Upgrade or Lose - The Nvidia Strategy

NVidia came up at a time that Intel was a full sized juggernaut. Investments needed to even enter the market were humungous. Desktop 3D graphics was almost not a thing. There was no single standard to build graphics computation products. So what was the kernel of the strategy that has propelled Nvidia to twice the size of Intel by 2020?

The Diagnosis

Running out of cash, up against intel, Nvidia saw a few things differently. 1990's was an era where Personal Computers, Desktop Gaming and Internet were all going through their growth spurts. Nvidia saw an opportunity at the intersection of these 3. It carefully crafted its mission as providing competitive advantage in multiplayer gaming. If you didn't have the latest Nvidia graphics card, there was a good chance that you were dead before the first bullet left your gun.

The other part of the story was about technological advances. Intel had defined and worked with Moore's law and so did the entire industry. Which meant that new products came out every 18 months. That was acceptable for CPU power as no one wanted a significantly faster word processor. But faster, more detailed, more interactive games just gulped up the additional megaflops. The demand for graphics was virtually limitless. Nvidia predicted that the GPU demand would rise 3x faster than the CPU demand.

Guiding Policy

Propelled by these insights, Nvidia decided to break the industry cycle and work at 3x speed. This meant releasing a product every 6 months.

Realizing its cash position and need for speed, Nvidia decided to outsource all manufacturing and be a fabless semiconductor company.

Coherent Actions

Releasing a product every 6 months meant some fundamental organizational changes. Nvidia created 3 separate teams that worked on an 18 month cycle each at a phase difference of 6 months. Being fabless meant increased dependence on external entities for prototyping and feedback. To beat that, it invested heavily into simulation and emulation tools. It decided to bet on Microsoft DirectX as the standard of the future.

Powered by this strategy, Nvidia drove 157% YOY gain in performance from 1997-2001 period and 62% YOY performance gain for the 5 years after that. These gains compounded with the fast release cycles led to Nvidia obliterating competition when it came to GPUs and Gaming.

Inspired by

Good Strategy / Bad Strategy - Richard Rumlet