CHF Income Plus and USD AI Yield Plus are being rebuilt on isolated policy, cash, position, checkpoint, and evidence contracts. Historical and live performance figures are temporarily withheld while the resolved mark-to-market backtests and January 2026 live ledgers are independently reconciled.
The isolated USD book is being replayed from 1 January 2026 against its versioned policy and exact execution record. Live figures will return only after positions, cashflows, model calls, and weekly MTM reconcile.
The 2016 to 2025 backtest is being recomputed with no leverage, gross accounting, a maximum six-month tenor, resolved positions only, and fresh MTM. AI receives the identical locked Quant shortlist and may rank or veto; it cannot change product terms, notional, or risk limits.
The isolated CHF book is being replayed from 1 January 2026 under an explicit CHF product and currency profile. No USD or historical artifact is accepted as CHF live evidence.
The CHF mandate is optimized for breach avoidance, low drawdown, and diversification. Its 2016 to 2025 evidence will be published only after the complete CHF cash, hedge, lifecycle, and MTM ledger reconciles with no unresolved positions.
A CHF mandate for conservative capital, with breach avoidance, drawdown control, and diversification ahead of coupon maximisation. The currency hedge, note lifecycle, and resolved ten-year evidence are being reconciled before any return or loss statistic is republished.
A higher-yield USD mandate. Quant builds the causal shortlist; a temperature-zero AI desk may rank or veto candidates inside fixed product, notional, concentration, and risk limits. The resolved ten-year result is being recomputed before publication.
Structured-product issuance relies on standardised pricing engines built for distribution scale. They are calibrated to the average. The firm operates the underwriting layer those engines omit.
The firm operates a systematic underwriting cycle, not a discretionary trade desk. The same quantitative quality gates run every week against the same as-of-that-day market data. Allocation decisions are escalated to human review only at predefined inflection points; routine deployment is mechanical. Every historical retrospective uses the exact code that runs today, with no forward-looking inputs.
Every deployable position is traced from the canonical opportunity surface, through the deterministic MC-ranked selection pipeline, to executor validation. The deployment authority is the systematic ranker; no hidden side-channels can bypass it. For strategy variants that route through AI review, every AI selection is similarly required to clear the same executor validation before deploy.
Each position is executed in the standardised institutional autocall format. Capital recycles on a known schedule.
Every future performance number on this page will bind to exact input objects, policy and code versions, a resolved event ledger, fresh weekly MTM, and an independent recomputation. Historical options coverage and coupon calibration are disclosed rather than silently filled.
The strategy itself is simple to state. We sell insurance against large falls in pairs of quality stocks, collect the premium as coupons, and hold notes that repay in full unless a stock halves. Most weeks the system looks at hundreds of candidates and buys a handful; saying no is most of the work. The AI fund is allowed to reach for more yield inside hard limits it cannot change. The franc fund refuses anything the market prices as risky, and in ten simulated years it never paid for that caution with a loss.
Pick your pace, or hold both: dollars with an AI desk working for yield, or francs with a rulebook that has not produced a loss in ten simulated years. Either way, you can read exactly what your money is doing every week.
Two products, one engine: CHF Income Plus (hedged francs, live fund) and AI Yield Plus (AI managed, USD, live fund).