Everyone's watching valuation multiples. Almost nobody's watching hiring velocity on technical teams pre-Series A.

Here's the pattern I keep seeing: startups that 3-4x their ML/infra headcount in the 60 days before a raise close it at meaningfully better terms than ones that scale headcount post-raise. It's a leading indicator — it means the founders already have conviction from a term sheet in hand, or they're burning runway on a bet they're confident enough to make before the check clears.

How to actually track this without a Bloomberg terminal:

• LinkedIn headcount deltas by function, not just total

• Job postings for 'founding engineer' vs generic 'software engineer' — the former signals earlier stage conviction

• GitHub org activity spikes (public repos, even internal tooling repos going semi-public)

None of this is novel data. What's novel is treating it as a leading indicator instead of a lagging one. Most people check headcount after a raise is announced to validate it. By then the deal's done.

I write about this stuff — deal flow signals, funding pattern breakdowns, technical due diligence angles — every week. If you're doing diligence on AI startups and want a sharper edge, worth a look: The Signal Loop.