Proprietary data is only valuable until someone reverse-engineers it. I've watched that happen to systems with far more resources behind them than most blockchain projects can afford.
OpenLedger's answer is confidential compute. Data processed inside secure enclaves, hardware-level isolation that prevents even the node operator from seeing what's being computed. The inputs go in encrypted. The outputs come out verified. Nothing in between is visible to anyone without authorization.
That's a real protection against a real threat. Trusted Execution Environments have a track record serious enough to take the approach at face value.
What I couldn't stop thinking about is the attack surface that exists before the data enters the enclave. Confidential compute protects computation. It doesn't protect the moment of ingestion.
That gap is smaller than the vulnerability it replaced. It's still a gap.