When a solar company decides to mine AI instead of sunlight, the crypto world listens—but what does it hear? Sunrun, a publicly traded US solar installer, quietly launched a pilot to turn customers' home solar battery systems into distributed AI compute nodes. No token, no DAO, no smart contract. Just a traditional energy company repurposing idle hardware for inference tasks. The silence in the ledger speaks louder than code—not because nothing happened, but because what happened didn't need a blockchain.
For years, the DePIN (Decentralized Physical Infrastructure Network) narrative has been the architect’s dream: incentivize people to contribute their home routers, GPUs, or solar panels in exchange for tokens. Projects like io.net and Render Network built entire architectures around this premise. Yet here is Sunrun, a 20-year-old company with 800,000 customers, doing the same thing—without a single line of Solidity. The context matters. DePIN is not a technology; it’s a social contract. When a corporation steps into that niche, it tests whether the contract is about decentralization or just distributed compute.
I’ve audited projects that claimed decentralization but silently shipped centralization fallbacks. In 2017, I spent 120 hours on a whitepaper that promised governance tokens but hid a key distribution flaw. That experience taught me to listen to what the repository refuses to say. Sunrun’s pilot offers no code to audit. It offers no permissionless verification. But it offers something many Web3 projects lack: a working business model. The core insight is uncomfortable: distributed compute does not require a blockchain to be economically viable. The crypto layer only matters when trust is missing—and Sunrun has brand trust, legal contracts, and a billing relationship with users.
Let’s dig deeper. The pilot targets AI inference—tasks like image recognition or data preprocessing. These workloads are latency-tolerant and bursty, perfect for home hardware. Sunrun’s solar batteries already have inverters with modest compute capabilities; adding a lightweight software agent could turn 50,000 homes into a 50,000-node cluster. But here’s the catch: the coordination layer is centralized. Sunrun decides which node runs which job, validates the output, handles payments in fiat. No ZK-proofs, no slashing conditions, no on-chain dispute resolution. The void between tokens holds the true value—in this case, the value is operational simplicity.
During my time as a developer advocate at Aragon, I saw how even well-designed DAOs suffered 60% voter apathy among underrepresented groups. We fixed it by rewriting templates in plain language. Sunrun doesn’t need governance; it has a single decision-maker. That is efficient, but it is not resilient. A single point of failure? Probably. But for the users, it just works. They install the system, get a check, and never think about consensus algorithms. The contrast with Web3’s DePIN projects is stark: we built trust-minimized networks that are hard to use; they built trusted networks that are easy to use.
Now, the contrarian angle. The crypto community will rush to claim Sunrun’s pilot as validation for DePIN. I say: it’s the opposite. Sunrun proves that distributed infrastructure can succeed without crypto incentives—and that success makes the case for DePIN weaker, not stronger. If a centralized solar company can aggregate real compute supply at scale, why would an AI company pay premium token costs for verifiable but slower networks? The niche we nurtured—home compute for community—is being harvested by a corporation that doesn’t share our values. Growth without belonging is just noise.
But there is also hope. In 2022, I wrote a 10,000-word post-mortem on Luna’s algorithmic stabilizer flaws. I learned that fragility is often hidden in elegant design. Sunrun’s model is fragile in its own way: it relies on a single company’s infrastructure, a single billing relationship, and a single legal jurisdiction. A smart contract alternative, while harder to use, offers a hedge against that fragility. Open source is not a license; it is a covenant—a promise that the system can survive its creators. Sunrun can go bankrupt, sell the division, or change business priorities. A well-designed DePIN network, if it achieves true decentralization, cannot be shut down by corporate whim.
So what is the takeaway? We do not write code; we weave conviction. The Sunrun pilot is a mirror. It shows that distributed compute is real and valuable. It also shows that our obsession with token mechanisms may have distracted us from building usable products. Nurture the niche, and the forest will follow—but if the forest grows without us, we must ask: did we plant the right seeds? Or were we too busy writing whitepapers to notice the roots already spreading? Faith in the fork, hope in the merge. Sometimes the merge happens outside the ledger.