McKinsey: Generative AI Could Add $4.4 Trillion a Year

McKinsey Global Institute has put a hard number on a question that's been floating around every AI conference for months: what is all this actually worth? The consultancy estimates that generative AI could add between $2.6 trillion and $4.4 trillion a year to the global economy, according to its report The Economic Potential of Generative AI: The Next Productivity Frontier, published earlier this month.

To put the figure in perspective, McKinsey compares it to the United Kingdom's GDP, one of the largest economies on the planet. In other words, the firm argues that the value this technology adds, at the high end of its estimate, would be equivalent to creating a British-sized economy every single year — purely from the productivity gains unlocked by generative AI.

Where the number comes from

The calculation didn't come out of nowhere. McKinsey has spent years analyzing the economic impact of "traditional" AI — prediction, optimization and classification systems trained for specific tasks. What's new in this report is that it isolates the additional effect of generative models, which can produce text, images or code from natural-language prompts, like the ones ChatGPT and Bard have popularized in recent months.

According to the consultancy, generative AI doesn't replace that prior impact — it amplifies it. The report estimates this technology could boost the total economic effect already attributed to AI as a whole by 15% to 40%, precisely because it automates tasks long considered resistant to automation: writing reports, coding, generating designs or holding conversations with customers.

Which sectors stand to gain the most

McKinsey identifies banking, tech, life sciences (pharma and biotech) and retail as the industries where the economic impact would be largest in absolute terms, both because of the volume of repetitive cognitive tasks they handle and their capacity to quickly invest in these tools.

The report isn't limited to value generated by new products or services. A large share of the figure comes from productivity gains — doing the same work with fewer hours of human labor. This is where McKinsey introduces the statistic that's circulated most widely among labor-market analysts: the firm calculates that generative AI has the technical potential to automate activities that currently take up between 60% and 70% of employees' time, up from the 50% the consultancy had estimated before these models existed.

Technical potential, not economic certainty

That figure needs to be read carefully: McKinsey is talking about the technical potential for automation, not a prediction that these tasks will vanish tomorrow. Between what technology can do and what companies actually adopt lie years of investment, workforce training, regulatory frameworks and, in more than a few cases, organizational resistance. The history of industrial automation itself shows that the gap between technical capability and real-world deployment can take a decade or more to close.

The $2.6 trillion to $4.4 trillion range also reflects that uncertainty: it's not a single figure but a wide interval, contingent on the pace of adoption, the regulations each country develops, and companies' ability to redesign processes around these tools — something McKinsey explicitly acknowledges as the main source of variation in its model.

Why it matters now

The timing of the report is no accident. Since ChatGPT's public launch in November 2022, investment in generative AI startups has surged, and Big Tech players — Microsoft, Google, Amazon — have announced plans to integrate these models into virtually their entire product lineups. A report of this scale, carried by a consultancy with McKinsey's credibility among executives and boards, serves as ammunition for those pushing to accelerate corporate adoption of these tools.

At the same time, the figure feeds the jobs debate that's been simmering in the sector for months. If a significant share of today's cognitive work is automatable, the question left open — one the report itself doesn't answer — is what happens to the people currently doing that work: whether they're reassigned to new roles, whether positions are cut, or whether the net outcome depends, as in previous waves of technological change, on political and business decisions that have yet to be made.

What McKinsey has put on the table is an estimate of potential value, not a roadmap for how it gets distributed. That part of the story is still being written, and it will hinge less on the technical capability of the models than on the decisions governments and companies make in the years ahead.