Raphael Schaad: Frontier Labs Won't Dominate Everything, Huge Opportunities Remain in Vertical Fields
Raphael Schaad stated that despite the rapid iterations of frontier labs like OpenAI and Anthropic, startups can easily avoid their "blast radius" and build excellent businesses.
He cited examples: frontier labs will not develop cattle herding drones (Graze Mate), integrate dental insurance verification systems (Lance), or handle NATO procurement processes (Milliray). The value lies in the "last mile"—understanding the sales cycle and domain expertise of customers.
Market mechanisms are accelerating venture funding towards vertical industries, domain specialization, and outcome-oriented AI applications. Vertical AI companies and full-stack service startups benefit, while general chatbots and thin UI layer products face pressure, with capital shifting towards AI projects that can truly address industry pain points and generate business results.
Source: Public Information
ABAB AI Insight
Raphael Schaad's viewpoint directly addresses the common anxiety among MIT students that "frontier labs will take over everything." He emphasizes that AI is not just competing for the existing 1% of software spending, but is unlocking an additional 5-6% of the market previously handled by humans, indicating significant overall market expansion rather than contraction.
In terms of capital pathways, startups should focus on "full-stack AI companies" (such as AI law firms, AI accounting firms, AI uranium exploration companies) or outcome-oriented products, rather than merely tool layers. Cases like Graze Mate from YC W26 demonstrate that hybrid models combining real-world execution (driving to farms) and AI are rapidly growing.
Similar cases include Notion and Figma, which delve deeply into verticals based on general tools, as well as numerous AI + traditional industry (such as healthcare, legal, manufacturing) full-stack startups; current AI entrepreneurship is transitioning from "chasing model capabilities" to "deepening the last mile and industry outcomes."
Essentially, this represents capital concentration: AI value is shifting from general models to the last mile in vertical fields, with the mechanism being that frontier labs struggle to cover all industry-specific processes and human sales relationships, leading to pricing power concentrating from pure model providers to startups with domain expertise that can deliver real business results, while opening up a much larger space for the next generation of AI entrepreneurship than the existing software market.