When Will AGI Arrive? Even AI CEOs Don't Agree

Hundreds of billions of dollars have flooded into AI development with one ultimate goal: artificial general intelligence. Yet the very people building it can't agree when—or even if—it'll arrive.

Elon Musk and Anthropic'sDario Amodei predict AGI by late 2026. Google DeepMind'sDemis Hassabis says we might wait another decade. Most intriguingly, OpenAI'sSam Altman claims "AGI kind of went whooshing by" already—now he's chasing "superintelligence" instead.

Just two years ago, these same leaders largely agreed AGI would emerge in the late 2020s. Today, not only are their timelines scattered, but their fundamental understanding of what AGI even means has completely diverged.

The Slippery Definition Problem

This confusion isn't entirely surprising. AGI has always been a "squishy" concept. Alan Turing's1950 test for machine intelligence has been passed multiple times by programs nobody would call truly intelligent—they just happened to be convincing enough to fool some humans on that particular benchmark.

The modern notion of AGI emerged in the early 2000s not as a specific threshold but as a field of study—developing generally intelligent algorithms rather than narrow, targeted ones. There was never agreement on how to actually test for such general abilities in machines.

OpenAI popularized the term when it launched ChatGPT in late 2022, with its founding mission to ensure AGI "benefits all of humanity." The company's former chief scientist Ilya Sutskever famously encouraged employees to "feel the AGI." This ambiguity became a marketing goldmine—companies could sell "intelligence" without defining it, leading to endless questionable ads claiming chatbots make ideal travel agents.

Reality Check: AI Hits Its Limits

But the case for imminent AGI is getting harder to make. Large language models excel at technical tasks like software engineering and competition math problems, yet they still struggle with seemingly trivial challenges—drawing clocks or completing simple logic puzzles.

Throughout 2023, each new generation of AI models yielded only marginal improvements rather than dramatic leaps forward on standard benchmarks. And those benchmarks are highly gameable—it's unclear whether labs are measuring genuine general capabilities or just optimizing for specific tests.

The industry is quietly acknowledging this reality. Microsoft'sSatya Nadella now describes AI as "a tool," with success measured not by achieving AGI but by driving 10% global GDP growth. White House AI adviser Sriram Krishnan recently called AI "a very useful technology" that "has nothing to do with 'general intelligence.'" Even AI researchers like Arvind Narayanan and Sayash Kapoor are calling chatbots a "normal technology"—useful but hardly revolutionary.

The Product Pivot

This shift is playing out in real time. Around San Francisco, billboards now advertise AI accounting tools instead of world-changing intelligence. Google DeepMind touts how its latest model Gemini 3 can improve your "shopping experience" and organize your inbox. Both OpenAI and Anthropic brag about making corporate employees more efficient at "writing sales emails."

OpenAI kicked off 2026 by announcing ads in ChatGPT. The company's applications CEO Fidji Simo recently wrote that the winning AI company will be the one that turns "frontier research into products." The firm has launched a web browser, social media apps, and numerous other AI tools over recent months.

Strategic Positioning Through Predictions

Here's where it gets interesting: each CEO's AGI timeline aligns perfectly with their company's strategic needs.

OpenAI, having lost its technical lead, now claims AGI has already arrived—justifying its pivot to product development. Anthropic'sAmodei maintains that "powerful AI" is imminent, reinforcing his company's reputation as the responsible, anxious alternative that enterprise customers prefer.

Google'sHassabis can afford longer timelines because the company has massive revenue streams supporting patient development. "I don't think AGI should be turned into a marketing term for commercial gain," he said Thursday.

Musk, true to form, makes grand promises while pivoting his entire empire toward AI. This week, Tesla announced it's abandoning major car lines to produce humanoid robots, accelerating its transformation from automaker to AI company. The company recently committed $2 billion to xAI, and Musk is reportedly considering merging SpaceX with his AI venture.