Estimates for Achieving Artificial General Intelligence

Artificial General Intelligence (AGI)—AI that matches or exceeds human-level intelligence across a wide range of tasks—remains a moving target. Estimates for when we might reach it vary sharply among researchers, company leaders, and forecasters. Definitions of AGI itself differ, and today’s most advanced systems still lack the kind of true generalization, autonomy, and understanding that would qualify as AGI in the strict sense.¹² Even so, rapid progress in large language models and compute has led many to shorten their timelines. Here’s a concise overview of where the debate stands.

Expert predictions: industry leaders

High-profile AI leaders often give relatively near-term estimates, usually tied to their own labs’ roadmaps:

• Elon Musk (xAI) has predicted AGI by 2026, revising earlier 2025 forecasts.³⁴
• Sam Altman (OpenAI) has pointed to AI “research interns” by 2026 and full automated researchers by 2028—steps that could lead toward AGI.⁵⁶
• Dario Amodei (Anthropic) has suggested superhuman AI in 1–2 years; Demis Hassabis (DeepMind) has given 5–10 years or roughly a 50% chance by 2030.⁷⁸

These views are optimistic on timing and reflect the incentives and information of people building the systems. They are not a scientific consensus.

Survey averages: researchers and forecasters

Broader expert surveys and prediction markets paint a more spread-out picture:

• AI researchers: Median timelines often put a 50% chance of AGI by 2040–2050, and around a 90% chance by 2075.¹
• Forecasters and prediction markets (e.g. Metaculus, updated in 2026): 50% by around 2041 in some aggregates; some forecasts cluster near 2031 for a 50% chance.¹⁹
• Entrepreneurs: Median estimates often sit around 2030.¹

Key challenges: why timelines are uncertain

Scaling current models has produced impressive gains, but it has not yet closed core gaps: robust reasoning, long-horizon planning, and reliable real-world adaptability. Many analysts argue that additional breakthroughs beyond today’s LLMs will be needed for AGI.¹⁰¹¹¹

At the same time, aggregate timelines have shortened. Pre-2023, some median estimates for a 50% chance of AGI sat around 2060; more recent surveys and markets have pulled that forward. Skeptics, including in communities like Reddit, still argue that 8–15+ years or more could remain, and that hype is compressing estimates.¹⁰¹¹

For the public and policymakers, the important point is that estimates are spread across decades and depend heavily on definition and methodology. Planning for risk and governance should not rely on any single forecast—whether very soon or far in the future.