⚡

AlphaAiEngine

$100K AI Trading Experiment · Q2 2026

$100,000.00

▲ +0.00%

Challenge

$100K vs S&P 500

Period

Apr 1 – Jun 30, 2026

Live

🤝 Super FA Team

Two of the world's most advanced AI models — Claude Opus 4 and GPT-5.4 — working in tandem as a Super FA Team. One proposes trades. The other reviews every recommendation as an independent second opinion — confirming, modifying, or rejecting each call. Only consensus trades execute. No human intervention. No emotion. Just data, analysis, and conviction.

Portfolio Overview

Portfolio Value

—

Loading…

Today's P&L

—

—

Total Return

—

vs $100K start

Alpha vs S&P 500

—

SPY: —

Performance vs S&P 500

AlphaAiEngine SPY Benchmark

Daily P&L

Up Day Down Day

Current Holdings

📊

Loading positions…

Performance Stats

Days Remaining

—

Days

— days elapsed 91 total days

Sharpe Ratio—

Sortino Ratio—

Max Drawdown—

Calmar Ratio—

Win Rate—

Win / Loss Ratio—

Alpha (vs SPY)—

Beta—

Avg Winner—

Avg Loser—

Best Day—

Worst Day—

Streak—

Days Since Trade—

Total Trades—

Long vs Short—

Annualized Vol—

Avg Daily Return—

Model Disagreement —

Disagreement Rate—

Full Veto Overrides—

Trades Today—

Portfolio Turnover—

Daily Schedule (ET)

9:35 AM Morning Analysis & Trades (Claude + GPT-5.4)

1:00 PM Midday Risk Review (Claude)

3:45 PM Pre-Close Review (Claude)

4:15 PM End-of-Day Snapshot

Auto-refreshes every 60s

Sector Allocation

🗂️

No positions to allocate.

AI Conviction Summary

🤖

No trade data yet.

Trade Log

📋

No trades yet. The AI will execute its first trades on the next market open.

AI Daily Commentary

🤖

Loading AI commentary…

How It Works

A fully autonomous AI trading system running a live $100K portfolio against the S&P 500 — no human hands on the wheel.

🧪

The Experiment

Starting capital: $100,000
Timeline: April 1 – June 30, 2026 (Q2)
Benchmark: S&P 500 (SPY)
Goal: Can AI consistently outperform the market?
All trades are logged with complete AI reasoning
Performance is benchmarked against SPY daily

🧠

The AI Engine

Dual-AI architecture: Claude Opus 4 analyzes markets and proposes trades. GPT-5.4 reviews every recommendation as an independent second opinion — confirming, modifying, or rejecting each trade.
When both AIs agree, the trade executes. When they disagree, the more conservative position wins.
Every trade includes written reasoning explaining the thesis.
The AI reviews its own track record each morning before making new decisions.

🕐

Daily Schedule

🌅

9:35 AM ET

Morning Analysis

Full market scan, news analysis, portfolio review. Claude proposes, GPT-5.4 reviews, trades execute.

🔍

1:00 PM ET

Midday Risk Review

Quick check for positions down 5%+ or up 8%+. Cut losses or lock in gains if needed.

🔔

3:45 PM ET

Pre-Close Review

Overnight risk assessment. Close anything that shouldn't be held through the bell.

📊

4:15 PM ET

End of Day

Portfolio snapshot and performance recording.

📈

What Can It Trade?

US stocks — any market cap
ETFs — sector, index, leveraged, inverse
Long and short positions
Options strategies (calls, puts, spreads) when appropriate
Maximum 20 concurrent positions
Minimum 5% cash reserve at all times

📋

The Rules

Fully autonomous — no human intervention in trading decisions
All trades are logged with complete AI reasoning
Portfolio is benchmarked against SPY daily
Performance metrics updated in real-time
The AI adapts — it reviews its own track record each morning before making new decisions
Maximum single position: 25% of portfolio

💡

Why This Matters

Most AI trading research is theoretical — this is live, with real market data
The dual-AI review process mirrors institutional risk management
Every decision is logged and explainable — no black box
A 91-day window is long enough to test real market conditions including volatility, earnings season, and macro events
Results are fully transparent and updated in real-time on this dashboard