semiconductor

Should Intel breakup and go “fabless”?

Question: Should Intel go “fabless” and spin-off its semiconductor fabrication? In other words, should Intel break up into two companies, one focusing on design and the other on manufacturing?

Quick Answer: From financial and operational perspectives, such a move by Intel can deliver cost savings, flexibility and competitive advantage.

Fabless Trend: There are many examples of successful enterprises who have solely focused on designing semiconductor chips or have divested the fabrication component of the business. Apple and Qualcomm are two good examples for design-only companies. AMD, Intel’s main rival, spun off its chip manufacturing as Global Foundries (“GloFo”) in 20091 and has reaped many benefits. Samsung also engaged in a similar move last year.

Why go fabless: While designing and manufacturing microchips offer great benefits in end-to-end vertical integration, such a business model —Integrated Design Manufacturer or IDM — suffers from two drawbacks:

  1. Cost. Fabrication plants – also known as foundries – are very expensive in terms of CAPEX and OPEX. For example, Intel’s new plant, Fab 42, in Arizona will cost more than $5 billion. 2
  2. Single Point of Failure. While fabless companies have two or more choices to manufacture their chips, Intel’s designers are stuck with their in-house manufacturing. Thus, any missteps or unforeseen problems in Intel’s foundry will have profound impact on timely product releases, which is very critical in a competitive market. On the other hand, a company like AMD enjoys the flexibility of contracting with Samsung or pure-play foundries such as GloFo and TSMC.

Leadership Loss: Intel enjoys unprecedented brand recognition, thanks to its pioneer work and unrivaled leadership in the microprocessors market for many decades. Intel manufactured the world’s first microprocessor in 1971 and easily captured 90% of the market share for CPU’s in personal computers and servers for a long time. However, Intel missed out on mobile computing, which came to be dominated by the likes of Samsung and Qualcomm. Intel’s Atom processor was late to the game and slowly faded away. Furthermore, Intel lost out in the high-margin and high-growth markets for memory and GPU to challengers such as Micron and NVidia.

Even in the desktop/server market, Intel has been recently losing the market share to AMD. Intel’s processors have been stuck at 14nm technology for the last five years. Its new processors based on 10nm won’t be released until the end of 2019, while AMD will be releasing its 7nm products at the end of 2018.3 This is why AMD’s stock price has tripled in the last few months while Intel’s share has languished.

Other signs of trouble for Intel:

  • Samsung overtook Intel as the largest semiconductor company in 2017 in terms of revenue. 4
  • TSMC has not only exceeded Intel in market cap, but has also beaten Intel in chip manufacturing technology. 5

Conclusion: Selling off its foundry will inject Intel with billions of dollars, which it can use for R&D to design processors for profitable and emerging markets such as IoT, AI and 5G. Being a smaller and more focused company will help Intel be more agile. Intel’s shares will also likely get a massive boost from such a divestment. Finally, the newly spun-off subsidiary can also increase its revenue and skill sets by working with multiple vendors.

Author – Chris Kanthan