Ramp AI spending data shows that businesses are using more artificial intelligence without significantly increasing their spending on the technology. Corporate AI token spending rose just 1.5% across August and September 2026, compared with a 42% monthly increase in July, according to the financial technology company’s latest research.
The slowdown comes as Anthropic prepares for an initial public offering and OpenAI faces growing pressure to justify its valuation. Rising usage alongside relatively flat spending suggests that cheaper models, competitive pricing and stronger cost controls are changing the economics of enterprise AI.
Ramp AI Spending Data Reveals a Shift in Business Demand
Ramp’s AI Index, which analyzes transactions from more than 85,000 businesses, shows that spending growth has slowed despite increasing token consumption. Its October research report found that the highest-spending customers were beginning to reduce their expenses.
The top 1% of companies buying AI services spent approximately $9,568 per employee in September, down 3.1% from their July peak. Ramp economist Ara Kharazian told TechCrunch that spending among these customers matters because they account for a disproportionate share of AI model revenue.
Other businesses continued increasing their spending, but their purchases were insufficient to offset the slowdown among the largest customers. Ramp’s data also showed some recovery in October, with token spending rising 8.9% between Sept. 23 and Oct. 23.
OpenAI and Anthropic Face Growing Pricing Pressure
Competition between AI developers is helping businesses process more information at lower prices. OpenAI reduced API pricing for GPT-6 Sol and GPT-6 Luna by 50% compared with their GPT-5.6 promotional pricing, as detailed in its model announcement.
Anthropic has also reduced pricing for several Claude models. According to its financial disclosures, price cuts reduced revenue growth by approximately $600 million. Both companies have offered enterprise discounts as customers negotiate lower costs for larger workloads.
Businesses are also tightening spending controls by limiting access to expensive models, setting employee budgets, and routing simpler tasks to cheaper alternatives. These measures allow companies to increase AI usage without proportionately increasing their bills.
Anthropic IPO Valuation Faces Investor Scrutiny
The spending figures come as AI valuations face a difficult period. Bloomberg reported on Nov. 6 that prospective Anthropic investors were discussing a valuation between $1.2 trillion and $1.3 trillion, below earlier expectations of as much as $2 trillion.
OpenAI is facing similar pressure. The Financial Times reported that the company reduced its fundraising target by approximately $10 billion while reconsidering its valuation expectations.
TechAmerica previously examined Anthropic’s rising computing costs, which consumed 58% of its third-quarter revenue. Those expenses add pressure as AI developers invest heavily in infrastructure while customers seek lower prices.
AI Adoption Continues Despite Slower Spending
The slowdown does not necessarily indicate declining demand. Ramp found that 55.9% of businesses on its platform paid for AI services in September, while Anthropic and OpenAI recorded adoption rates of 43.8% and 40.6%, respectively. Some businesses subscribed to both providers.
Kharazian said the challenge for AI companies is that additional usage no longer guarantees comparable revenue growth. Coding agents have generated substantial commercial demand, but other applications may not support the same spending levels.
For Anthropic and OpenAI, the central question is whether new products can generate enough additional revenue to offset lower model prices. Ramp’s findings suggest that enterprise AI adoption remains widespread, even as the pace of spending growth becomes harder to sustain.
