My thesis here is that we are going to see a bifurcation based on model type. Text based models will become a commodity since so much data is publicly available. The lack of a moat (proprietary data) and stiff competition will drive down margins. As a result, companies will increasingly open source their text based models.