Business AI spending rises 406% at Revolut Business

Revolut Business has published data showing monthly business spending on artificial intelligence tools rose 406% over the past year. The figures cover spending patterns across more than 850,000 business accounts.

The data shows companies are spending far more heavily on AI than retail users. Average monthly spend per business user reached USD $416, compared with USD $42 for the average retail user. Consumer AI spending rose 259% over the same period.

The figures suggest the market is moving beyond initial trials. Businesses are increasingly concentrating their spending with a smaller group of providers rather than abandoning AI tools after short experiments.

Retention rates among business users were highest for OpenAI at 89.9%, followed by Anthropic at 89.7% and Cursor at 85.4%. At the same time, 31.9% of business users on smaller platforms switched to rival providers, while 10.1% stopped using tracked AI tools altogether.

This points to two parallel trends: consolidation around a handful of established names, and a continued willingness among companies to change suppliers when smaller services fail to hold attention. It also suggests AI spending is being treated more like a recurring operating cost than a one-off test budget.

Micro firms

Smaller businesses account for most AI adoption by volume. Micro-businesses made up 77.9% of AI-using firms in the data and represented 59.7% of total spend.

However, spending rises steeply with company size. Firms with one to 10 employees spent an average of USD $236 on AI in July 2026, while mid-market businesses with 51 to 200 employees spent USD $2,031 on average.

The gap shows that larger organisations are committing bigger sums as AI becomes embedded across more functions. It also underlines how quickly very small firms have adopted the technology. For micro-businesses, software that can automate drafting, coding, analysis or customer support may offer a way to expand output without hiring at the same pace.

Sector trends were equally pronounced. Consulting, IT and business services accounted for 66.4% of all business AI spending in the dataset, despite making up 25.8% of active businesses on the platform.

These industries tend to rely heavily on knowledge work, documentation and digital workflows, making them early candidates for spending on language models, coding assistants and related tools. The concentration also suggests AI adoption remains uneven across the broader economy, with service-led sectors moving faster than businesses whose operations are less centred on digital tasks.

Changing budgets

Revolut Business tracks anonymised, aggregated transaction data from its global business customer base, covering commercial AI adoption, spending, spend per user, vendor retention, and trends by sector, geography and company size. The index includes both core AI applications and related infrastructure such as GPU cloud services and model-serving platforms.

The methodology defines an AI user as a business that makes at least one completed payment to a tracked AI vendor in a calendar month. Spending totals include subscriptions, token purchases and prepaid credits, meaning the figures reflect overall outlay on AI suppliers rather than a single product category.

This matters when comparing business and consumer patterns. A retail user may subscribe to one chatbot or image tool, while a company may pay for a mix of staff licences, application programming interface usage and specialist software layered on top of foundation models.

James Gibson, Head of Revolut Business, commented on the pace of growth in the category.

"We've never seen a category of spend scale as quickly as AI spend, and we see no sign that it's slowing. Through this transition, we are continuing to build and adapt Revolut Business for our hundreds of thousands of ambitious, technologically advanced businesses," Gibson said.

He also pointed to possible shifts in how businesses buy AI as the market develops.

"We expect AI spend to continue scaling in the coming years, but as more AI-first products come to market, some of that spending could shift away from direct token purchases and back towards third-party software," Gibson said.