Why the Prover Business Model Is Shifting from Software to Services

Open-source proving stacks like Risc Zero, SP1, and Jolt are making the core software a commodity. The value is shifting to the operational layer where performance, reliability, and cost matter.

• Key Benefit: Teams avoid the ~2-year R&D cycle of building a custom prover.
• Key Benefit: Competition shifts to prover uptime (>99.9%) and cost per proof (<$0.01).

Applications like on-chain gaming, high-frequency DeFi, and intents (e.g., UniswapX, Across) require sub-second state finality. A service-based prover network can provide this as a utility, unlike standalone software.

• Key Benefit: Enables new application classes with ~500ms latency guarantees.
• Key Benefit: Creates a defensible moat via network effects and integrated liquidity.

General-purpose provers are inefficient. The future is verticalized services optimized for specific workloads: ZK rollups, private computation, or oracle attestations. This mirrors the shift from generic cloud to AWS AI chips or Snowflake's data cloud.

• Key Benefit: 10-100x efficiency gains for targeted use cases (e.g., EVM, SVM).
• Key Benefit: Service providers capture value through recurring revenue and enterprise SLAs.

Open-source proving stacks like Risc Zero, Jolt, and SP1 are rapidly improving, making the core proving software a low-margin commodity. The real value shifts to operational excellence and integration.

• Hardware Advantage: Access to specialized hardware (GPUs, FPGAs) creates a ~10x performance moat.
• Economic Scale: Aggregating proofs from multiple chains amortizes fixed costs, driving down price per proof.

Platforms like Risc Zero's Bonsai and Espresso Systems abstract complexity, offering proving as a managed API. This mirrors the cloud compute evolution from on-prem servers to AWS.

• Developer UX: Teams deploy circuits, not infrastructure. Reduces time-to-market from months to days.
• Cost Predictability: Shifts model from CapEx (hardware) to OpEx (pay-per-proof), aligning costs with usage.

Networks like Succinct's SP1 Prover Network and Geohot's Axiom don't just sell software; they sell guaranteed, verifiable compute. They become trust layers for applications like ZK coprocessors and light clients.

• Network Effects: More applications → more proof volume → better hardware rates → lower prices.
• Vertical Integration: Control the full stack from client SDK to final proof settlement on-chain.

The trajectory leads to proving becoming a standardized, low-cost utility, similar to AWS EC2 or Google Cloud Run. Winners will be those who build the most resilient, cost-effective platforms with the deepest integrations.

• Commodity Pricing: Proof costs trend toward the marginal cost of electricity + hardware depreciation.
• Strategic Bundling: The service becomes a loss-leader for higher-margin offerings like sequencing or interoperability (see Polygon AggLayer, Avail).

Open-source proving stacks like Risc Zero, SP1, and Jolt are making core proving logic a public good. This erodes the moat for pure software vendors, turning their IP into a commodity. The value shifts to operational expertise and integration.

• Key Risk: Software licensing revenue collapses.
• Key Pivot: Monetize managed services, slashing client setup time from weeks to hours.

The rise of specialized hardware (e.g., Ulvetanna, Cysic) and cloud-based GPU/FPGA clusters is creating a race to the bottom on cost-per-proof. This turns proving into a low-margin utility business, similar to AWS compute.

• Key Risk: Capital-intensive hardware arms race with thinning returns.
• Key Pivot: Offer "Proofs-as-a-Service" with SLA guarantees on cost and latency, abstracting the hardware complexity.

Major L2 stacks (OP Stack, Arbitrum Orbit, zkSync Hyperchains) and RaaS providers (Conduit, Caldera) are bundling proving services into their core offering. This makes standalone proving vendors redundant for the majority of new chain deployments.

• Key Risk: Being disintermediated by vertically-integrated infrastructure.
• Key Pivot: Specialize in cross-chain or app-chain specific proving (e.g., for Avail, Celestia rollups) where bundling is less prevalent.

End-users (bridges, exchanges) don't buy proofs; they buy finality and capital efficiency. Services like Across and Circle CCTP need guaranteed settlement, not just a validity certificate. A pure prover fails to solve the full problem.

• Key Risk: Being a middleman in a stack that demands full-stack ownership.
• Key Pivot: Build or partner to offer bonded services with slashing insurance and fast-lane liquidity, moving up the value chain.

Open-source frameworks like Risc Zero, Jolt, and SP1 are making the core proving software a commodity. The value is shifting to the operational layer where performance, reliability, and integration matter.

• Key Benefit 1: Builders can launch a prover in weeks, not years, but face intense competition.
• Key Benefit 2: Investors must look beyond "another zkVM" to sustainable service moats.

The real revenue is in prover-as-a-service (PaaS) and shared sequencer models, not licensing. This mirrors the cloud transition from on-premise software to AWS.

• Key Benefit 1: Recurring revenue from throughput fees and MEV capture, not one-time sales.
• Key Benefit 2: Network effects from attracting rollups (like AltLayer, Eclipse) to your service stack.

With software commoditized, the winning prover services will compete on latency, cost efficiency, and prover diversity. This requires deep hardware optimization (GPUs, ASICs) and efficient proving markets.

• Key Benefit 1: Sub-second finality attracts high-frequency DeFi and gaming apps.
• Key Benefit 2: Cost leadership via optimized hardware passes savings to rollups, creating a flywheel.

Standalone provers will struggle. The winning model bundles proving with shared sequencing (like Espresso, Astria) and data availability (like EigenDA, Celestia). This creates a full-stack "rollup OS" with superior UX.

• Key Benefit 1: Single integration point for rollup developers simplifies stack decisions.
• Key Benefit 2: Captures value across the entire settlement stack, not just one layer.