Sam Altman, the CEO of OpenAI, claims that AI token usage has grown by a factor of one million over the past six years. That alone would be a staggering number. But here’s the thing: he expects it to happen again.
The math behind the madness
Altman’s projection isn’t pulled entirely from thin air. Industry expert David Sacks has broken down where these exponential gains might come from: roughly 100x improvements in AI model algorithms, another 100x from advances in hardware like chips, and yet another 100x from raw compute power. Multiply those together and you get a million-fold increase in capability from those three vectors alone.
The real-world data supports the trajectory, even if it hasn’t yet reached Altman’s full vision. Google reported a 14-fold increase in AI token usage over just eight months during the 2024-2025 period.
OpenAI’s own ecosystem tells a similar story. Engineers and projects tied to OpenAI have reportedly processed between 100 billion and 600 billion tokens in a single month. One notable project racked up $1.3 million in API usage costs alone.
What AI tokens actually are (and why crypto cares)
In the AI context, a token is a chunk of text, roughly three-quarters of a word, that language models process. Every time you ask ChatGPT a question, it’s consuming tokens. Every API call from a developer burns tokens. The volume of these tokens is essentially a measure of how much the world is using AI.
Altman sits at the intersection of both worlds. He’s the CEO of OpenAI and also the co-founder of World, formerly known as Worldcoin, which uses biometric verification to distinguish humans from AI agents.
The World Network’s WLD token surged over 27% in January 2026 following news that OpenAI was exploring biometric tools that could integrate with World’s verification systems. The token had already posted gains exceeding 80% back in September 2025 when the project announced its treasury launch.
The ripple effects across crypto
WLD isn’t the only crypto asset riding the AI wave. Tokens like GRASS posted gains after Nvidia’s commentary about an “agentic future” for AI, where autonomous AI agents handle tasks independently and consume even more compute resources in the process.
For investors watching the AI-crypto crossover, the key metric to track is actual token processing volume, not promises about future growth. Google’s 14x increase over eight months is verifiable. Monthly processing volumes in the hundreds of billions are measurable. These are the data points that separate genuine demand from narrative-driven speculation.