Editor's Choice

Nvidia's Fatal Flaw

Steven is a member of The Motley Fool Blog Network -- entries represent the personal opinion of the blogger and are not formally edited.

While conducting research on my holdings, I recently uncovered a deep-rooted, fundamental flaw to Nvidia’s (NASDAQ: NVDA) business model, which led me to sell my entire position.  This development could perhaps explain the recent short-term underperformance and lower valuation bullish writers have been touting.

It is my belief that Nvidia shares are likely to remain range bound for the foreseeable future, and perhaps as far as 2017.  This major headwind brings into question the sustainability of Nvidia’s earnings growth over the long term; I am not even sure if Nvidia will be able to deliver stable earnings results over a similar period. 

Business Basics

Nvidia is known as a “fabless” chip designer, meaning they do not own the manufacturing fabrication plants (aka “fabs”) for the chips they market.  This business model allows Nvidia to focus more on the design (R&D) of their products and less time worrying about production.  Currently, Nvidia utilizes Taiwan Semiconductors (NYSE: TSM) to produce their designs.  This is common practice within the semiconductor space where a designer sends out their blueprints to a semiconductor foundry like Taiwan Semiconductor.  For example, Qualcomm (NASDAQ: QCOM), AMD (NYSE: AMD), and Broadcom (NASDAQ: BRCM) all are “fabless” semiconductor companies that employ similar approaches to chip development.   

While it’s a positive that Nvidia does not have to master manufacturing processes in order to compete, they end up relying heavily on a third party to produce their designs in a cost effective manner.  This point is especially important to remember when designers like Nvidia are first in line for next generation processes (e.g. 40nm -> 28nm) from the foundry.  It is all but certain that production issues will occur when transitioning a foundry to next generation manufacturing processes alongside new, untested chip designs. 

Given that Nvidia is a design company by nature and lacks the manufacturing expertise, production issues are likely to present themselves during initial runs.  This almost always ends up resulting in added costs, delays, and compromises for Nvidia.  All fabless manufacturers deal with the same obstacle, so this isn’t necessarily the fatal flaw in their business relative to the competition.         

The issue here isn’t about the increased cost when manufacturing problems assert themselves – it’s related to the ever increasing cost of development with each generation, especially being a first adopter for a new foundry process.

Moore’s Second Law

While Moore’s first law is well understood, Moore’s second law is less well known.  It states that as chip densities double, so does the capital required for producing them.  Common logic would put fabless designers at a competitive advantage, right?  Wrong.         

Who do you think bears the greatest impact of production issues at the foundry, the customer or the foundry?    

Fatal Flaw   

Last month, the details of Nvidia’s presentation given at the International Trade Partner Conference back in November were made public and expressed their frustration with Taiwan Semiconductor.  The following slides were presented:

<img src="/media/images/user_12856/nvda-transistor-cost_large.jpg" />

<img src="/media/images/user_12856/nvda-wafer-cost_large.jpg" />

The first slide illustrates how the current model is unsustainable by showing the company’s transistor cost at current and future nodes.  This affirms Moore’s second law by demonstrating that as process nodes shrink, it takes more time for the cost per transistor to fall below the previous generation.  In fact, Nvidia does not expect the cost per transistor to fall more than marginally below previous generations at 28nm, essentially losing out on potential profits.  In other words, when Nvidia invests in future generations, the return on investment diminishes significantly.  That’s the first sign of trouble. 

The second slide speaks to the pricing of wafers as densities increase in future generations. Simple logic would conclude, the smaller the process, the more chips you can fit on each wafer, making it more cost effective for each future generation.  This logic overlooks the reality that the smaller the design, the more complex and difficult it becomes to manufacture, which in turn directly increases future input costs for Nvidia. 

Fabless Not So Fab

Historically, Nvidia has been an aggressive first adopter when it comes to transitioning to next generation manufacturing methods.  When you couple the diminishing returns on investment between future generations with the fact that each generation increases in cost, complexity, and time, you don’t find Nvidia being able to sustain its earnings power.  In short, there is little to no financial incentive for Nvidia to transition to smaller, more complex, and more dense components; the fierce competition within their sub industry is essentially forcing them to, otherwise they risk losing major business.

However, all is not lost.  To combat this major headwind, Nvidia has redesigned their Fermi product blueprints with a renewed emphasis on energy efficiency and performance per watt.  While admirable, I am not willing to wait around for the results of their attempt to steer the ship away from the land of increasing input costs and vanishing returns on investment. 


TopDownTrends has no positions in the stocks mentioned above. The Motley Fool owns shares of Qualcomm. Motley Fool newsletter services recommend NVIDIA. Try any of our Foolish newsletter services free for 30 days. We Fools may not all hold the same opinions, but we all believe that considering a diverse range of insights makes us better investors. The Motley Fool has a disclosure policy.

blog comments powered by Disqus