The Disconnect Between Semiconductor IP and Production
We are currently in the midst of a global semiconductor shortage.[1] While this shortage certainly isn’t the most serious global crisis at the moment, it does have considerable ramifications on everything from international trade to whether you’ll be able to get that new phone you’ve been saving for.[2]
Backing up for a minute: what is a semiconductor? In short, a computer chip.[3] They’re the brain of essentially any device or product that one might consider ‘smart,’ meaning they’re at the heart of most consumer electronics like phones and computers.[4] These chips are also increasingly necessary for vehicles, which have grown more complex over the years.[5]
The semiconductor industry is incredibly complex but can generally be divided into two roles for industry players to fulfill: design and manufacture.[6] Many companies focus on one or the other, while some do both (although this is increasingly rare).[7] Design companies spend most of their time and money on R&D, trying to create advanced chip schematics that are incrementally smaller and smaller.[8] Some design companies like ARM focus entirely on licensing those design schematics out to other companies like Apple, who adapt and build on them for their devices.[9] Those licensor companies, as well as other chip designers who don’t license their designs, will then outsource to a manufacturer which churns chips out in huge foundries known colloquially as ‘fabs.’[10] Currently, Samsung, Intel, and Taiwan Semiconductor Manufacturing Company (TSMC) are the only companies capable of producing the industry’s most advanced chips due to the high cost of the process and technical knowledge required.[11]
So why is there such a dramatic shortage of these chips? A main cause of the current shortage is easy enough to identify; it’s the same cause behind most of the world’s woes for the past year and change: COVID-19.[12] However, there are multiple and multifaceted ways in which COVID-19 has impacted the chip manufacturing industry. Uncertainty will negatively affect any industry, but it is felt acutely in the chip industry because of how involved the manufacturing process is for semiconductors.[13] Manufacturing a computer chip is one of the most difficult processes in all of manufacturing because it involves an intricate process working with incredibly small components.[14] From start to finish, it can take a manufacturer around 26 weeks to complete a chip for a customer.[15] So, when the pandemic first began to grip the world in late 2019 to early 2020 and governments began to shut things down, uncertainty about the future necessarily made it difficult for some industries to justify placing orders for goods that wouldn’t be ready for another half-year.[16] This was especially true of the automotive industry, which anticipated that demand for cars (and their capacity to manufacture cars) would take a sharp downturn for the foreseeable future.[17] Chip manufacturers responded by switching their focus to manufacturing for the much more lucrative consumer electronics market, which thrived during the pandemic due to the increased need for effective personal computing and home diversions like video game consoles during lockdowns.[18] Either fortunately or unfortunately, the dire projections for the automotive industry turned out to be somewhat incorrect.[19] The industry began ramping up vehicle manufacturing again but found that most of them lacked the necessary computer chips to do so as aggressively as they might have liked.[20] At that point, they had to start competing with consumer-electronics companies for limited quantities of computer chips that manufacturers were capable of outputting during the pandemic.[21]
Putting all of that aside, it’s worth considering whether the current complex web of chip manufacturers and designers brought us to a point where a chip shortage like this could happen. Concentrating all manufacturing power in a few key players is an excellent way to ensure that when things go wrong, they go very wrong. So, is there any role for the holders of semiconductor design IP to play in ensuring large-scale shortages like this don’t happen down the road? Perhaps, but they certainly don’t seem interested in doing so. For example: Intel is currently one of the few companies left that still both designs and manufactures chips.[22] This allowed it to have a leg-up over its main competition for consumer computer processors, AMD, during the ongoing shortage.[23] Though AMD’s processors have been better received recently, Intel’s have been on the shelves.[24] This enabled Intel to get back some market share it had lost to AMD.[25] However, even Intel’s longstanding practice of integrating both parts of the business may be changing.[26] Intel is lagging in design and is seemingly considering outsourcing its manufacturing to TSMC – like everyone else – so it can spend more on R&D and catch up with the competition.[27] As long as design companies focus more of their time and energy on strengthening their IP portfolios – and less time on ensuring that IP translates into tangible goods – the more problems with production there will be down the line.
Footnotes
