Editorial illustration for Business AI Usage Climbs 50% Since July Peak, Ramp Index Shows
Business AI Usage Surges 50% Since July, Ramp Data Shows
Ramp's latest AI Index, compiled by company economist Ara Kharazian, tracks spending and usage data from more than 70,000 US businesses that run transactions through Ramp's corporate card and expense platform. The October update shows a pattern that cuts against the usual story about AI costs. Usage is up roughly 50 percent since spending peaked in July, and it hit a record high at the end of September. At the same time, what companies pay per unit of AI consumption has been falling.
Kharazian attributes the drop mainly to price cuts from the two dominant API providers, OpenAI and Anthropic, along with a shift toward cheaper standard and lite model tiers that do the job for less. Open-source alternatives haven't been the driver. They still account for less than five percent of business AI spending in Ramp's data, which only captures API usage and leans toward larger customers.
The index has logged pricing dips before, but Kharazian notes this one has stretched on longer than past corrections. The next section breaks down what's actually splitting the market between the two leading providers, and how that split has moved week to week through late September.
Why this matters
For anyone building on top of these models, the Ramp numbers confirm what procurement teams have suspected for months: the price of frontier AI is falling faster than usage is rising, and that gap is the whole story. A 50 percent jump in usage since July paired with declining spend means businesses are getting more tokens per dollar, which is good news if you're a founder running inference-heavy workloads on tight margins. But the concentration stat is the one worth watching closely.
Anthropic grabbing 51 percent of token spending in a single week shows this isn't a diversified market yet, it's a two-horse race between Anthropic and OpenAI, and open-source models stuck under five percent of spend suggests the "escape the API tax" narrative hasn't materialized at scale. If you're betting your infrastructure on open-weight models becoming cost-competitive anytime soon, this data says otherwise. The real question for Q4: does Anthropic's lead hold, or does the next price cut from OpenAI reshuffle that 51 percent number entirely.
Related Reading
Research & Benchmarks Microsoft AI Chief: "If It Isn't Safe We Shouldn't Build It
Research & Benchmarks Google's FACTS benchmark shows 70% factuality ceiling across four tests
Research & Benchmarks Nvidia's DLSS 4.5 beta adds 6x Multi Frame Generation for RTX 50 GPUs
From Archives AI solves accounting tasks flawlessly in benchmark with 160 tasks
From Archives Microsoft's New AI Transcribes 60 Languages in 100 Milliseconds
Common Questions Answered
What does the Ramp AI Index measure and what was the key finding in the October update?
The Ramp AI Index tracks spending and usage data from over 70,000 US businesses using Ramp's corporate card and expense platform. The October update revealed that AI usage has climbed approximately 50 percent since peaking in July, while simultaneously the cost per unit of AI consumption has been falling, indicating businesses are getting more value from their AI spending.
Why are US companies paying less per unit for AI consumption according to Ramp economist Ara Kharazian?
The decline in per-unit AI costs stems almost entirely from competition between OpenAI and Anthropic, which has driven price cuts on top models and led to the development of cheaper, more efficient standard and lite model alternatives. This competitive pressure has made frontier AI pricing fall faster than usage is rising.
What is the significance of the gap between AI usage growth and spending decline for founders building on AI models?
The 50 percent jump in usage since July paired with declining overall spend means businesses are achieving more tokens per dollar, which is particularly beneficial for founders running inference-heavy workloads on tight margins. This efficiency gain allows startups and companies to scale their AI operations more cost-effectively than previously possible.
When did AI spending peak according to the Ramp AI Index data?
According to the Ramp AI Index, AI spending peaked in July before declining in subsequent months. However, usage hit a record high at the end of September, demonstrating the divergence between spending trends and actual AI consumption patterns among US businesses.
Further Reading
- Businesses are using more AI and paying less for it, Ramp AI Index shows - The Decoder
- August 2026 Ramp AI Index: Cracks in the AI thesis - Ramp
- How much does it cost to be AI-pilled? - Ramp
- Cracks in the AI thesis, part 2 - Ramp
- Top AI spenders cut per-employee costs by nearly 10 percent in August - The Decoder
