The advent of transformers, higher computational budgets, and big data has engendered remarkable progress in Natural Language Processing (NLP). Impressive performance of industry pre-trained models has garnered public attention in recent years and made news headlines. That these are industry models is noteworthy. Rarely, if ever, are academic institutes producing exciting new NLP models. Using these models is critical for competing on NLP benchmarks and correspondingly to stay relevant in NLP research. We surveyed 100 papers published at EMNLP 2022 to determine whether this phenomenon constitutes a reliance on industry for NLP publications. We find that there is indeed a substantial reliance. Citations of industry artifacts and contributions across categories is at least three times greater than industry publication rates per year. Quantifying this reliance does not settle how we ought to interpret the results. We discuss two possible perspectives in our discussion: 1) Is collaboration with industry still collaboration in the absence of an alternative? Or 2) has free NLP inquiry been captured by the motivations and research direction of private corporations?