The AI expansion
Boundless began by building infrastructure for zero-knowledge proving. In doing that work, we learned how to coordinate GPU supply, route demanding compute jobs, and make distributed machines behave like reliable infrastructure.
That foundation now gives Boundless a larger opportunity: AI inference. We believe inference is the most important compute market in the world, and the same network capabilities we built for proving can be extended to serve it.
As Boundless expands into inference, ZKC is expected to remain central to network participation, operator staking, access, and long-term alignment between compute suppliers, users, and the protocol.
As Boundless expands into inference, ZKC is expected to remain central to network participation, operator staking, access, and long-term alignment between compute suppliers, users, and the protocol.
Boundless plans for operators serving inference workloads to stake ZKC. Staking is expected to create accountability for reliability, uptime, and quality, while tying network participation to long-term commitment.
Staking requirements are expected to vary by hardware class, reliability tier, and expected earning potential. Higher-capacity nodes or nodes serving higher-value workloads may require more ZKC at stake.
Boundless plans to let ZKC holders and stakers access inference through the network, including the ability to pay using ZKC. More details on access, credits, discounts, or staking benefits will be shared as the product rollout progresses.
Boundless intends to introduce a ZKC buyback program connected to revenue from the inference network. The goal is to link commercial growth in AI inference back to the token economy. Details on timing, structure, and allocation will be shared once finalized.
Our expansion into AI broadens Boundless's focus while the ZK marketplace continues as it operates today. Prover software will remain open source, allowing provers to keep participating and serving marketplace demand.