he told reporters that “over the next five to 10 years, a lot of those capabilities will start coming to the fore and we’ll start moving towards what we call artificial general intelligence,” NBC News reported. At Google I/O in May 2026 he said: “When we look back at this time, I think we all realize that we were standing in the foothills of the singularity.” Sherwood News reported that his expected window had narrowed to 2029 to 2030.

Elon Musk

In March 2024, Musk wrote on X that “AI will probably be smarter than any single human next year. By 2029, AI is probably smarter than all humans combined.” His AI company, xAI, is now part of SpaceX as SpaceXAI.

Mark Zuckerberg and Ilya Sutskever

Meta’s chief executive wrote in July 2025 that “developing superintelligence is now in sight,” as the company built Meta Superintelligence Labs. Safe Superintelligence Inc., co-founded by former OpenAI chief scientist Ilya Sutskever, opens its website with the line “Superintelligence is within reach.”

What AI researchers predict

Company leaders are a small group. The largest regular survey of AI researchers, run by AI Impacts, asks people who publish at leading AI conferences when “unaided machines can accomplish every task better and more cheaply than human workers”. Their combined forecast for a 50% chance of that has moved closer with every survey:

Survey year	50% chance of machines beating humans at every task by
2016	2061
2022	2059
2023 (2,778 researchers)	2047
2024 (1,580 researchers)	2042

In the 2023 survey, the same researchers put a 10% chance on it happening by 2027. When the question was asked about every job being fully automatable, the answers were much later: a 50% chance only by 2116. The way a question is worded clearly changes the answer.

Not everyone agrees it is close

Some leading researchers doubt that today’s approach will lead to human-level AI at all. Yann LeCun, a Turing Award winner who was Meta’s chief AI scientist until the end of 2025, has argued that large language models like those behind ChatGPT will not reach human intelligence, because they lack reasoning, planning, lasting memory and an understanding of the physical world. He left Meta to found AMI Labs, a company building a different kind of AI based on “world models”.

Why the predictions are so far apart

• Different goals. “Powerful AI”, “AGI” and “superintelligence” mean different things, and even the surveys get different answers when they change the wording.
• Different views on progress. Leaders like Amodei expect the steady improvement of the past decade to continue. Skeptics expect current methods to hit limits that need new ideas.
• Different vantage points. Company leaders see systems that are not public yet. Critics point out that they also have commercial reasons to sound optimistic, for example when raising money or hiring.
• Real uncertainty. Progress depends on computing power, energy, data and research breakthroughs that no one can schedule.

Predictions at a glance

Who	What they said	When
Sam Altman (OpenAI)	Superintelligence possible “in a few thousand days”	September 2024
Dario Amodei (Anthropic)	Powerful AI “could be as little as 1–2 years away”	January 2026
Demis Hassabis (Google DeepMind)	AGI over “the next five to 10 years”; later reported as 2029 to 2030	March 2025, May 2026
Elon Musk	AI “smarter than all humans combined” by 2029	March 2024
AI researchers (survey)	50% chance by 2042	Published September 2026

How to read a superintelligence prediction

When you see a new forecast, ask three questions. What exactly is predicted: human-level AI, AI that beats humans at everything, or AI far beyond that? When was it said: timelines shift quickly, so a two-year-old prediction may already be out of date. How confident is it: most of the leaders quoted here added hedges such as “could” and “it may take longer”, which headlines often drop.

We will update this page as new predictions and surveys are published.