Chipageddon: The Chip Shortage in ~1500 Words
Unless you’ve been living under a rock, you must have heard about the chip shortage. By now there’s even a Wikipedia page for the 2020–2021 global chip shortage — so you know it’s a real thing.
But why is this happening, who should care, how does it affect semiconductors founders, and are there any signs of relief?
Well, let’s start from the obvious: if someone wants something and there isn’t enough of it, that means that someone (else) miss-calculated supply and demand balance. And indeed, the shortage stems from some severe miss-calculation of demand on the world’s chipmakers’ side — all are now overwhelmed by a sharp increase in demand.
Let’s see: we had a pandemic, causing supply chain problems all across the globe. If that wasn’t enough, the blocking of the Suez Canal, one of the world’s busiest trade routes connecting Europe and Asia caused a jam of hundreds of cargo ships in March, and the February Texas storm that forced Samsung, NXP Semiconductors, and Infineon to shut down factories temporarily. The Taiwan drought (fabs are huge water consumers) didn’t help and a fire last March in the Renesas chip factory in Japan, manufacturing 2/3 of the automotive industry chips, significantly impacted the entire supply chain. Add to that the fact that there is a tight geographic concentration of supply chain and of course the trade war between the US and China isn’t helping — it seems the god of Semiconductors are angry.
COVID-19: If there’s one good thing about COVID, it’s that you can blame it for just about anything. The remote-everything trend fueled by lockdowns increased sales of computers and other electronic devices on one hand (not only does every employee now require a laptop, but it also must be a stronger model to support virtual- everything), while the automotive industry witnessed a dip in demand that wasn’t expected to change in the short term: No one was going anywhere. As a low margin industry, the automotive industry sticks to a “just-in-time” manufacturing policy as opposed to “just in case” philosophy: our ability to track down to the minute the global movement of ships and trucks enables manufacturers to ensure that just the part they need arrive exactly when needed, reduce redundant costs, and increase efficiency. But it also means there’s very little room for error. Or for a global pandemic. So, chipmakers switched over their production lines from the automotive lines, which meant that in Q3 2020, when demand for consumer electronics continued to grow but sales of cars came soaring back unexpectedly, the demand could not be met.
The semiconductor industry is a highly specialized global supply chain. Each region performs its part based on its comparative advantage, with over 50 points across the supply chain where one region holds over 65% of the global market share. While the US, with its world-class education system and an abundance of engineering talent, leads in core IP and chip design, east Asia took the lead in wafer fabrication, which requires massive capital investment supported by government incentives. Based on a BCG report from April 2021, 100% (!!) of the world’s most advanced semiconductor manufacturing capacity (below 10nm) is currently located in South Korea (8%) and Taiwan (92%). China, on its end, ruled in assembly, packaging, and testing.
This model did not only work well, it also saved a lot of money. But it clearly suffers from its vulnerabilities as it relies on free trade compromised by geopolitics tension, mother nature, and your current global virus.
When the US government restricted technology exports to China, some Chinese manufacturers have been stockpiling, preparing for further restrictions. Huawei, for example, used to be one of the industry’s biggest chip buyers. In anticipation of being put on a US trade blacklist (which indeed it was, followed recently by Xiaomi), it began stockpiling chips from 2019, which led to tight capacity at Huawei’s main chip supplier, TSMC. Trump’s restrictions on SMIC also meant that companies had to shift to other manufacturers increasing the bottleneck.
In short, everyone. The world runs on microchips, and not just the obvious ones in our phones or computers (which by itself means that half the world population should care — a recent study by strategy analytics claims that 4 billion people own a smartphone). Our shopping experience, our health system experience, our driving experience, and our learning experience — every aspect of our day-to-day is becoming more digital. And when it becomes digital, it runs on semiconductors.
The first to take a hit was the automotive industry — new cars typically include dozens of microprocessors. Ford, BMW, Volkswagen, have all reported bottlenecks in manufacturing, some carmakers are leaving out high-end features (Nissan, for example, is reportedly leaving navigation systems out of some of its cars. If you’re anything like me, that hurts). But recently, electronics manufacturers in Asia have also warned that the chip shortage was beginning to spread to consumer electronics and home appliances.
How Will I be Affected as a Semiconductors Startup Founder?
As a tech-savvy fund, Grove Ventures has a pretty wide semiconductor portfolio. The picture as it is painted by our founders is consistent: It’s not only that fabs’ willingness/ability to invest in R&D (as opposed to production) is at an all-time low, but also that price and timeline of just about everything are impacted. If you’re selling a development kit, your capacitors, resistors, boards, FPGA’s, all of these parts are now more expensive if you’re lucky, or out of stock (with a year-and-a-half lead time) if you’re not. That means your BOM changes and you’re back at the drawing board for some significant redesign. A large corporate executive I spoke with last week said that 40% (!!) of his R&D is busy with redesign. What can you do? Overstock (if you think you’ll need 20 kits, plan for at least double. This way, even if your supplier reallocates and cuts your supply by half, you’re still good. You’re also prepared for positive traction), be as creative as possible with your sourcing, and be mentally ready for some major changes in design.
Who’s Enjoying The Situation?
In face of demand — we pay. So not surprisingly, the revenues of the 10 semiconductor foundries broke records in Q1 reaching $22.75B with TSMC holding a phenomenal 55% market share. And of course, if you are a semiconductor investor, you’re also in luck. The SOX index (composed of companies primarily involved in the design, distribution, manufacture, and sale of semiconductors) has increased 14% since the beginning of the year (compared to a 7.7% increase in the technology sector of the S&P 500 index).
When Will This End??
Well, it depends on who you ask. Some manufacturers forecast the shortage will last until at least mid-2022. Flex, for example, the world’s third-biggest electronics contract manufacturer, said that the manufacturers it relies on for semiconductors have pushed back their forecasts for when the shortage will end. Some of the more pessimistic analysts (Forrester, for example) believe the shortage could last until 2023. The most optimistic estimation I could find was by Cisco’s CEO expecting a relief by early 2022. What we know for sure is that the only realistic solution is to increase the supply of semiconductors, and indeed — it’s response time.
Chipmakers are investing more and more in new production capacity: Intel announced in March that it plans to spend $20 billion on two new chip factories in Arizona. TSMC announced in April it plans to invest $100 billion in the next three years in expanding its manufacturing capacity and GlobalFoundries announced in June a $6B expansion in Singapore, the US, and Germany.
Many car companies have adopted a more assertive approach to sourcing, such as stocking and paying for chips in advance, and some are being more creative. Tesla, as the Elon-company that it is, has been reported to explore buying a chip plant outright. Bosch, Europe’s largest auto parts supplier, opened a $1.2B chip factory in Dresden last June, the single largest investment in the company’s history.
Countries are trying to become more independent when it comes to semiconductor production. The US is upping its game. The US Senate just recently passed a bill to support domestic chip production in the US. Just this March, Biden called on Congress to invest $50 billion in semiconductor manufacturing and research to support the CHIPS Act (Creating Helpful Incentives to Produce Semiconductors. Masterpiece of copywriting). Immediately following Biden’s statement, a cross-sector alliance of semiconductor companies and major downstream users of semiconductors (from the immediate suspects such as AMD, Intel, IBM Nvidia to AWS, Apple, Google, and Microsoft) announced the formation of the Semiconductors in America Coalition (SIAC) and called on congressional leaders to allocate $50 billion for domestic chip manufacturing incentives and research initiatives.
The EU’s updated its industrial strategy, now including semiconductor technologies as an area of key vulnerability for Europe’s strategic interests and declaring that the EU should produce more chips locally to help prevent future supply issues. In response, Intel announced in March that it will open a foundry plant for cutting-edge semiconductors in Europe.
There’s no reason for the demand for semiconductors to slow down (cloud computing growth, cryptocurrency mining, 5G, AI, and the list goes on and on), and having 80% of its supply concentrated in a single region should probably have never made sense in the first place (never put all your eggs in one basket and all that). The only realistic solution is geographic diversification and increasing the supply of semiconductors. There is, however, a place for optimism — billions are being spent to shorten the waitlist for my new Playstation5.