Cursor launches CFO Council: 88% of companies have adopted AI, but only 39% see EBIT growth

Cursor has announced the Cursor CFO Council, a working group of chief financial officers that will develop a common approach to measuring returns on corporate AI spending. Worldwide, that spending reached $1.5 trillion in 2025.

Why is it hard for companies to measure AI returns?

A recent McKinsey study found that 88% of organizations have already adopted AI in at least one business function, but only 39% can link those investments to EBIT growth at the company level. AI spending has grown from a line item for experimental pilots into a major recurring operating expense, yet no established discipline for measuring its effectiveness has emerged. That is the problem the new council is meant to solve.

• Global AI spending reached $1.5 trillion in 2025
• 88% of organizations have adopted AI in at least one function (McKinsey data)
• Only 39% of companies can trace AI investments to EBIT growth
• The Cursor CFO Council will meet quarterly in cities around the world

How are AI returns distributed across companies?

A BCG analysis based on Cursor data found that companies in the top quintile by token usage posted median annual revenue growth of 16.5%, compared with just 5.1% for companies in the bottom quintile. A separate study found that after major model updates in late 2025, employees sent 44% more messages to AI agents per week, and 68% more on complex tasks.

Why are the gains so uneven?

Cursor's own Developer Habits Report found an even sharper gap within companies. Top-1% (p99) developers generated 46 times more lines of AI-written code per day than the median active user, and merged 15 times more pull requests per week than the median PR author. By Cursor's calculations, measured by the Gini coefficient, the distribution of spending, token consumption and AI-generated code inside companies is more unequal than the income distribution of any country in the world.

Costs vary widely too: the price of a single agent request differed by almost 9x across model families, and the cost per accepted line of code by roughly 7x. That is one reason companies benefit from access to multiple models and providers at once: different models are better suited to different types of tasks.

What does this mean?

Corporate AI spending is growing faster than the understanding of how to measure it. Most of the returns are concentrated among a small share of "power users" rather than spread evenly across the organization, and Cursor intends to turn measuring those returns into a distinct corporate discipline through a regular dialogue among CFOs.

FAQ

How much did companies spend on AI in 2025?

Global AI spending reached $1.5 trillion in 2025, shifting from experimental pilots to a permanent operating line item in the budget.

What is the Cursor CFO Council?

It is a working group of chief financial officers launched by Cursor. It will meet quarterly in cities around the world to develop a common approach to measuring the effectiveness of AI spending.

How unevenly are AI gains distributed among employees?

According to Cursor's Developer Habits Report, top-1% (p99) developers produce 46 times more AI-assisted code per day and merge 15 times more pull requests per week than the median user. That spread exceeds income inequality in any country in the world.