Data Startups

Startups that rely on data models and AI/data science algorithms for running their business will only get better with time. Moreover, as time passes they have leverage over newcomers. Making it really hard to overthrow a leader in a vertical. In this space I think there can be 2 strategies. Either the company can choose to niche down or expand to other verticals. If they niche down, over time they will collect niche specific data and develop models that are effective in that small space. One way such a company can be overthrown is if there is a new technology that doesn’t require as much data to predict consumer behaviour or whatever. While it sounds possible that still seems quite hard. Even if models get efficient and computers become more powerful and be cloud-first. There is a bottleneck in available data. There is only so much real world data that you can harvest. Out of that only so much that you can label and sanitize it so it can be used for training purposes.