Talk notes
Climate tech is the sector where our signal works least well, and that's worth being honest about. We expect this to be our weakest sector, primarily because climate-tech companies are hardware-heavy and do less of their engineering in public repositories. A battery-chemistry team or a direct-air-capture team often has a small public-facing tools repo and a much larger private engineering codebase we never see.
The trade-off is that when the signal does fire in climate-tech, the lead time may be the longest in the panel. The reason is structural: climate-tech fundraises are slower, more diligence-heavy, and more sensitive to government grant cycles. So when our signal catches a real climate-tech fundraise, it tends to catch it earlier in the lifecycle.
The dominant pattern is the simulation-first team. A founding team builds a custom simulation framework (battery-cell modeling, atmospheric chemistry, grid optimization) and publishes it as a tools repo. Commit velocity is sparse for 6-12 months, then spikes sharply when the team is preparing for a major demo or fundraise milestone. The spike is often accompanied by a flurry of patent filings, which we don't measure directly but recommend integrating as a secondary signal.
The Q1 2026 cohort included 6 named climate-tech orgs that flagged the composite signal. Four of them announced a fundraise within the 90-day window. Two are still pending, climate-tech timelines are slow enough that 'still pending' is consistent with a true positive. We'll know more by Q3.