Tiffany Luck, a partner at New Enterprise Associates, outlined how vertical AI startups build defensible advantages against dominant platforms. Luck identified the core challenge facing founders: competing in AI when OpenAI, Google, and Anthropic control foundational models. The path forward centers on domain expertise and data moats rather than raw compute power.
Startups win by solving specific problems in narrow verticals. A vertical AI company targeting radiology gains an edge through specialized training data, regulatory knowledge, and workflow integration that generalist platforms cannot replicate. Luck emphasized that founders must own their data pipeline and build network effects within their industry.
The market opportunity is substantial. Enterprises across healthcare, finance, manufacturing, and law spend billions solving repetitive, knowledge-intensive tasks. Vertical AI startups capture this demand by embedding themselves into workflows where switching costs run high and accuracy directly impacts revenue.
Luck's thesis demands founders think long-term about moats. Quick wins in fine-tuning fade when platforms add features. Real defensibility comes from accumulated customer data, industry relationships, and regulatory certifications that competitors cannot easily copy.
NEA backs vertical AI companies betting that specialization beats generalization in enterprise software, especially where domain knowledge determines product quality.