Author

@jiahanjimliu jiahanjimliu

Deep AI-infrastructure specialist reasoning from compute, power, and stack economics

Publishes original, mechanistic fundamental theses on AI dat

trader score
+0.46
hit rate
40%
mean α
+0.45%
signals 14d
31

Grade = how their written analysis reads (A best). Trader score = how their last-20 timestamped calls performed vs SPY.

Their picks, scored

Across their last 20 scored bets: 40% hit rate, +0.45% mean alpha, trader score +0.46. Their last-14d mentions, direction-adjusted, have moved -4.7% since posting (mean over 26 mentions with price data).

Recent signals15receipts included
date (PT)tickerauthorsentwhat they saidsince thenreceipt
2026-08-27IREN@jiahanjimliu+0.60Bullish thesis that owned data centers create a financing flywheel for future AI infrastructure expansion.-12.5%
2026-08-26IREN@jiahanjimliu+0.00Flags IREN earnings for tomorrow and an accompanying revenue estimate explanation.-10.4%
2026-08-26ORCL@jiahanjimliu-0.30Reports a credible rumor of a second Oracle layoff round scheduled for September 1.+1.3%
2026-08-25NVDA@jiahanjimliu+0.10Ranks photonics above memory and storage, favoring LITE, MU and SKHY over SNDK.+2.1%
2026-08-25SNDK@jiahanjimliu-0.55Ranks photonics above memory and storage, favoring LITE, MU and SKHY over SNDK.+0.3%
2026-08-25LITE@jiahanjimliu+0.70Ranks photonics above memory and storage, favoring LITE, MU and SKHY over SNDK.+1.1%
2026-08-25SKHY@jiahanjimliu+0.55Ranks photonics above memory and storage, favoring LITE, MU and SKHY over SNDK.·
2026-08-25MU@jiahanjimliu+0.55Ranks photonics above memory and storage, favoring LITE, MU and SKHY over SNDK.-0.0%
2026-08-24NVDA@jiahanjimliu+0.55Argues Rubin Ultra memory capacity steers model design toward FLOP-heavy techniques benefiting Nvidia.+4.4%
2026-08-23NVDA@jiahanjimliu+0.30Discloses an IREN investment while analyzing NVDA's capture of GPU-efficiency margins.+1.3%
2026-08-23IREN@jiahanjimliu+0.20Discloses an IREN investment while analyzing NVDA's capture of GPU-efficiency margins.-15.4%
2026-08-23NVDA@jiahanjimliu+0.10Relays a labeled rumor that Kimi K3 was trained on Nvidia GB300 systems in Thailand.+1.3%
2026-08-23·@jiahanjimliu·Argues open-source models win token volume while closed models retain superior profit margins.·
2026-08-21NBIS@jiahanjimliu+0.10Frames power as the AI bottleneck and favors IREN and BE among potential beneficiaries.-4.5%
2026-08-21IREN@jiahanjimliu+0.65Frames power as the AI bottleneck and favors IREN and BE among potential beneficiaries.-15.4%

Grade is our human read-worthiness rating; trader score is a rolling 20-bet hit-rate/alpha composite — different things, often disagreeing. “Since then” is direction-unaware in the table; the summary line above adjusts for which way they leaned.