Glonce

Ramp AI Index: US companies spend less on AI as usage climbs about 50%

businessThe Glonce editorsThe Decoder🔥71Glonce rating71Impact65Novelty52Relevance98Surprise60Players78Credibility82
Ramp AI Index: US companies spend less on AI as usage climbs about 50%

US companies are spending less on AI, according to the latest Ramp AI Index from Ramp economist Ara Kharazian. Spending peaked in July. Since then, usage has climbed about 50 percent and hit a record high at the end of September. In other words, businesses are getting more AI for less money.

Kharazian points to two causes: price cuts on top models, and cheaper, more efficient standard and lite models. He says the drop comes almost entirely from competition between OpenAI and Anthropic. The source notes that this is not the first decline in spending, but that it has lasted unusually long.

On the vendor split, Kharazian says open-source models still make up less than five percent of business spending. In the last week of September covered by the data, Anthropic took 51 percent of token spending and OpenAI took 44.5 percent.

The Ramp AI Index tracks transaction data from more than 70,000 US companies, and its token data comes from a subsample. According to Kharazian, the analysis only covers API spending and skews toward large AI customers.

Key facts

  • The latest Ramp AI Index, from Ramp economist Ara Kharazian, shows US companies spending less on AI; spending peaked in July.
  • Usage has climbed about 50 percent since the July spending peak and hit a record high at the end of September.
  • Kharazian says the drop comes almost entirely from competition between OpenAI and Anthropic, with price cuts on top models and cheaper standard and lite models.
  • In the last week of September covered by the data, Anthropic took 51 percent of token spending and OpenAI 44.5 percent; open-source models make up less than five percent of business spending.
  • The index tracks more than 70,000 US companies, but covers only API spending and skews toward large AI customers.

Why it matters

The index points to falling prices for AI: spending is down from its July peak while usage rose about 50 percent and reached a record high at the end of September. Kharazian ties this almost entirely to competition between OpenAI and Anthropic, which together took 51 percent and 44.5 percent of token spending in the last week of September covered by the data. The source adds that this is not the first spending decline, but it has lasted unusually long.

Who it affects

Mainly US companies that buy AI through APIs, since that is the spending the index measures. The two vendors named are Anthropic and OpenAI, which split nearly all of the token spending in the data. Open-source models, at under five percent of business spending, remain a small slice.

How to use it

Read it as a market gauge rather than a price list. The source names no specific price cuts, models or price levels, so it gives a buyer no figures to plan a budget around. What it does show is the direction: top-model price cuts and cheaper standard and lite models are, per Kharazian, behind lower spending alongside higher usage.

How solid is it

The Ramp AI Index is built on transaction data from more than 70,000 US companies, and the author is a Ramp economist. The vendor shares come from a subsample, and Kharazian himself notes the analysis skews toward large AI customers. The 51 percent and 44.5 percent figures cover one week, the last week of September covered by the data. The cause (competition between OpenAI and Anthropic) is Kharazian's reading.

Risks and caveats

The analysis covers only API spending, so it says nothing about consumer or subscription spending. It skews toward large customers and covers US companies only. The 50 percent figure refers to usage, not spending, and the source does not say how usage is measured. No absolute dollar amounts or size of the spending decline are given, and the vendor shares come with no earlier comparison, so the data does not show whether Anthropic or OpenAI is gaining or losing ground.