It is commonly agreed that the next era of software opportunity lies in using AI to power products that automate or enhance processes previously handled by humans. What sets the next wave of billion-dollar startups apart will be their ability to harness proprietary data to train AI models. Companies that can not only utilize publicly available or paid-for data (like everyone else) but can also generate unique and meaningful proprietary data are creating data feedback loops that lead to long-lasting competitive advantages. These advantages have always been valuable, but with the advent of LLMs, their power is explosive. At ValueStream Ventures, we have been building portfolios of data-diff