We're building thought at scale.

What we believe

1. 01

   Intelligence should be abundant.

   For most of history, the amount of thinking the world could do was limited by people and time. AI gives us the ability to create intelligence at a scale that has never existed before.

   We want to help make that abundance real.

2. 02

   The cost of thought matters.

   When intelligence becomes cheaper, it gets used in more places, on smaller problems, and by more people.

   A model that is ten times cheaper isn't just cheaper. It makes entirely new products and use cases possible.

   That's why we obsess over the cost of inference.

3. 03

   AI will become infrastructure.

   AI won't remain a feature people occasionally use. It will sit underneath software, businesses, and eventually most industries.

   That infrastructure needs to be fast, dependable, and economical at enormous scale.

   That's what we're building for.

4. 04

   Models are only part of the system.

   Useful intelligence takes more than running a model.

   Routing, caching, tools, context, reliability, and infrastructure all determine how well AI works in the real world.

   We build the systems around models that make them more useful and less expensive.

5. 05

   More intelligence is a good thing.

   We believe making useful intelligence dramatically more abundant will help people build better products, solve harder problems, and spend less time on work machines can do better.

   And we think we're still at the very beginning.