**Compute as a commodity. AI adoption is constrained by hardware **

Many companies face significant challenges in adopting AI due to inadequate infrastructure and a lack of investment in necessary talent. Consequently, they often resort to leveraging APIs from major tech companies or utilizing open-source models to enhance their capabilities. A notable trend is the increasing reliance on compute resources for training large models, which is essential for applications like recommender systems. The transition from smaller organizations to larger tech entities highlights the stark contrast in access to compute resources and the ability to train sophisticated models, emphasizing that smaller companies may struggle to keep pace with advancements in AI.