NVIDIA
NVIDIA
USA, Founded 1993
Society
10/10
Customer
10/10
Self
8/10
14 above the excellence line
Society
10/10
Customer
10/10
Self
8/10
14 above the excellence line
Gaming GPUs became the heart of AI. 30 years of accumulated technology blossomed overnight.
12 criteria
52/60
3 axes
28/30
Bonus
+9
NCG
89/100
Bonus: [Impact+Capital+Strategy] 15yr vision taking GPUs→the heart of AI+top market cap; 33yr tenure is mid
Context
Jensen Huang founded the company as a graphics-chip maker for games. In 2006, opened the GPU to general-purpose computing with CUDA. The deep-learning revolution made the GPU indispensable infrastructure for AI training and inference. Now the world's most valuable company, with a market cap over $3T.
Teles' Insight
NVIDIA's bet was CUDA in 2006 — in an era when nobody outside of games needed it, they opened up the GPU to general-purpose computing. Jensen Huang is a rare leader who has been CEO for 30 years, and when the AI boom arrived, 20 years of software compatibility and a 4M-developer ecosystem were already ready. They effectively monopolize 3nm data-center GPUs, with gross margins over 70% and net margins over 55% — outliers by semiconductor-industry standards. The foresight 'a GPU is a general-purpose parallel processor' was rewarded at a $3T market cap.
12 Criteria (52/60)
StrengthsA Market, B Problem, C Solution, D Team, E Traction, F Moat, G Revenue, H Financial, I Efficiency, J Exit
AI semiconductors. Over 80% share of data-center GPUs.
AI training and inference is massively compute-starved.
GPU + CUDA + software stack. Dominates both hardware and software.
Jensen Huang. A rare founder-CEO who's held the role for 30 years.
Data center revenue $190B+/year (FY2026). Every AI company runs on NVIDIA GPUs.
The CUDA ecosystem. 20 years of software compatibility. 4M developers.
Data center + gaming + automotive + robotics.
Gross margin over 70%. Net margin over 55%. Abnormal for semiconductors.
$215B+ revenue from about 42,000 employees. The efficiency of the fabless model.
Market cap over $3T. Surpasses Apple and Microsoft.
The decision to build CUDA in 2006 was everything. The foundation was laid before the AI boom.
'The GPU becomes the general-purpose processor of parallel computing.' Saw uses beyond games.
