Google processed nearly a quadrillion tokens across its AI services.

Tokens are units of data processed by AI models during training and inference, giving some context into the scale of the usage of Google services. Tokens can be made up of text, images, audio clips, videos, or another modality.

In an OpenAI blog post, the rival generative AI company said that one token was usually equivalent to four characters in English, while one paragraph usually meant around 100 tokens. The transcript of the US Declaration of Independence contains 1,695 tokens.

“We processed almost 1,000,000,000,000,000 tokens last month, more than double the amount from May,” Google DeepMind CEO Demis Hassabis said on X.

The increase in tokens processed may not necessarily mean a doubling in usage, but more that reasoning models process significantly more tokens per use. Google began rolling out a family of Gemini 2.5 reasoning models from the end of March.

Increasing the number of tokens processed requires an increase in data center infrastructure. This month, Google said that it expected data center capex for the year to increase by $10 billion to $85bn.

“With compute, there’s the amount of compute you have for training, often it needs to be colocated – so actually even bandwidth constraints between data centers can affect that,” Hassabis said on the Lex Fridman podcast.

“But because now AI systems are in products being used by billions of people around the world, you need a ton of inference compute now. And then on top of that, there’s the thinking systems – the new paradigm of the last year – where they get smarter the longer amount of inference time you give them at test time.

“So all of those things need a lot of compute, and I don’t really see that slowing down. And as AI systems become better, they’ll become more useful and there’ll be more demand for them.”