Google DeepMind's Demis Hassabis told WSJ Magazine he now expects artificial general intelligence (AGI) within two to three years, down from the three-to-five-year window he cited earlier in 2026. He defines AGI through an "Einstein test" and says his estimate keeps shifting closer to the present.
Demis Hassabis, the co-founder and chief executive of Google DeepMind, told The Wall Street Journal Magazine that he now expects artificial general intelligence (AGI) to arrive within two to three years. The interview was published on October 21, 2026, as part of the magazine's "Innovator in Tech" awards package, where Hassabis was named the 2026 honoree in technology.
Asked about the timeline, Hassabis said it was coming soon and that his estimate kept shifting closer to the present, according to the profile. The new window marks a tightening from the three-to-five-year range he had publicly cited earlier in 2026.
What Hassabis means by AGI
Hassabis defines AGI as human-level cognition and has described a specific benchmark he calls the "Einstein test." Under this test, a system would qualify as AGI if, trained only on human knowledge available up to 1901, it could independently produce the kind of breakthroughs Albert Einstein made in his 1905 "annus mirabilis" papers — covering the photoelectric effect, Brownian motion, special relativity and mass-energy equivalence.
In the WSJ Magazine profile, Hassabis also described his broader mission in two steps: first solving intelligence itself, then using that intelligence to address other major scientific problems. He pointed to physics research and cancer treatment as early applications he hopes to pursue once more advanced AI systems are available.
A timeline that has moved during 2026
Hassabis's public statements about AGI's arrival shifted over the course of 2026. At Google's I/O developer conference on May 20, 2026, he said AI had reached what he called the foothills of the singularity, clarifying that he meant the current era rather than a single moment of transformation. At that event he also said agents and coding tools had begun working effectively during the year, describing it as an early stage of practical AI use.
Hassabis's AGI timeline statements through 2026
When
Stated timeline / framing
Context
Earlier 2026
Three to five years
Public estimate cited before the WSJ Magazine interview
May 20, 2026
"Foothills of the singularity"
Remarks closing Google I/O
October 21, 2026
Two to three years
WSJ Magazine "Innovator in Tech" profile
Isomorphic Labs and the regulation debate
The WSJ Magazine profile also touched on Isomorphic Labs, a drug-discovery company founded by Hassabis that applies AlphaFold-style AI models to drug discovery. The company was reported to be in discussions to raise more than $2 billion in funding at a valuation of roughly $40 billion to $50 billion.
On the question of AI oversight, the profile described Hassabis arguing against premature regulation of the technology, a position consistent with his push during 2026 for industry development to continue while governance frameworks are worked out alongside it.
Why Hassabis's view carries weight
Hassabis leads Google DeepMind's AI research. In October 2024, he and DeepMind colleague John Jumper shared half of the Nobel Prize in Chemistry for AlphaFold, an AI model that predicts three-dimensional protein structures from genetic sequences; the other half of that prize went to David Baker. The Nobel committee at the time described the achievement as fulfilling a long-pursued goal in protein science.
I'm updating nearer and nearer.
— Demis Hassabis, to WSJ Magazine
Because AGI has no single agreed scientific definition, Hassabis's timeline reflects his own benchmark — the Einstein test — rather than an industry-wide standard. Other AI leaders have offered different definitions and timelines, which the sources reviewed for this article did not detail in comparable terms.
Questions
What did Demis Hassabis say about AGI's arrival?
He told WSJ Magazine, published October 21, 2026, that he expects artificial general intelligence within two to three years, saying his estimate keeps moving closer to the present.
How does Hassabis define AGI?
He uses an "Einstein test": whether a system trained only on knowledge available before 1901 could independently produce Einstein's 1905 breakthroughs, including special relativity and the photoelectric effect.
Has Hassabis's AGI timeline changed during 2026?
Yes. Earlier in 2026 he cited a three-to-five-year window; by May 2026 he said AI was at the "foothills of the singularity"; by October 2026 he narrowed the estimate to two to three years.
What is Isomorphic Labs and why is it mentioned?
Isomorphic Labs is a drug-discovery company founded by Hassabis that applies AlphaFold-style AI to drug discovery; the WSJ Magazine profile reported it was seeking more than $2 billion in funding at a $40–50 billion valuation.
What is Hassabis known for besides this prediction?
He leads Google DeepMind's AI research and in 2024 shared the Nobel Prize in Chemistry with John Jumper for AlphaFold, an AI system that predicts protein structures.
This piece was prepared with artificial intelligence and passed an automated source check before publication. If you see an error, ask for a correction. Request a correction or reply · How we make this
Neural networks learn by making a prediction, measuring how wrong it was with a loss function, and then adjusting internal values called weights backward through the network — a process called backpropagation paired with an optimisation method called gradient descent. The same mechanics, scaled to billions of parameters, underlie how systems such as ChatGPT and Gemini are first trained and then adapted to specific tasks through fine-tuning.
AI image tools like Midjourney and Stable Diffusion build pictures by learning to reverse a noise process, while music tools like Suno and Udio compress audio into tokens and predict them step by step, guided by a text prompt. Whether either process counts as genuine creativity is unresolved: researchers are split between calling it pattern-based "remixing" and crediting it with measurable creative performance.
Chinese AI lab Moonshot says it removed cyber-related training data from its new Kimi K3.1 Pro model so the open-weight system would not push the "frontier" of offensive hacking capability. The move, announced Saturday, October 24, 2026, comes as governments and security researchers warn that increasingly capable AI could make sophisticated cyberattacks cheaper and easier to carry out.
AI text detectors do not prove who wrote something; they estimate a probability by measuring statistical patterns such as word predictability and sentence-length variation. A detector called Pangram 4.1 recently attributed 99% of a €20,000 prize-winning Spanish-language novel to artificial intelligence, prompting its publisher to suspend promotional events while it investigates.
