By Thomas | financial enthusiast
What We Are Testing
Seven weeks ago, I launched three autonomous AI trading bots — one for Micron Technology (MU), one for Solana (SOL), and one for Alphabet/Google (GOOGL). Each bot runs five distinct algorithmic strategies simultaneously, starting with $100,000 in capital per strategy. All trades are fully autonomous, executed without human intervention.
The five strategies:
| Strategy | Style | Core Logic |
|---|---|---|
| Trend-Following | Directional | Buys momentum, rides sustained price moves |
| Mean-Reversion | Counter-trend | Fades extremes, expects return to average |
| Momentum | Quantitative | Ranks assets by recent relative performance |
| The Reader | News-driven | Trades on sentiment extracted from financial news |
| The Thinker | Macro reasoning | Uses fundamental and macroeconomic context |
No cherry-picking. Every strategy runs on every asset. Every trade is logged. This is the live data.
Micron Technology (MU) — 17 Days Live
Start date: 2026-05-19. Starting price: $698.70 per share.
After 17 trading days:
| Strategy | Current Equity | Return |
|---|---|---|
| The Reader | $109,422 | +9.42% |
| Trend-Following | $108,037 | +8.04% |
| Mean-Reversion | $102,881 | +2.88% |
| Momentum | $100,000 | 0.00% |
| The Thinker | $97,551 | −2.45% |
The Reader leads by a meaningful margin. It processes financial news and earnings-adjacent sentiment signals — and Micron has been a high-news environment since May, with DRAM demand commentary, AI server orders, and guidance revisions moving the stock. The Reader is built for exactly this kind of noise-rich environment.
Trend-Following is close behind at +8.04%. MU has moved in a sustained directional pattern over the test window, which is the environment where trend strategies shine. Mean-Reversion has generated positive returns but underperformed: betting against the trend in a trending market is structurally punished.
The Thinker is the only strategy in negative territory at −2.45%. Macro reasoning has struggled to translate broad semiconductor thesis into timely entries. The Thinker is a slow-process strategy — its edge may appear over longer time horizons.
Momentum has not opened any positions yet. By design it waits for a cross-asset ranking threshold before deploying capital.
Solana (SOL) — 13 Days Live
Start date: 2026-05-23. Starting price: $82.30 per SOL token.
After 13 trading days:
| Strategy | Current Equity | Return |
|---|---|---|
| Trend-Following | $100,897 | +0.90% |
| The Thinker | $100,350 | +0.35% |
| The Reader | $100,180 | +0.18% |
| Momentum | $99,530 | −0.47% |
| Mean-Reversion | $98,860 | −1.14% |
| Benchmark (buy-and-hold) | 1,215.07 SOL bought at $82.30 | — |
SOL returns are tightly clustered near zero — this reflects how choppy the Solana market has been since late May. Trend-Following leads with +0.90%, but the spread between best and worst is only 2.04 percentage points. No strategy has made a decisive move.
Mean-Reversion is the worst performer here as well: −1.14%. SOL does not have the same mean-reverting character as large-cap equities. The strategy is fighting the asset's volatility structure.
At 13 days, SOL results are inconclusive. The real differentiation will emerge over 30–60 day windows when the strategies accumulate enough trade cycles to show edge.
Alphabet/Google (GOOGL) — 6 Days Live
Start date: 2026-05-30. Starting price: $380.39 per share.
After 6 trading days:
| Strategy | Current Equity | Return |
|---|---|---|
| Momentum | $100,000 | 0.00% |
| The Reader | $100,000 | 0.00% |
| Trend-Following | $99,570 | −0.43% |
| The Thinker | $99,450 | −0.55% |
| Mean-Reversion | $98,950 | −1.05% |
| Benchmark (buy-and-hold) | 262.89 shares bought at $380.39 | — |
Six days is not enough to draw conclusions. Momentum and The Reader have not yet deployed capital. Trend-Following and The Thinker have small open losses. Mean-Reversion is the worst performer at −1.05%.
GOOGL is a fundamentally different animal from both MU and SOL: a $2.3 trillion mega-cap with lower daily volatility and different liquidity dynamics. Whether strategies built for smaller-cap or crypto environments can translate to GOOGL at the same return profile is one of the central questions this test will answer.
What the Data Shows After 17 Days
One pattern is already visible: Mean-Reversion is the weakest performer across all three assets. It is negative in two of three bots and the weakest positive in the third. This is not random noise — it is structurally consistent. Mean-reversion strategies require a range-bound, oscillating price environment. All three assets have been trending (MU sharply upward, SOL with directional moves, GOOGL in a tight range that punishes the fading of small moves).
The Reader is the standout on MU at +9.42% in 17 days. Whether this reflects genuine news-signal edge or a favorable stretch of news flow is the question. The strategy needs to be tested through a quiet news period to verify the alpha is real.
Trend-Following is consistent: leading on SOL and second on MU. If the pattern holds across GOOGL, it will be the most robust general-purpose strategy in the portfolio.
Open Positions and Trade Count
All current equity figures reflect unrealized P&L only. No strategy has closed a completed round-trip trade. Every position is still open. This is important context: the returns shown above are mark-to-market, not locked-in gains. They can and will fluctuate.
The absence of closed trades means the realized performance column is empty. Actual realized returns will differ from current mark-to-market figures — often significantly, as exits are where strategy design is most severely tested.
What I Am Watching
The 30-day marks will be the first meaningful performance checkpoint. At that point, MU will have enough trade cycles to show whether The Reader's edge is repeatable, and SOL will have accumulated enough data to separate skill from noise.
Key questions for the next 30 days:
- Does The Reader maintain its lead on MU as the AI chip news cycle evolves?
- Does Trend-Following translate its SOL and MU performance to GOOGL?
- Does Mean-Reversion find a period of range-bound price action where it can recover?
- When does Momentum finally deploy capital — and in which direction?
I will publish updated performance tables monthly. Every number above is live, sourced directly from the bot databases, updated as of 2026-06-05.
