The AI–Robotics Productivity Revolution 

Posted on September 11th, 2026

Advances in artificial intelligence and robotics have the potential to drive a step-change in productivity over the coming decades.

Unlike previous waves of change, which largely automated routine manual tasks, this technology is increasingly capable of performing cognitive and professional functions: analysing data, drafting documents, diagnosing conditions and supporting complex decision-making. The result could be a broad-based transformation of work across much of the economy.

Estimates from institutions such as the OECD13, IMF14 and World Economic Forum15 suggest that a significant share of current tasks—across manufacturing, logistics, professional services and parts of healthcare—could be automated or augmented by AI systems.

For businesses, this offers the prospect of higher output with fewer inputs, supporting margins and enabling new forms of scale. For economies, it raises the possibility of a sustained uplift in productivity growth after a prolonged period of stagnation, but also significant labour-market disruption.

The range of views on how far this will go is wide. At one end of the spectrum are more cautious assessments, in which AI reshapes tasks rather than eliminates jobs wholesale. At the other are more dramatic claims. Some economists have argued that advanced AI could eventually reduce the demand for human labour across large parts of the economy, raising the prospect of a “world without work.”

As with past technological revolutions, the eventual outcome will depend not just on capability, but on adoption, regulation, social acceptance – and investment.

For someone currently in their 20s, the lived experience of this transition may be one of continual adaptation over a long working life. Roles are likely to evolve repeatedly. Tasks that define a job today may be automated within a decade, requiring workers to shift towards areas where human capabilities—judgment, empathy, creativity and interpersonal skills—remain valuable. Careers may become more modular, combining periods of work, retraining and partial retirement.

This will place new demands on education systems. The traditional model—front-loaded education followed by decades of work—may prove inadequate. Instead, learning is likely to become continuous, with reskilling and mid-career training becoming routine. The ability to adapt may become more valuable than any specific body of knowledge.

Yet there are limits to automation. Human preferences matter. Experience to date suggests that, even where technology can substitute for human interaction, it is not always the preferred outcome. Japan’s efforts to deploy robots in elderly care, for example, have delivered mixed results, with many users continuing to favour human carers despite labour shortages.

Distributional effects remain uncertain. Productivity gains do not automatically translate into broadly shared prosperity, and there is a risk that benefits accrue disproportionately to a select few. This could widen inequalities and prompt policy responses, from education reform to more interventionist approaches to income and labour markets.

For investors, there will be opportunities in the companies developing and deploying AI and robotics—across semiconductors, software and automation. But the enabling infrastructure may prove just as important. Data centres, high-capacity networks and power systems are becoming foundational to the AI economy. These assets could take on a role analogous to that of railways in the 19th century. There is potential for vertical integration as technology companies develop and manage their own energy networks.

Despite uncertainties, in the next half-century, the interaction between human workers and intelligent machines will be one of the defining features of economic life—reshaping not just productivity, but how individuals work, learn and earn across their longer lives.

Back to all News & Insights