Seed Stage · AI-Sector Equity Rotation
SafeAITravels systematically ranks and holds the strongest name in each layer of the AI buildout — chips, cloud, energy/cooling, cybersecurity, automation — using a repeatable scoring formula, then manages every position through a pre-defined, volatility-adaptive tier ladder. Not a stock picker's opinion. A system.
The problem
Retail and even professional portfolios pile into a handful of mega-cap AI names with no systematic entry, sizing, or exit rule — conviction stands in for process.
A single bad earnings print becomes an all-or-nothing decision: hold and hope, or panic-sell. There's rarely a pre-committed plan for what happens at -20%, -40%, -60%.
Tracking 22+ names across 8 sub-sectors, rescoring fundamentals, and watching tier triggers daily is more than a part-time investor — or most advisors — can sustain.
The solution
SafeAITravels systematically ranks and holds the strongest name in each layer of the AI buildout using a repeatable scoring formula, then manages every position through a pre-defined, volatility-adaptive tier ladder for entries, adds, and exits.
22 names across 8 AI-economy categories, scored on revenue growth, valuation, analyst upside, and beta — rebalanced quarterly.
Every position sizes in through volatility-adaptive tiers (25/32/43% of allocation) with hard stops — no unbounded exposure to a single thesis.
Signal generation, scoring, and monitoring are fully automated; no order is ever submitted without an explicit human confirmation.
How it works
22-name AI universe re-ranked each quarter on a composite of growth, valuation, analyst upside, and low-beta factors.
Candidates must clear both a profitability bar (positive historical expectancy) and a correlation screen (max 0.4 pairwise).
Positions enter at 25% of slot allocation (Tier 1), adding only on further favorable volatility-adjusted moves.
Pre-committed profit targets and a two-level stop replace discretionary hold/sell decisions.
Research evidence
Per-trade expectancy re-tested on a recent (2023–2026) window — not just a favorable historical episode — to separate a durable edge from a lucky backtest.
PLTR — avg. per-trade return, most durable candidate (2023–2026 window)
CVS — avg. per-trade return, recent-window survivor of both filters
Max pairwise correlation enforced across concurrent positions
Candidates that failed the profitability filter outright and were excluded before ever reaching the correlation screen
Of simulated trades ever reached the terminal hard-stop across the full historical test set
AI-economy categories tracked, from chips and cloud to energy/cooling and cybersecurity
Backtested, hypothetical results based on historical data; not a guarantee of future performance.
Where we are today
We are deliberately early. The priority so far has been proving the mechanics — scoring, tiering, risk controls, and reporting — work end-to-end with real broker data, at deliberately small live capital, before scaling.
Direct IBKR connectivity for pricing, positions, and execution — currently running at intentionally minimal live capital to validate real-world fills against the model.
A larger paper account runs the identical rule set alongside the live account, so live execution can be compared against theoretical performance before capital is scaled.
Weekday tier-checkins run automatically, screening all open positions against the ladder and producing a status report — but every trade still requires explicit human confirmation.
Every trade, tier trigger, and ranking refresh is logged — the discipline needed for a real fund structure and for outside diligence.
Let's build the track record together
Not a single bet — a system. We're raising seed capital to take it from a validated framework to a proven track record.