Sunrun's Solar Compute: DePIN's False Prophet or a Lesson in Centralized Efficiency?

The announcement is clean. Sunrun, a solar panel installer, launches a pilot. Homeowners with solar power now feed their excess energy into AI compute tasks. Distributed compute, they call it. The crypto press pounces. Another validation of DePIN. Another step toward decentralized infrastructure. I do not trust the announcement. I audit the logic. The pilot sounds familiar. Home solar → spare electricity → run AI inference jobs. No tokens. No blockchain. No trust-minimized settlement. Just a company scheduling compute on its own hardware. This is not DePIN. This is distributed compute. And the difference is not semantic—it is structural. DePIN, as defined by the Web3 playbook, requires permissionless participation, on-chain incentive alignment, and cryptographic verification of contributions. Sunrun's pilot offers none. Users sign a contract. Sunrun manages the backend. The trust model is classical: you trust the company. You trust their API. You trust their billing. The proof is silent; the code screams the truth. But here, the code is closed. The audit is impossible. Let me unpack the technical gap. In 2017, I spent six months optimizing the Groth16 prover in Zcash's Sapling. The core challenge was not computation—it was verification. For a distributed compute network to be trustless, the network must verify that a node actually performed the assigned work. This requires either a TEE (which introduces hardware trust), a ZK-proof (which is expensive), or a replicated execution (which is redundant). Sunrun's pilot likely uses a simple API call: "job sent". If the node returns garbage, you only find out when the client complains. Web3 projects like io.net and Render Network attempt this with on-chain proofs. But they face high gas costs and latency. Sunrun avoids that complexity entirely—by being centralized. The question is: does that make them better or just cheating? From a cost perspective, centralization wins. No gas. No proof generation overhead. No slashing. No governance disputes. Sunrun can offer cheaper compute than any permissionless network. The pilot will likely succeed on unit economics. The crypto community will then celebrate it as a "real-world use case". They will miss the irony. But the contrarian angle is sharper. If Sunrun's centralized model succeeds, it validates distribution, not decentralization. The narrative around DePIN gets diluted. Investors may confuse the two. They will pour money into tokens that promise "the Sunrun model with crypto", ignoring that crypto adds friction without offering tangible advantages in this context. The core of DePIN—trustless coordination—is irrelevant when one party controls both infrastructure and payment. I have seen this pattern before. In 2020, Compound's reentrancy vulnerability looked like a timing bug. But the root cause was trust in immutable logic. The same illusion now: "distributed" feels like "decentralized". It is not. Sunrun can flip a switch and shut down the whole compute network. No validator set. No exit game. No fork. Where does this leave Web3? The infrastructure side of DePIN—iotex, helium, hive mapper—has always struggled with the physical world: device cost, battery life, real estate permissions. Sunrun solves one part: cheap electricity. But they replace the trust model with a corporation. That is a step backward for crypto maximalists. I do not trust the contract; I audit the logic. The logic of Sunrun's pilot is clean. It will probably generate revenue. But it is a fortress, not a network. The moment Sunrun decides to raise prices or censor certain clients (regulatory pressure, for instance), the users have no recourse. No smart contract enforces the deal. No DAO votes. In 2022, I wrote a 10,000-word report on Lido's validator centralization. The risk was not technical—it was governance capture. Lido's node operators were concentrated. The same risk reappears here: one company, one backend, one point of failure. So what is the takeaway? DePIN advocates should stop claiming every distributed compute initiative as their own. If Sunrun's model works, it will reveal that the market values cheap compute over trustlessness. That is a bitter truth for those who believe "code is law". The crypto-native solutions must compete on latency, cost, and settlement speed—not just on ideology. I am not bullish on this pilot for crypto. I am neutral. It is a sandbox for traditional players. Web3 gets a benchmark. If Sunrun later integrates a token or an on-chain settlement layer, I will revisit. Until then, the proof is silent.