Google DeepMind CEO Demis Hassabis delivered a stark message this week. Artificial general intelligence stands just a few years from reality. Maybe 2030, plus or minus a year. The implications stretch far beyond faster software or smarter assistants. This technology, he argues, marks the start of something far larger.
“Maybe 2030, plus or minus a year, which is astounding to think, really. I think that will be such an enormous transformative technology; it’s gonna effectively be a new human era,” Hassabis said during a fireside chat at the Stanford Graduate School of Business. The remarks, highlighted by Business Insider, underscore a narrowing window for society to absorb what comes next. Short timeline. Enormous stakes.
Hassabis has refined his forecast in recent months. He now sees 2029 as possible. Agents that act autonomously serve as an early test. They offer a “practice run” for the deeper changes ahead. Speaking after his appearance at Google I/O, he told Axios that current systems already demonstrate real momentum. “We can see agents really happening now and imagine what they will be in another year, and how useful they’ll be.” The pace feels tangible. Industry insiders sense it too.
Yet urgency mixes with caution. Hassabis describes humanity as standing in the “foothills of the singularity.” Preparation time feels short. Economists, in his view, still underestimate the shift. “My economist friends, I feel, are still not taking this seriously enough. That needs to change.” Safety work must accelerate alongside capability gains. The next few years will shape whether the transition favors broad benefit or concentrated risk.
His optimism remains intact. Done right, AGI could deliver radical abundance. Cures for diseases. New energy sources. Materials that transform industries. Longer lifespans. Hassabis sketched this vision clearly in an earlier conversation. “The reason I’ve worked on AI and AGI my entire life is because I believe, if it’s done properly and responsibly, it will be the most beneficial technology ever invented,” he told TIME. The upside includes space travel and a new golden era of discovery. But only if stewardship matches ambition.
He quantifies the scale in striking terms. The Industrial Revolution altered economies and societies over a century. AGI could deliver ten times the impact and unfold ten times faster. “It’ll be 10 times bigger than the Industrial Revolution – and maybe 10 times faster,” Hassabis explained to The Guardian. Incredible productivity. Widespread prosperity. Yet the speed introduces friction. Job displacement. Energy demands. Questions about fairness in distribution. These issues demand attention now, not later.
Recent public appearances reinforce the message. At events tied to Google I/O and other forums, Hassabis has stressed that current agentic systems preview recursive improvement. Soft self-improvement already boosts engineer productivity through coding tools. Full recursive loops carry both gains and hazards. “There’ll be clear gains in terms of speed of your research. But there are also risks with that type of system,” he noted in the Axios discussion. Labs across the sector focus on these dynamics. Alignment between capability and control grows more pressing each quarter.
DeepMind’s own work supplies concrete examples. AlphaFold transformed biology. New world models generate consistent interactive environments. Multimodal systems interpret context with growing nuance. These advances feed into the larger arc toward general intelligence. Hassabis believes one or two additional breakthroughs may still be required. Large foundation models will likely remain central. Scaling alone might not suffice. Insight matters too. The blend of computation, architecture, and fresh ideas will decide the exact arrival.
Industry voices differ on timing. Some predict sooner. Others push the date back by decades. Hassabis critiques excessive certainty on all sides. “The future, in my view, is still to be written, but these next few years are going to be very critical as to which way that will go and how we collectively want that to look like.” Collective choice. Not inevitability. That distinction matters. Governments, companies, researchers, and citizens all hold pieces of the decision.
Calls for global coordination have grown. Hassabis has floated the idea of an international body, similar to the United Nations, to guide AGI development. Recent X discussions echo the tension. Sergey Brin reportedly pressed for faster delivery in a public exchange, while Hassabis reiterated the need for reliable, safe systems first. The exchange captured a core friction. Builders who grasp the technical risks often hold less sway over schedules than funders chasing deliverables. Laughter from the audience masked a serious point.
Post-scarcity enters the conversation as both promise and puzzle. If AI handles cognitive work at superhuman levels across domains, economic assumptions change. Goods and services could become far cheaper. Yet societies must decide how to allocate gains. What follows abundance? New pursuits. Scientific exploration on an unprecedented scale. Or fresh inequalities if access remains uneven. Hassabis acknowledges the unknowns. “Let’s say we get radical abundance… what happens next?” The question lingers.
His track record lends weight. Nobel recognition for AlphaFold. Decades spent studying neuroscience, games, and learning systems. DeepMind’s founding vision centered on general intelligence from the start. That consistency shows. He does not treat AGI as distant speculation. It forms the north star for current research priorities. Agents. World models. Reliable reasoning. Each step builds the foundation.
Critics worry about hype. Timelines have slipped before. Capabilities sometimes plateau. Hassabis counters that recent progress validates the path. Confidence has grown. The technical route looks clearer. Still, he avoids overclaiming. True AGI means systems that match or exceed human cognitive abilities across the board. Not narrow excellence. Broad competence. Measurable progress toward that bar remains an active research focus at DeepMind, including new cognitive frameworks for evaluation.
Energy demands add another layer. Training and running ever-larger models consume significant power. Data centers strain grids. Solutions may come from the very technology under development. Better materials. Optimized fusion pathways. AI-assisted design of efficient systems. The feedback loop could prove self-reinforcing. But only if risks around control and alignment stay managed.
Students received direct advice in the Stanford session. Lean into the technology. Adapt skills. Combine domain expertise with fluency in AI tools. Humanities and sciences both matter. The coming era will reward those who understand both human context and machine capability. Society at large needs the same message. Preparation cannot wait for perfect clarity. The window is closing.
Hassabis positions himself as a cautious optimist. Human ingenuity has solved hard problems before. Adaptability runs deep. Yet the speed and scale of this transition exceed prior shifts. “I’m a cautious optimist… I believe in human ingenuity. I think we’ll get this right,” he said. The bet rests on deliberate action. Safety research. Inclusive dialogue. Policies that distribute benefits widely. Technical work that prioritizes reliability from the outset.
Recent coverage captures evolving nuance. A May 2026 Axios report detailed his updated views post-Google I/O. Predictions tightened. Emphasis on agents as preview increased. Safety acceleration became more explicit. These threads connect to earlier statements. The 2025 Guardian interview laid out the tenfold comparison. TIME captured the long-term flourishing scenario. Together they paint a coherent picture. Opportunity sits beside responsibility. The technology will arrive. The shape of its influence depends on choices made today.
Debate continues on exact definitions. When does a system qualify as AGI? DeepMind has published frameworks using cognitive science to track progress. Benchmarks matter. Consensus across labs would help. Yet first-mover advantages could discourage full transparency. The competitive dynamic adds complexity. Public pressure for safety standards may prove as important as internal research.
And the clock ticks. Three to four years until possible arrival, by some readings of Hassabis’s words. Five years at the outside for initial forms. The difference feels small against historical timescales. Preparation spans regulation, education, infrastructure, ethics. Each domain requires focused effort. Delay in any area narrows options later.
Hassabis returns repeatedly to stewardship. Build it right. Deploy it responsibly. Maximize benefit. Minimize harm. The phrase appears across interviews because the concern does not fade. Technical mastery alone falls short. Societal readiness completes the equation. Whether that readiness materializes will define the new human era more than any single model release.
Observers on X and in recent analysis note the tension between acceleration and care. Public exchanges between tech leaders reveal it plainly. One side sees deliverables and timelines. The other sees unknown unknowns and irreversible steps. Both perspectives carry truth. Reconciling them determines the outcome.
The coming years will test institutions, companies, and individuals. Hassabis has sounded the alert. AGI is not science fiction. It approaches as engineering reality. The question now shifts from if to how. How societies absorb the change. How gains spread. How risks stay contained. Answers will emerge through action, not speculation. The era begins sooner than most expected. Readiness cannot lag behind.
WebProNews is an iEntry Publication