Enterprise AI spending: 91% plan to spend more, coding takes 72%
Updated
Illustrative photo
What the new enterprise AI studies found
Three separate studies of how businesses use AI were published within days of each other. They tell a similar story about enterprise AI: companies now have more options and are moving around to find better deals. The sources are the new Databricks AI Index, Arcana Research's first annual State of Enterprise AI study, and the Ramp AI Index for October and November.
The common thread is a shift in leverage. According to the podcast's summary of the studies, buyers hold the leverage in what is increasingly a buyer's market. The leading companies are using increased token price competition to secure good deals, move toward efficiencies, and raise their AI usage without increasing costs proportionally.
Arcana's survey found that 91% of respondents expected to spend more in 2026, with most expecting their spend to rise between 20% and 40%.
The gap sits in oversight. A full two-thirds of the leaders Arcana surveyed said they did not have visibility into what they spend on model and API usage. The podcast notes that this is likely to change quickly. It also describes a gap between vanguard organizations, such as those in the Ramp Index that operate more nimbly with AI, and many middle-of-the-range Global 2000 companies that appear to still operate with older methods.
Switching vendors is becoming normal
Loyalty is loosening. In Arcana's survey, a full third of OpenAI customers said they would be open to switching, frequently considering a move to Anthropic. Among GitHub Copilot customers, 43% were open to switching, generally naming Claude Code.
The podcast frames this as part of why buyers have leverage: switching is getting increasingly normal. The studies describe a market where businesses have more options and are more comfortable moving around to find the best deals.
A few firms spend hundreds of times more
Spending is extremely concentrated. Ramp found that as of November, the top 1% of firmsspend about 625 times what the median firm spends per employee per month. The median sits at $16.32; the top firms spend over $10,000 per employee per month.
Databricks shows a similar spread: its median user spends $2.49, while the 99th percentile spends almost $700 a week, a 281x difference. Ramp's sample is a specific group of generally technology-based and tech-forward companies, so its top 1% is a very differentiated group.
Coding still eats most of the budget
Across all the studies, coding remains dominant. On Ramp, 72% of AIspend comes from engineering teams.
Affiliate links: we may earn a commission from qualifying purchases, and it never changes our data or rankings. They open Amazon's search results: prices and availability are shown there.