“While most current investment is focused on foundational LLMs for text, there are significant opportunities for foundational models in other modalities like video, biology, chemistry, audio, and image.”