The Signal Behind the Spending

Ramp's AI Index is not a survey. That distinction matters. Rather than asking executives what they plan to spend or how bullish they feel about artificial intelligence, the index pulls from actual transaction records — corporate card charges and bill-pay flows across more than 70,000 U.S. firms that use Ramp's platform. Published monthly, it measures the proportion of those companies actively subscribing to paid AI models, platforms, and tooling. The result is one of the few ground-level datasets that reflects what businesses are actually doing with AI, rather than what they say they intend to do.

For practitioners inside technology organizations, that methodological distinction is worth dwelling on. Sentiment surveys have a long history of overstating adoption during hype cycles and understating it during quiet, infrastructure-building phases. Spend data, by contrast, tends to lag intention slightly but carries a cleaner signal: money left an account, a contract was signed, a seat was provisioned. It is harder to fake than a checkbox on a questionnaire.

The Top 1%: A Category of Their Own

The most striking figure in the index is the spending profile of the cohort Ramp labels the "AI-pilled" firms — the top one percent of companies on the platform by AI expenditure. These businesses are spending $7,500 per employee per month on AI tools and services, according to Ramp's AI Index. That is not an annual budget figure normalized to a monthly rate. It is a recurring monthly per-capita outlay.

To put that in industrial context: $7,500 per employee per month works out to $90,000 per employee per year dedicated solely to AI subscriptions and services. For a 50-person team, that is $4.5 million annually flowing toward AI tooling — roughly the salary cost of several senior engineers in a high-cost market, spent not on headcount but on capability procurement.

Who are these firms? The index does not publish a named roster, and the "AI-pilled" label, while catchy, is Ramp's own framing. What the data suggests, structurally, is a cohort running AI as core infrastructure rather than as a discretionary productivity add-on. At $7,500 per seat per month, you are not talking about a ChatGPT Teams subscription and a couple of Copilot licenses. You are looking at organizations likely running proprietary or fine-tuned model deployments, high-volume inference workloads via API, specialized vertical SaaS built on foundation models, or some combination of all three.

Worth flagging: the per-employee figure, while dramatic, does not distinguish between companies where every employee is an AI power user and companies with a small specialist team driving the bulk of spend. A 10-person AI-native startup with a $75,000 monthly model API bill will look identical to a 10-person team inside a larger organization that has been given a dedicated AI budget. The index, as a spend-per-employee ratio, compresses that heterogeneity.

Subscription Penetration as the Baseline Metric

Beyond the outlier cohort, the index's month-over-month subscription penetration figure is the more broadly actionable number for anyone tracking enterprise AI rollout. By measuring the fraction of Ramp's 70,000-plus firms that carry at least one paid AI subscription in a given month, the index provides a running baseline for how far commercial AI adoption has diffused across the U.S. business population — not just among tech-forward early adopters but across the full distribution of company types and sizes that use Ramp's corporate card infrastructure.

This is meaningful because Ramp's customer base skews toward small and mid-market companies — the segment that is often invisible in enterprise AI coverage, which gravitates toward Fortune 500 deployments and hyperscaler announcements. Penetration data from this cohort gives a more honest read on whether AI tooling is crossing the chasm into mainstream business use, or whether it remains concentrated among the digitally native.

We have seen this pattern before. In the mid-2000s, cloud infrastructure adoption followed a nearly identical diffusion curve: a small cluster of startups ran everything on AWS from day one, a broader layer of tech-forward SMBs followed within two to three years, and the long tail of conventional businesses took the better part of a decade to move meaningful workloads off on-premise hardware. The Ramp index, applied consistently month over month, is essentially an early-warning instrument for where AI sits on that same S-curve — and how quickly the slope is steepening.

What the Data Does and Does Not Tell Us

Transaction-level spend data has its own blind spots. It captures what flows through Ramp's platform, which means companies using other corporate card providers, direct bank transfers for large vendor contracts, or internal infrastructure budgets — GPU clusters, self-hosted models, internal tooling built on open-weight models like Llama or Mistral — are not reflected. Firms running significant AI workloads on cloud compute billed directly through AWS, Azure, or GCP reserved-instance agreements may also be undercounted if those charges do not route through Ramp's bill-pay system.

The index is also U.S.-centric by design, which limits its utility as a proxy for global AI adoption but sharpens its value as a domestic business barometer. For policy analysts, investors, or enterprise vendors trying to calibrate U.S. market sizing, a 70,000-firm transaction panel is a meaningful sample — large enough to be statistically robust, granular enough to surface segment-level variation.

The Infrastructure Implication

The $7,500-per-employee-per-month figure for the top one percent is not merely a curiosity about outlier behavior. It is an indicator of what AI-as-infrastructure looks like in practice for organizations that have made the commitment fully. At that spend level, AI is no longer a line item under "software and SaaS" — it is a capital-equivalent investment in operational capability, sitting alongside compute, networking, and security in the core infrastructure budget.

For the majority of technology organizations that have not yet reached that threshold, the more relevant question is what the subscription penetration curve looks like across the full 70,000-firm panel, and how fast it is moving. That monthly cadence — the proportion of companies crossing from zero paid AI subscriptions to at least one — is where the adoption story is actually being written, one corporate card charge at a time.

The Ramp AI Index does not tell you which AI tools are winning, which vendors are losing, or whether the productivity gains justify the spend. Those are harder questions that transaction data alone cannot answer. What it does provide is a relatively clean, high-frequency read on the breadth and intensity of commercial AI adoption in the U.S. business population — grounded not in aspiration but in the rather less glamorous evidence of what companies are actually paying for.