For years now, chip giant Intel (INTC +7.06%) has talked about how it hopes to compete in the contract chip manufacturing market dominated by Taiwan Semiconductor Manufacturing Company (TSM +1.32%) and Samsung (NASDAQOTH: SSNLF).
Even as recently as last fall, Intel hosted a Technology and Manufacturing Day in China and spent a great deal of time laying out the opportunity that it saw in becoming a contract chip manufacturer for mobile and network infrastructure applications.
Image source: Intel.
Although the company says that the total addressable market for contract chip manufacturing services was around $53 billion in 2016, with so-called "leading edge" technologies -- the subsegment of the market that Intel hopes to serve -- comprising $23 billion of that, I believe that Intel's efforts here will fail, and it won't capture even 1% of that $23 billion served addressable market.
Here's one simple reason why.
Unreliable technology road map
Companies like TSMC and Samsung deliver new manufacturing technologies like clockwork. They need to do so because their major customers, such as mobile processor makers, have to hit very tight schedules because smartphone cycles stay the same each year.
Apple (AAPL +0.33%), for example, launches new iPhones in September of each year, so any contract chip manufacturer that Apple uses to build its chips needs to have its latest technologies ready for mass production by around April or May of that same year. Any delay and Apple's iPhone launch plans will be derailed, adversely impacting its business prospects.
While TSMC and Samsung consistently execute in bringing new technologies to market each year, Intel has -- over the last five years -- failed to deliver new technologies into mass production on time.
Intel's 14nm technology, originally slated to go into production by the end of 2013, didn't begin production until mid-2014. Even worse, the company's 10nm technology, which was originally scheduled to go into production by the end of 2015, still isn't in production and won't go into production until sometime in 2019 (Intel refuses to publicly commit to when in 2019 it expects to start production).
A four-year delay in a product would be nothing short of disastrous for any chip or systems company. Can you imagine if Apple saw a four-year delay in its latest A-series processor?
Until Intel can show multiple generations of consistent technology execution -- something that'd likely only prove out over the next six to 10 years -- no company run by halfway competent management would bet its businesses on Intel's chip manufacturing capabilities. Such a move would be, at best, reckless, and at worst career suicide for the decision-makers who chose to go with Intel.
Intel's best course of action
At this point, Intel's best course of action would be to completely shut down its contract chip manufacturing efforts. This means the company should no longer invest any resources in trying to acquire new companies as customers (this would likely be an exercise in futility, anyway) and it should no longer allow the needs of potential third-party customers influence its technology development.
Any engineering resources that it had allocated to this effort should be reassigned to work either on Intel's internal chip manufacturing technology development or its internal product development. Any marketing resources should be similarly redirected to support the company's large portfolio of in-house-designed products.
