AI Spending Falls 10% Among Top Businesses
Ramp’s September 2026 AI Index shows that AI spending among the top 1% of corporate spenders fell nearly 10% month on month, from about $7,976 per employee in July to $7,205 in August. The data covers transactions from more than 70,000 US businesses.
Despite the decline in AI spending, paid AI adoption rose 0.4 percentage points to 56% of Ramp’s customer base. Median AI spending across all customers remained about $12 per employee per month, highlighting the wide gap between typical users and the largest enterprise adopters.
Ramp lead economist Ara Kharazian attributed the pullback partly to seasonal factors, including summer holidays, but also pointed to falling AI infrastructure costs. Average effective token prices dropped to $0.68 per million tokens in early September from $1.15 in March, a 41% decline. Lower prices may allow companies to obtain similar AI capacity with smaller budgets, meaning weaker AI spending does not necessarily indicate falling demand.
Anthropic led business adoption on Ramp’s platform in August with a 43.8% share, ahead of OpenAI at 39.8%. The figures suggest that AI tools are becoming standard enterprise software, while price competition and improving efficiency are reshaping corporate AI spending.
Neutral
The market impact is neutral because the report contains both negative and positive signals for technology and AI-related assets. The nearly 10% decline in AI spending among the largest corporate users could initially raise concerns about spending fatigue, slower enterprise demand and weaker revenue growth for AI infrastructure providers. Such concerns can weigh on risk appetite, particularly when traders are already monitoring high AI valuations and capital expenditure sustainability.
However, paid AI adoption increased to 56%, while token prices fell 41% over roughly six months. This suggests that lower spending may reflect improved efficiency and cheaper compute rather than a meaningful contraction in AI usage. Anthropic’s 43.8% business adoption share, ahead of OpenAI’s 39.8%, also points to continued competition and commercial adoption across the sector.
For cryptocurrency markets, the direct effect is limited. The article does not mention any cryptocurrency, blockchain network or token, and it provides no direct catalyst for buying or selling digital assets. Short-term trading reactions would therefore likely be indirect, mainly through changes in sentiment toward technology stocks, AI infrastructure and broader risk assets. If investors interpret the data as evidence of slowing AI demand, technology-linked markets could face pressure. If they focus on lower costs and rising adoption, the report could support a more constructive view of AI growth.
Over the longer term, falling AI costs and wider enterprise adoption could benefit companies providing computing, data-centre and software infrastructure. For crypto traders, the key indicators to monitor are AI-related capital expenditure, technology-sector valuations, macroeconomic liquidity and whether risk sentiment spreads from equities into digital assets. Similar past reports showing strong AI adoption but improving cost efficiency have generally produced mixed, rather than market-wide, reactions.