AI will make a Nobel-level discovery within a year, Anthropic co-founder predicts

Anthropic founder Jack Clark has laid out a series of ambitious predictions about the development of artificial intelligence. By his estimate, within a year AI systems will make a discovery worthy of a Nobel Prize, and within two years bipedal robots will help workers on construction sites and in industrial plants.

When AI becomes a scientific partner

Clark describes a "dizzying sense of progress" in AI technology. The first key milestone is a Nobel-level discovery. AI will work side by side with a human scientist, analyzing experimental data, proposing hypotheses and suggesting directions for further research at a fundamentally new level of understanding.

This implies that AI systems have advanced enough to drive science itself forward, rather than merely assisting with the tedious processing of large volumes of data. Such a partnership between human and machine could speed up breakthroughs in physics, chemistry, biology and medicine.

Robots leave the lab

Alongside scientific discoveries, Anthropic expects bipedal robots to show up in workplaces within two years. This is not about factories, where a robotic arm can operate in a tightly controlled environment. It is about practical help for electricians, plumbers, builders and industrial workers in real-world conditions.

Bipedal platforms will be able to do what machines do poorly or what is dangerous for humans:

• Work at height (roofs, building facades)
• Assembling and delivering heavy loads
• Repetitive operations in industry
• Help in accidents and emergencies
• Work in extreme conditions (high temperatures, chemicals)

This sounds futuristic, but large companies are already actively funding such projects in 2025-2026.

Companies run by AI without people in 18 months

Clark's most radical prediction: companies run entirely by AI systems, without human involvement, will begin generating millions of dollars in revenue within 18 months. That means that as early as late 2027 we may see autonomous AI startups that build products and attract customers and funding on their own.

By the end of 2028, according to Anthropic's founder, AI systems will be able not only to work without humans but also to design their successors: more advanced versions of themselves.

«This is not just a tool for humans that boosts their productivity.

This is becoming an agentic revolution, where AIs act independently»

What this means

Clark is candid: along with "dizzying" progress, the technology will bring "deep and significant changes" to society. AI carries both incredible opportunities for science, the economy and solving hard problems, and real risks to employment, concentration of power and the geopolitical balance.

The timelines Clark gives are very aggressive. If even some of them come true, it will reshape the economy and science in literally one to two years. That means companies, investors and policymakers need to start preparing for change now.