Principles
Our work begins with open infrastructure and continues through products that make it useful. These principles guide what we build, how we build it, and what we believe it should make possible.
The freedom to compute belongs to everyone
Talent exists everywhere. The resources to act on it should be within everyone’s reach. We believe compute should ultimately be free. Our ambition is to make it so abundant that cost stops determining who can participate. Every step toward broader access gives more people the power to turn an idea into something real.
Open foundations should outlast their builders
We build infrastructure that others can inspect, contribute to, and continue. Creating Akash was a step toward a compute network that could grow through the work of many independent participants. Overclock remains a builder within that ecosystem. The foundation should give others room to build things we have never imagined.
Openness must become usable
An open model needs somewhere to run. An open network needs a way in. Access becomes meaningful when people can accomplish something with it. Akash Console makes deployment easier; AkashML makes inference available as a service. Our responsibility is to make that infrastructure practical to use.
Convenience should preserve choice
Removing complexity should leave people with meaningful control over their models, workloads, and providers. We design for the ability to understand, change, and move. As agents act on people’s behalf, that control should include clear authority over what they can do and spend. A simpler experience should strengthen the freedom of the person using it.
Build for change
Models will change. Chips will change. The infrastructure beneath them should accommodate that change. We favor clear interfaces and components that work together, so a useful improvement can be adopted without rebuilding everything around it. Lasting infrastructure gives new technology a place to become useful.
Abundance requires care
Compute depends on physical resources. Hardware, electricity, and reliable operations have real costs. We work toward abundance by getting more useful work from those resources: improving utilization, matching workloads to appropriate hardware, and extending equipment’s useful life where it serves the task. Our ambition for universal access requires careful engineering.
Earn by making the system more useful
Open infrastructure can support sustainable businesses. AkashML is a commercial inference service built on Akash. Its value comes from operating a service people need and making the underlying compute easier to use. We believe revenue should reward useful work. Sustainable services and providers are part of making open compute last.
Be clear about what is real
People should understand what they are using and what they are paying for. Capabilities, costs, and limitations deserve plain language. We distinguish open infrastructure from commercial services, and available products from work still ahead. Trust grows when the experience matches the promise, including in the details.
Let evidence improve the work
Our principles should help us make better decisions. We judge progress through broader access, performance, reliability, affordability, and user control. Openness and decentralization should improve those outcomes in practice. When experience reveals a better way, we revise our approach and keep building toward the freedom to compute.