Evaluating AI Agents using the Price-to-Mindshare (P/M) ratio
tl;dr:
As with the stock market's Price-or-Earnings (P/E) ratio, a lower P/M ratio may indicate undervalued AI agents.
The Concept
This blog post introduces the Price-to-Mindshare (P/M) ratio. It’s the market cap over the mindshare for AI Agent tokens. The mindshare measures how often an AI Agent’s X account has been mentioned on Crypto Twitter. Different data providers apply their magic sauce on how they aggregate and weigh these mentions. As with the stock market's Price-or-Earnings (P/E) ratio, a lower P/M ratio may indicate undervalued AI agents.
The price-to-earnings (P/E) ratio compares a company’s share price with its earnings per share. I make two changes to arrive at the P/M ratio:
- I multiply the P/E ratio by the number of shares. That gives us the aggregate view, i.e., the Market cap over total earnings.
- I replace earnings with mindshare because an AI Agent’s social media traction is one of its key value drivers.
I keep the nominator as “P” because it’s still the valuation side.
P/M Ratio in Action
I pulled the three-day (3D) mindshare and the market cap from Cookie Fun on January 11, 2025. For the top-50 AI Agents (by 3D mindshare) with available data, I calculate the P/M ratio. It is the market cap (in $ million) over the mindshare (in %) for these 50 AI Agent tokens.
The average P/M ratio is 135 (the median is 87). It ranges from 5.8 to 1,400, showing the huge dispersion in the rapidly evolving market for AI Agents.
Most of the lower-ranked AI Agents are relatively small, with one notable exception: aixbt had a market cap of $480m and a P/M ratio of only 43. On the higher end are ai16z (178) and VIRTUAL (358) which enjoy a blue chip valuation because of their central position in their respective ecosystems.
Conclusion
The P/M ratio may become a useful metric to evaluate AI Agent tokens. One should bear in mind to only compare relatively similar agents (say within a certain ecosystem, agent category, and size bucket).
