1Q25: Our Research Methodology and AI Company Performance - Part 3

This earnings season, we published 15 preview reports on AI stocks. In chronological order, based on the accuracy of our earnings predictions versus stock price movements, our track record was: META (✓), MSFT (✓), AWS (✓), PLTR (✓), APP (✓), MNDY (✓), Tencent (✓), AliCloud (✗), SNOW (✓), OKTA (Mixed), C3.AI (✗), CRM (✓), NVDA (✓), MDB (Mixed), RBRK (✓).

Here, ✓ indicates our directional assessment aligned with stock performance, ✗ indicates our prediction contradicted stock performance, and "Mixed" indicates our preview contained highly mixed views that were difficult to judge accurately.

We also completed several in-depth reports including IT Budget 1Q25 survey, China AI - Demand Research, China AI - CSP, China AI - Fintech, Apple Payment Policy Changes, LLM 1Q25 Deep Dive, Robotaxi 1Q25 Deep Dive, and China AI - Robotaxi.

We aim to review the quarter chronologically to help everyone understand earnings trends, how our fundamental research helps interpret AI industry developments, and how our research improves earnings prediction accuracy.

In such complex macroeconomic conditions, 11 of our 15 preview reports correctly aligned with stock price movements, achieving an impressive 73% success rate. Notably, when most observers were pessimistic about the Magic 7 stocks entering earnings season, we predicted from the outset that META and MSFT results would surprise everyone positively, which became the prelude to this quarter's U.S. market reversal.

Deep|OKTA 1Q25: Modest Beat the Upper-end of Guidance, But High Potential Ahead in AI Agent Era