At 2 a.m., I finished configuring Newton’s VaultKit SDK with the first strategy: automatically trigger liquidation when the collateral ratio falls below 110%. I pushed it to the test environment and ran a few simulated trades—everything worked fine. But before turning off my computer, I inexplicably checked the parameter table for the strategy configuration, just to see which chain link the price data was actually pulled from.
Then all sleepiness vanished.
RedStone. Only one RedStone.
All real-time price judgment for the entire Newton policy engine is wired to a single RedStone oracle link. There is no Chainlink backup, no Pyth cross-verification, and no second data source for double confirmation. Just one line—connecting the entire treasury’s liquidation logic.
Admitting that Newton’s starting point is indeed correct. On June 23, 2026, when the mainnet Beta went live, Newton also launched the VaultKit SDK, allowing developers to set rules such as spending limits, collateral requirements, and counterparty checks. RedStone and Credora as the initial data partners were onboarded. RedStone itself says, “A policy is only as strong as the data it checks against.” That’s true—the strength of a strategy depends on the quality of the data it reads. But the problem is: when a strategy only reads one data source, the reliability of that data becomes the lifeline of the entire system.
In KuCoin’s report, it says it very plainly: “The risk that needs attention is concentration. If Newton’s policy engine overly relies on RedStone’s provided price data, any oracle outage could lead to trading freezes on the platform.” BingX’s report also mentioned the same risk: “Once oracle supply experiences an interruption, it may trigger a chain-reaction of trading freezes at the platform level.”
RedStone currently covers more than 100 blockchains and claims that “no mispricing event has occurred to date.” But “no mispricing to date” doesn’t mean “it will never happen.” Oracle downtime, data sources being poisoned, price-feed delays—these scenarios have repeated throughout DeFi history, and Newton’s strategy engine has no backup plan.
What chills me even more is this: after Newton’s TEE execution environment reads RedStone’s market data and completes the off-chain computation, it only generates a simplified ZK proof and submits it on-chain. The complete original data is not stored on-chain.
What does this mean?
If one day an automated strategy executes the wrong actions due to abnormal price data and causes users to lose money, you can’t find the complete original data on-chain. You can only see that simplified proof—it tells you that “the calculation process is fine,” but it doesn’t tell you whether the input data itself is wrong.
Was it the oracle feeding an incorrect price, did the TEE node tamper with the data, or did the contract logic have a bug? Users can never fully verify everything on-chain. And in Newton’s whitepaper, I didn’t see a detailed design for multi-oracle and multi-layer verification.
The market has already provided the answer. NEWT fell from its historical high of $0.717 in July 2025 to around $0.05 in July 2026, a drop of more than 93%. Its market cap is only about $10.18 million. For a project that has been under the halo of $90 million in financing, the quietness of this number is unsettling.
When an automated trading strategy of a protocol relies entirely on a single oracle data source to drive it, and the complete original data on-chain cannot be verified—then when we say “verifiable automation,” what exactly is being verified?
All of the above are my personal opinions and do not constitute investment advice. Feel free to discuss your thoughts in the comments.