Most people tracking AI stocks are looking at the wrong numbers.

They're watching revenue growth, user counts, maybe GPU orders. Those are fine — but they're lagging indicators. By the time revenue shows up, the move already happened.

Here are 3 leading indicators I track before a stock moves:

1. API call volume growth (week-over-week)

Forget quarterly revenue. If an AI company's API usage is compounding weekly, developers are building on it. That's future revenue locked in. You can infer this from developer community activity, GitHub integrations, and partner announcements.

2. Enterprise pilot → production conversion rate

Every AI company brags about "enterprise pilots." What matters is how many convert to full production contracts. A company with 50 pilots and a 60% conversion rate is worth more than one with 200 pilots and 10% conversion. Earnings calls bury this — dig into it.

3. Talent flow direction

Track where top ML researchers are moving. When senior researchers leave Google DeepMind for a startup, that's a signal. When they leave a startup for a big tech co, that's also a signal. LinkedIn is an underrated tool here.

I cover these signals (and more) regularly in The Silicon Brief — curated AI investing insights for people who actually want an edge.