Google DeepMind chief Demis Hassabis is sharpening his timeline for artificial general intelligence (AGI), warning that society has only a few years to prepare for systems he says could reshape civilization.

At Google’s I/O developer conference in mid‑May 2026, Hassabis told attendees that humanity is standing "on the foothills of the singularity," describing a looming turning point where AI could outpace human intelligence and begin improving itself.¹ He pointed to the rapid rise of "agentic" AI systems that can build tools and products for people, noting that he now uses AI to create mini video games in hours that once would have taken months.¹

In follow‑up remarks, Hassabis said he expects AGI—machines roughly as intelligent as humans—to arrive "as soon as 2030," and possibly as early as 2029, arguing that recent progress shows the industry has likely found the right technical path.¹² He predicted AI’s impact would be "100 times" that of the Industrial Revolution and cast the coming "agentic era" over the next year as a "practice run" or stress test for far more powerful systems.¹²

By late May, speaking with Axios, Hassabis emphasized that this accelerated timeline is meant to jolt governments, economists and the public into action, saying society has "only a few years left" to prepare for AGI and calling current safety efforts a "good warning shot" that still need to be "accelerated."² He described himself as a "cautious optimist," arguing that while AI poses risks, humans can harness it to solve hard problems in science and healthcare rather than lose control to machines.¹

Meanwhile, outside Google, other AI leaders are underscoring how fast the field is moving. OpenAI’s Sam Altman recently highlighted that "a general-purpose model solved a major open problem in mathematics," calling it "a kinda big milestone" that leaves him with "complicated feelings."Sam Altman (sama)³ OpenAI cofounder Greg Brockman similarly described the math breakthrough as "a milestone in new knowledge generation by AI" and said it is "very exciting" to imagine comparable advances across science, even as the pace makes it "very hard to sleep."Greg Brockman (gdb)⁴

These parallel developments—Google’s push toward AGI and OpenAI’s problem‑solving milestones—are feeding a shared sense among AI leaders that transformative capabilities may arrive sooner than many policymakers and institutions are prepared to handle.