Back to top

Image: Bigstock

Rising Token Usage to Favor AI Investing? ETFs in Focus

Read MoreHide Full Article

Key Takeaways

  • Token usage is becoming a key indicator of AI adoption and demand.
  • Anthropic generates about 80% of revenue from usage-based pricing.
  • ANTW, OAIW and XLC offer exposure to the expanding AI ecosystem.

As AI companies move toward public listings, token usage may become a key growth metric —similar to subscriber counts for streaming platforms. AI providers commonly charge customers for the tokens their models process — the units of text, code, or other data read and generated.

But investors still have limited visibility into the revenue generated per token and how much customers actually use, as mentioned on a Yahoo Finance article. Still, investors should note that Anthropic generates most of its revenue through usage-based pricing, charging customers according to the AI model and the number of tokens processed. And Anthropic's reported annualized revenue run rate is quite strong on a top-line basis.

In the first half of 2026, consumption-based revenue accounted for about 80% of sales, with subscriptions contributing the rest. Anthropic reported $33.2 billion in revenue in the first nine months of 2026, up from $16.3 billion in the first half, and put its run-rate revenue at about $81 billion as of Oct. 31.

OpenAI and Anthropic’s combined run-rate revenue surged 18-fold from December 2024 to Q3 of 2026, while Google’s monthly token volume grew 330-fold, per data provided by Report AI.

Rising Token Usage

The volumes involved in this context are enormous. Google said it processed more than 3.2 quadrillion tokens per month across its products in May 2026 compared with 9.7 trillion per month in spring 2024, as mentioned on Report-AI.

At GPT-3.5 quality, $1 bought 50,000 tokens in November 2022 versus 14.3 million in October 2024 — a 286-fold increase. At GPT-4 quality, $1 bought about 5 million tokens by October 2026, per Report AI.

At OpenAI’s flagship list price, $1 bought about 250,000 tokens in October 2026, against about 229,000 for GPT-4o in 2024, according to Report AI's calculation.

Why Should Investors Care About Tokens Used?

Token growth can signal rising demand, not financial quality. A company may process more tokens while cutting prices, granting credits, or absorbing higher computing costs. But then, falling token prices in recent times could actually boost demand by making AI cheaper to use.

OpenAI CFO Sarah Friar recently said usage of its cheapest model increased tenfold following a price cut, highlighting the strong price sensitivity of customers, as mentioned on the same Yahoo Finance article.

Bottom Line

Overall, investors need to clearly know revenues per token, margins, customer retention, and infrastructure spending to judge whether usage growth is profitable. The industry still lacks a standardized disclosure for revenue earned per token.

For now, Anthropic’s disclosures offer one of the clearest windows into the economics: token-based consumption generated about 80% of first-half revenues, with the company reporting that the run rate reached roughly $81 billion.

Against this backdrop, one needs to keep a close tab on ETFs like Anthropic AI Lab Ecosystem ETF (ANTW - Free Report) and OpenAI Lab Ecosystem ETF (OAIW - Free Report) , launched by Harbor Capital in August 2026.

Because OpenAI remains private and Anthropic has been private so far, these ETFs invest in public companies linked to the OpenAI and Anthropic ecosystem rather than holding actual companies’ shares directly.

Alphabet-heavy ETFs, such as Communication Services Select Sector SPDR Fund (XLC - Free Report) , should also be considered due to Google's presence in the field through Gemini AI models.

Published in