Succinct's Prover Network Is Doing More Than Rollups
Succinct has honestly built up a narrative as one of the foundational infrastructure layers for the Ethereum rollup ecosystem: 6.3M proofs generated, $3B secured value across 31 live provers running on chain is a rollup play through and through. Coverage I've seen of succinct crypto revolves almost entirely around this rollup narrative. Partnerships with Optimism, Arbitrum, Base, Polygon cover ~90% of the rollup market by TVL. But SP1 zkVM (any program compiled to a virtual machine that can produce zk proofs, written in Rust) can prove block transitions isn't the only game in town. The builders are already putting it to work proving applications well outside of the rollup narrative. What use cases are they looking at? They paint a very interesting picture of potential PROVE demand outside of rollups.
Five proving demand categories beyond rollups. Sources: Succinct project documentation, public partnership announcements.
Private DeFi Settlements Without Centralized Coordinators
Traditional DeFi settlements announce their transactions to the entire mempool audience. Solvers, searchers, and MEV bots extract value from that transparency. Succinct's new private proving technology, where proofs are generated inside of a Trusted Execution Environment (TEE), allow for settlements that can be publicly verified without revealing the details of the trades. Production ready. Merging TEE-based proving with SP1 Hypercube (the formally verified zkVM that launched on mainnet in Feb 2026), developers can build settlement layers with the DeFi protocol itself producing a proof that trades happened at correct prices, accounts were settled properly, and slippage parameters were upheld. Without any centralized oracle viewing raw order flow. Formal verification of all 62 core RISC-V opcodes was done in partnership with Nethermind Security and the Ethereum Foundation. Giving these private proofs a security baseline that prior zkVMs didn't have. The privacy layer changes the game for DeFi protocols.
DEX which privately settles trades using the Succinct prover network can credibly claim MEV resistance without trusting a sequencer. Each proof requires network resources to process, so each settlement represents computational demand for provers (and the PROVE token which organizes the network). Private settlements are just one component of the privacy explosion. The same TEE proving architecture unlocks a whole new category of cross chain communication.
Cross-Chain Messaging That Preserves Privacy
Most bridges read some state on a source chain, then publish that same state on a destination chain. This verification step is public - a user who bridges assets from Arbitrum to Polygon, for example, would have a readable footprint on both chains. SP1's creation of these proofs inside a TEE allows a bridge protocol to verify source chain state, prove that asset locks happened, and generate a proof of validity, without disclosing the user's address or amount on the destination chain. Succinct Labs has already integrated its product into most of the largest L2 ecosystems. The Optimism partnership grants OP Succinct preferred proving status on the Superchain. The Base Azul upgrade set for May 13 will power $7.4 billion in deposits using SP1.
At the time of writing, Arbitrum's exclusive partnership with Tandem will last until August 2026. Each of these integrations lay down the rails necessary for privacy preserving cross-chain messages to flow. If a dApp wants to build on any of those L2s, they can route their cross-chain calls over Succinct and prove validity of the transaction without leaking any metadata. This allows users to transact across chains with the cryptographic guarantee that their transaction will be correctly relayed, without creating a publicly visible connection between their source and destination wallets.
The growth to 54 million transactions traversing the network already is a testament to the throughput capability. It remains to be seen if the teams behind those bridges want to lock-in at scale to private proving functionality. That decision will be largely dictated by user demand for privacy, which is a highly elastic function as on-chain analytics solutions become more sophisticated. Privacy and verification is not limited to money movement. Google recently deploying SP1 opens up a completely different vertical.
AI Model Verification on Blockchain
A paper published on April 13 by Google's Quantum AI team showed that quantum resources needed to attack some cryptographic primitives could be reduced by a factor of 20. Leveraging SP1 to produce zero-knowledge proofs of the output of quantum circuits (enabling a third party to verify that result without seeing the circuit itself), an independent collaboration between Google, Ethereum Foundation, Stanford, and UC Berkeley later showed SP1 can be used to prove claims about computations outside of the traditional realm of blockchain. In this case the zero-knowledge element is not being used to obscure a token transfer. Instead, it's being used to prove that some complex AI or quantum computation returned a certain result in a way that can be verified by others, but without revealing the model architecture itself. Blockchain security firm Trail of Bits found memory safety bugs in Google's Rust implementation of the ZK proof system during an audit in April 2026, but Google addressed these issues promptly and they did not affect the scientific claims.
The attack nevertheless highlighted formal verification's usefulness: SP1 Hypercube's formally verified opcode set limits attackers' attack surface area compared to non-verified opcodes. The use-case for AI model developers is intuitive. A machine learning model makes an inference (a prediction, classification, recommendation), SP1 creates a proof that the inference was generated by a certain model with certain weights, without revealing the weights themselves. On-chain contracts can accept that proof and act upon it. Fraud detection models, credit scoring algorithms, insurance risk assessments: all can be used as candidates for trustless on-chain verification. Every proof generation transaction is sent to the Succinct prover network, which requires computational resources that provers are rewarded for with PROVE. If AI verification sits at one high-complexity end of the spectrum, gaming sits at the other. Instead of complex proofs, the challenge here is fast proofs.
Gaming State Proofs That Don't Kill Performance
On-chain games face a Catch-22. State transitions (move a character, update an inventory, resolve a battle outcome) need to be proven, but if each update required its own transaction it would be a horrible experience for the player. SP1 Hypercube shatters this trade-off with its performance. Researchers from Succinct proved that 16 RTX 5090 GPUs were able to prove 99.7% of Ethereum blocks in less than 12 seconds. That kind of performance allows a game server to run hundreds of state transitions locally, generate one proof that all of the transitions were valid per game rules, then only post the proof on-chain. Players get nearly instant feedback as they play, and the blockchain receives a small cryptographically verified record of the final game state.
This FPGA acceleration work done by AntChain OpenLabs that enabled SP1's 20x speedup pushes the baseline even higher on purpose-built gaming infrastructure. The model can scale to any game where trustless state verification is applicable: competitive esports with prize pools, prediction markets backed by game outcomes, NFT-based economies where the game needs to prove that it's honestly playing by the rules around item creation, etc. Game devs don't need to build any custom circuits. They write their game logic in Rust, compile it for SP1, and send the proving jobs to the network. The prover network takes care of the rest.
What These Use Cases Mean for PROVE Token Demand
Private DeFi settlements, cross-chain privacy, AI verification, gaming proofs, and every application using the rollup that's taking off right now all require the same thing: computation on the Succinct prover network. The 31 live provers in network earning PROVE by doing jobs right now creates a tangible connection between network demand and token utility. Trading at $.2694 per PROVE token with a market cap of $52.5 million (of the 1 billion token supply, only 200 million tokens are currently in circulation), PROVE has been mostly priced by the already compressed crypto market as a rollup infrastructure play. The tokens price increased 10.6% in the past seven days creating short term momentum. PROVE is 84.2% below its all time high price of $1.71.
But could non-rollup use cases push that valuation much higher? It's all a question of scale. Rollup proving alone is already handling billions of dollars worth of secured value. Use cases like private settlements or AI verification are just starting out in comparison. The work with Google made waves, and Zcam's recently announced iPhone app, which verifies photo authenticity against deepfakes, launched on April 24 as a real-world consumer-facing application. But neither have created anything close to the kind of daily, baseline proving volume that integrations like Optimism or Base will generate on day one. Another wild card is the token's vesting schedule. 65%+ of total supply is held by "Core Contributors, Investors and R&D" which is subject to a 12 month cliff, followed by a 36-to-48 month linear vesting schedule. New use cases would have to generate the required proving demand in short order to absorb that sudden increase in supply.
Driving proving demand to governments, financial institutions, and Fortune 500 companies is a major part of Succinct's stated vision for 2026. Mainnet launch of Base Azul on May 13 will be the first opportunity we'll have to see if expectations for broader proving demand outside of rollups for the Succinct token network pan out in a way that can be reflected on-chain. And whether the PROVE token price begins to reflect the utility of such massive increases in usage of infrastructure orders-of-magnitude bigger than rollups.
