Terravek Labs

Terravek Labs was founded in February 2018 and is 28 people across Delft and Toulouse. It exists to make the measurement better, and to make sure Terravek is not caught out in five years by a capability it should have started building today.

What we fund

Three categories, in descending order of budget:

1. Methods that improve existing products. Atmospheric correction, uncertainty quantification, co-registration. Clear path to product.
2. Capabilities customers will need in three to five years and cannot buy today. Persistent scatterer interferometry was in this category in 2019 and is now in three products.
3. Work with no commercial thesis at all, deliberately, because a research group that only does directed work stops attracting the people who make the directed work good.

How we stop

Every project has a written hypothesis and a review at six and eighteen months against pre-registered criteria. We killed four projects in 2025. Killing them well — with the negative result written up and published — is the part most organisations skip. Élodie Marchand on the funding model.

Publication policy

Labs publishes openly. Where a method reaches product, the method is published; where a result is negative, it is published. Every paper with a computational component has a reproduction repository with a pinned environment and a single command that reproduces the headline figure. Open data and code.

People