McKinsey & Company (QuantumBlack) · Survey

The State of AI in 2026

The annual benchmark: AI use keeps spreading, but enterprise-level profit impact has barely moved in two years.

Published
Reading time
About 25 min
Authors
Dan Tinkoff, Lieven Van der Veken, Michael Chui, Tara Balakrishnan
Read the original

McKinsey reuses this URL every year. Cite it as “The state of AI in 2026 (August 2026)”.

What is it about?

McKinsey's global survey of 1,719 respondents in 97 countries, fielded in May and June 2026, finds AI spreading faster and wider than ever. Among organizations with more than $1 billion in revenue, 40% of respondents say they are already scaling AI agents.

The headline tension is the gap between individual and enterprise gains. Eighty percent say AI improved their own productivity, yet only 37% attribute any EBIT impact to it. The roughly 6% of “high performers” are separated less by budget than by behaviour: they redesign workflows end to end, aim AI at growth and innovation rather than efficiency alone, and treat AI operating cost as a design constraint.

Key numbers

  • 80% say AI improved their individual productivity; 37% attribute any EBIT impact to AI.
  • About 6% qualify as high performers (5% or more of EBIT from AI), flat versus 2025.
  • 32% decided against buying at least one software product because agentic coding tools let them build it in-house.
  • Around one in five say AI operating costs, including tokens, already constrain their use of AI.
  • High performers hit cost limits on coding agents three times as often as everyone else (18% vs 6%).

Why read it

  • It is the most-quoted annual dataset on enterprise AI. Your board and CFO will cite it, so you should know exactly what it says.
  • The tokenomics section is the clearest explanation of why AI bills grow even while per-token prices fall.
  • The build-versus-buy finding should change how you review SaaS renewals.

Who should read it

  • CIOs and AI leads preparing a board update
  • CFOs building or challenging an AI business case
  • CEOs who want a second look at their AI plans

About the author: Michael Chui

Senior fellow, QuantumBlack, AI by McKinsey

One of McKinsey's longest-standing voices on technology and the economy, at the firm since 2002. He co-authored its widely quoted 2023 estimate that generative AI could add $2.6–4.4 trillion a year to the global economy and co-hosts the McKinsey Global Institute podcast Forward Thinking.

Before consulting he was the first CIO of the city of Bloomington, Indiana, where he moved city IT toward open-source software, and he founded a cooperative nonprofit internet provider. His PhD dissertation on web search is titled “I Still Haven't Found What I'm Looking For”.

Read it carefully: common misquotes

  • Every figure is a share of respondents, not of companies, and all of it is self-reported.
  • The sample is executive-heavy. “80% more productive” is mostly leaders describing themselves.
  • “40% scaling agents” applies only to companies above $1 billion in revenue, not to the whole sample.

How it shapes the CIO's AI Playbook

Notes by Pavel Veselov, last reviewed . Summaries are my own; quotes and figures belong to the original authors.

All 11 sources in the CIO's AI Library