NewsMacroRamp Expands AI Index to Track Token Spend and Model-Level Usage Across 70,000 Businesses

Ramp Expands AI Index to Track Token Spend and Model-Level Usage Across 70,000 Businesses

Author: CryptoBriefing·

Key Takeaways

  • •Ramp's AI Index built from the spending activity of more than 70,000 U.S. businesses, recorded paid AI adoption of 56.1% as of August 2026, compared with the Census Bureau's 22.1% survey-based estimate.
  • •The roughly 34-percentage-point gap between the two figures reflects differing methods: surveys rely on businesses self-reporting AI use between collection waves, while transaction data captures payments as they clear.
  • •AI spending is highly concentrated, with the median adopting firm spending about $12 per month while the top 1% of firms spent between $7,200 and $7,450 monthly.
  • •Anthropic held a 43.8% share of business AI adoption among Ramp's client base as of August 2026, ahead of OpenAI at 39.8%, though the shares reflect only Ramp's sample rather than the overall market.
  • •Ramp launched a Token Spend Management product that connects to business accounts and delivers weekly breakdowns of AI spending by provider, model, and individual user.
Ramp Expands AI Index to Track Token Spend and Model-Level Usage Across 70,000 Businesses

Corporate spend management platform Ramp has expanded its AI Index to include token volumes, spend breakdowns by model and provider, and weekly updates, giving businesses a far more granular view of what AI is actually costing them.

Ramp, a fintech best known for its corporate cards and expense management software, compiles the index from the spending activity of businesses on its platform. Because AI model providers typically price API access by the token — a small unit of text processed by a model — token spend offers a direct measure of how heavily companies are actually using AI tools.

The numbers the Census Bureau missed

Ramp's index, drawn from more than 70,000 U.S. businesses, puts paid AI adoption at 56.1% as of August 2026. The Census Bureau's equivalent estimate, derived from its periodic business surveys, sits at 22.1%.

That is not a rounding. The 34-percentage-point discrepancy suggests government surveys are capturing a fundamentally different picture of the market than what corporate card and spend data actually shows. The divergence also turns on method: survey estimates depend on businesses self-reporting AI use between collection waves, while transaction data records payments as they clear — a difference large enough that the source a company cites can shift the answer by tens of percentage points.

Spending distribution

The spending distribution inside that 56.1% adoption figure is striking. The median firm on Ramp spent roughly $12 per month on AI, while the top 1% of firms spent between $7,200 and $7,450 per month. The spread matters as much as the headline: adoption is broad, but by the index's own numbers the typical adopting firm's outlay remains modest, with spending concentrated in a small top tier.

The index also noted a trend of disciplined spending among top users, with certain firms showing declining per-employee AI expenditure month over month.

Anthropic leads OpenAI in business adoption share

As of August 2026, Anthropic held a 43.8% share of business AI adoption among Ramp's client base, ahead of OpenAI at 39.8%. These shares describe Ramp's 70,000-plus sample specifically rather than the market at large, so they are best read as a window into one large cross-section of U.S. businesses.

Token Spend Management as a product category

The index expansion is tied to Ramp's launch of a dedicated Token Spend Management product, which connects to business accounts and provides breakdowns by provider, model, and individual user on a weekly basis. That gives finance teams per-model and per-user line items for AI spending that can be reviewed alongside other operating expenses.

Because the index is updated weekly, provider and model rankings will shift regularly, making it one of the most frequently refreshed reads on where business AI budgets are actually flowing.