Boundless pivots 4,000-GPU crypto network to AI inference
Zero-knowledge startup Boundless is redirecting its 4,000-GPU network toward AI inference, leveraging a token-staking model to undercut major cloud providers by up to 50%.
Boundless, a startup originally built to settle zero-knowledge proofs on Bitcoin, is expanding its distributed network of 4,000 GPUs to handle artificial intelligence inference workloads. The company has optimized its infrastructure—previously used to connect the Ethereum mainnet and Base Layer 2 to Bitcoin—through hardware tuning, workload adaptation, and managed routing.
The pivot underscores a broader market shift as crypto firms look to monetize their compute infrastructure by targeting surging demand for AI. While Bitcoin miners have aggressively redirected their industrial data centers toward high-performance computing, it has been less common for protocol or Layer 2 developers to follow suit. Boundless joins a small but growing cohort of crypto-native infrastructure projects, such as Eigen Labs, shifting capacity toward AI and the agentic economy.
For enterprise buyers, the financial appeal lies in cost arbitrage. Early benchmarks from Boundless indicate inference costs up to 50% lower than those of major cloud providers, particularly for asynchronous workloads. The startup achieves this pricing advantage by utilizing a mix of lower-cost capacity, including consumer-grade GPUs and hardware originally purchased for crypto mining and proving operations.
To govern this expanded network, Boundless is integrating its native cryptocurrency, ZKC. AI operators will be required to stake ZKC tokens to join the network, with the size of their stake directly tied to their earning potential. This structure transforms the token into a gatekeeping mechanism for compute access, tying its utility directly to AI sector growth.
The company argues the underlying technology for zero-knowledge cryptography translates seamlessly to AI processing. “ZK proving and AI inference are both resource-intensive and require the same operational layer, one that matches jobs to available capacity, keeps utilization high, and routes work efficiently,” the announcement read.
Boundless will continue operating its zero-knowledge proving network alongside the new AI operations. “Four years ago, we set out to solve one hard compute problem,” CEO Shiv Shankar said. “In the process, we built something bigger: a network for coordinating distributed GPU capacity. AI now needs the same foundation at a much larger scale.”
