The move signals a strategic expansion for a network that launched its mainnet less than a year ago. Boundless went live on Base in September 2025 after an incentivized testnet earlier that summer. At the time, its core value proposition was straightforward: replace the small group of centralized proof-generation providers that currently handle more than 90% of ZK Layer 2 networks and, according to industry estimates, capture 60 to 70% of transaction fees in the process.

Analysts have put the dollar value of that bottleneck at roughly $97 million in the ZK proof market.

How the Network Works

Boundless runs on RISC Zero's zkVM, a zero-knowledge virtual machine built on the RISC-V instruction set. The system compiles programs, converts the output into a format called STARK proofs, aggregates them recursively, and then converts the result into Groth16 proofs suitable for verification on a blockchain. Settlement currently happens on Ethereum, Base, and, since a January 2026 integration, Bitcoin.

The Bitcoin layer uses BitVM, a framework for verifying off-chain computation on Bitcoin without changing the protocol, which launched in mid-2025, and connects through Citrea, described as Bitcoin's first production ZK rollup.

The design treats Bitcoin as a final verifier and trust anchor for ZK proofs generated elsewhere.

"This isn't about making Bitcoin like Ethereum," CEO Shiv Shankar said at the time of the Bitcoin integration. "It's about securing Ethereum-style execution with Bitcoin's security."

Live integrations include Wormhole, which uses Boundless to handle ZK verification for Ethereum consensus; BOB, a Bitcoin rollup focused on EVM-to-Bitcoin interoperability; and Lido, which uses the network to generate proofs around validator exit transparency. Earlier integrations cover EigenLayer, Celestia, and Taiko.

Token Mechanics and Market Position

Boundless operates under a model it calls Proof of Verifiable Work, which pays GPU operators for producing cryptographically verifiable computations, specifically zkVM cycles, rather than the useless hashes generated by traditional proof-of-work mining.

Provers receive 75% of ZKC token emissions; stakers receive the remaining 25%.

The ZKC token is trading at approximately $0.076 as of early July 2026, well below its ICO price of $0.29 and a significant distance from its all-time high of $1.87.

Market capitalization estimates vary by methodology. The project's self-reported figure stands at approximately $13.4 million, while third-party estimates from CoinGecko, Coinbase, and CoinMarketCap range from $15.9 million to $23.7 million.

Circulating supply sits at around 290 million tokens against a total supply of 1 billion.

The token faced a specific headwind in October 2025 when South Korean exchange Upbit issued an investment warning on ZKC, triggering a 46% price decline before the restriction was lifted.

The network had 363 active provers on its explorer at an earlier snapshot.

On pricing competitiveness, RISC Zero has claimed its system costs at least seven times less than SP1, the zkVM developed by competing firm Succinct Labs, and as much as 60 times cheaper for smaller workloads. Fenbushi Capital has independently confirmed RISC Zero's GPU benchmark performance, though Succinct holds a lead on developer adoption.

Why This Matters Outside the United States

The AI compute expansion is particularly relevant in markets where centralized alternatives are expensive or access-controlled.

In India, the government's IndiaAI Mission has deployed 34,000 GPUs at roughly 42% below commercial rates, but access requires an application reviewed by the Ministry of Electronics and Information Technology, prioritizing academic institutions and MSMEs (Micro, Small and Medium Enterprises) based on national importance criteria. The mission is also targeting 100,000 GPUs by year-end, a scale-up that may eventually narrow the access gap for developers who currently fall outside the program.

Developers building on Ethereum or Base who fall outside those criteria still face hyperscaler pricing of $2.50 to $4.00 per H100-hour from providers like AWS, Azure, and Google Cloud.

A permissionless decentralized compute market offering cost reductions of 50 to 70% versus hyperscaler rates, according to figures sourced from Spheron Network, is a practical alternative for that population.

In Africa, infrastructure constraints add a different dimension. Nigeria, home to one of the continent's largest developer communities, operates a national grid that has never exceeded 6 gigawatts, adding roughly 40% to operating costs at conventional data centers.

South Africa accounts for approximately 70% of the continent's existing data center capacity.

Large investments are in progress, including a planned $240 to $250 million GPU facility by MTN Group at its Ikeja data center in Lagos and a $1 billion package from Microsoft and G42, an Abu Dhabi-based artificial intelligence and technology company, that includes an Azure data center in Kenya. The MTN facility is targeted for delivery in the second half of 2026; a timeline for the Microsoft and G42 project has not been publicly disclosed.

For developers in Nairobi or Lagos building on Base today, a decentralized network where prover nodes can be operated globally addresses a gap that centralized alternatives have not yet closed.

What Comes Next

The decentralized compute sector has grown from a $5.2 billion market cap to roughly $19 billion over the past twelve months, with AI-focused DePIN projects now representing 48% of that total.

Akash Network reported a record $5 million in compute spend in Q1 2026; Render Network carries a market cap above $2 billion.

Boundless enters this segment with an existing GPU base and a proof-generation use case that is central to its original network design, one that larger general-purpose compute networks were not built around.

Whether the AI compute demand materializes at sufficient volume to sustain network growth will likely be the clearest test of the expansion's commercial logic, with utilization rates, active prover count, and revenue metrics providing more durable signals than token price alone.