NoahAI Technical Whitepaper v1.9: key changes explained
July 2026
What changed in v1.9
Whitepaper v1.9 is not a cosmetic update. It clarifies how NoahAI should be described: not as a short-term return machine, but as an operational financial AI infrastructure with explicit judgment, control, and replay.
1. HOLD is active judgment
In NoahAI, HOLD does not mean “nothing happened.” It means the system evaluated market state, expected value, volatility, recent performance, and risk conditions, then concluded that not trading is the better decision.
2. Decision layer and execution layer are separate
v1.9 explains NoahAI as a stack where a shared decision layer generates judgment and a shared execution layer standardizes broker, exchange, fill-state, and fee differences. This is what allows one operating structure to extend across multiple asset classes.
3. KPI is about operation, not bragging rights
The whitepaper also clarifies how public KPI should be read: not as a leaderboard, but as an operational validation surface. Win rate, drawdown, early exits, HOLD ratio, and the split between Trading PnL and Economic PnL should be read together.
Why this matters
v1.9 makes NoahAI easier to explain to partners, public R&D reviewers, and institutional diligence teams. The key claim is not “AI trades for you,” but “AI judgment is structured, controlled, recorded, and replayable.”