NVIDIA CEO Jensen Huang: NVIDIA Initially Bet on the Wrong Graphics Technology, He Bought Three Technical Textbooks for Engineers
NVIDIA CEO Jensen Huang stated at the Y Combinator Startup School 2026 that now is the best time in history to start a company, and young entrepreneurs should not overthink due to rapid technological changes or fear of failure.
He suggested that when facing challenges, one should first ask, "How hard can it be?" Entrepreneurs do not need to have all the answers from day one; the key is to face reality and keep learning. NVIDIA initially bet on the wrong graphics technology, and he bought three technical textbooks for engineers, emphasizing that as long as one maintains the ability to learn, the specific technology itself is not the most important factor.
Huang mentioned that he still works seven days a week and has long been worried about the company potentially going bankrupt; the sense of fragility and insecurity has never disappeared. He advised against imagining all difficulties in advance to avoid anxiety that could hinder action, and instead suggested letting difficulties come gradually and solving them one by one. Event-driven entrepreneurs accelerate action, with resources flowing to AI-related startups, benefiting young founders with learning abilities, while potential entrepreneurs who overanalyze and hesitate remain under pressure.
Source: Public Information
ABAB AI Insight
Since founding NVIDIA in 1993, Jensen Huang has publicly reviewed his early near-bankruptcy experiences multiple times, including the pivot after betting on the wrong graphics architecture by purchasing textbooks and communicating honestly with customers. This path has become a core narrative he emphasizes: "confronting reality and learning quickly."
This speech directs resources and attention towards the AI-driven entrepreneurial window, with capital encouraging founders to leverage current models and computing power for rapid trial and error, motivated by the desire to expand NVIDIA's downstream application layer and solidify its position as an infrastructure provider.
Similar statements about the "best time" have appeared in speeches by tech leaders before the internet bubble and during the mobile internet explosion. Currently, AI entrepreneurship is in an accelerated expansion phase and has not yet entered a period of large-scale elimination and consolidation.
Essentially, this represents a technological replacement: the computing paradigm is shifting from traditional software to AI-native, where the maturity of infrastructure lowers the barriers to entrepreneurship, making learning ability and speed of action the new core competencies.
ABAB News · Cognitive Laws
- The best time is always after action.
- Learning ability is greater than initial correctness.
- Difficulties should come in batches to be manageable.