How The Semiconductor Shortage Happened – And When It Will Recover

The chip shortage has constrained manufacturing and caused product delays. Here’s what to expect going forward.

This week Huawei sponsored a panel of experts, including one of their own, who explained what caused the semiconductor shortage of the last year – and why it will probably take until well into 2022 before it goes away. Already it is having an adverse impact on computers, automobiles, and, most recently, appliances, and, I expect, we’ll see more areas adversely impacted like aerospace, defense, and networking before this event is over.

The people on the panel were moderator Karen Brown, Principal Analyst at KL Brown Consulting; Mario Morales, VP of Enabling Technologies and Semiconductor Storage Research at IDC; Earl Lum, President of EJL Research; and Tim Danks, VP of Risk Management and Partner Relations for Huawei Technologies.

Chip Shortfall Causes

Morales did an impressive job laying out the complex causes of the semiconductor shortfall and why the segment was caught flat-footed. Planners weren’t prepared for the pandemic, initially overreacted, and then used past models, which proved inaccurate to forecast the market. As a result, orders were cut far more than sales fell, and manufacturing lines, which in fabs have to be kept at high capacity to cover fab overhead, shifted to produce for segments that were still consuming parts or shut down.

Since sales didn’t decline as sharply as planned, some areas like PCs saw significant growth from the shift to at-home work, parts reserves were consumed, and with sharply reduced manufacturing, these reserves couldn’t be refreshed. Stimulus payments were not anticipated, resulting in far higher sales than anticipated and, thus, a significant shortfall.

Those lines are not being converted back to product semiconductors for various markets. Mario estimated that by the 3rd quarter this year, they should be back up and running at capacity.  However, it takes time for parts to make it through the supply chain to finished products; thus, his expectation that we’ll continue to have finished good shortages due to this problem well into 2022.

Be aware that while IDC is one of the best places to get current numbers, their forecasting methodology is limited by the information they get from vendors. Often the vendors don’t share, or even know, critical information that would make a forecast accurate.  For instance, a 4th or 5th Pandemic wave, as yet unknown demand generation programs, future stimulus payments, and secret disruptive offerings can significantly impact sales but are generally not taken into account.

You might be able to close this gap with large-scale computer simulations, but the industry doesn’t yet fund those. For instance, Mario argued that autonomous cars wouldn’t reach critical mass until 2035, consistent with current demand, especially given the Tesla crashes. Still, none of the autonomous car demand generation programs have even been written, let alone deployed, or have the insurance companies been public with their plans to reduce insurance premiums massively while increasing premiums for human-driven cars, which could move this demand curve in sharply.

But his projection about 2022 being when this current shortage should mitigate should be more accurate since it is based on known plans already being executed.

Huawei in Good Shape

Tim Danks showcased why Huawei appears to be in far better shape than most. They maintain a much tighter relationship with their suppliers than most. They have a considerable number of those suppliers, many of which are redundant for supply chain resiliency, and they appear to do deeper risk analysis than most of their peers. Part of that analysis is looking at potential natural or human-made risks; economic instability; technology, cybersecurity, and privacy challenges; legal and regulatory impacts (of which Huawei has more than most); and geopolitical turbulence.

Even with that redundancy, they have simplified their supply chain, making it easier to manage, built-in fixability to their product lines, and they have a diverse portfolio of products so they can shift when necessary.

Planning for the Future

Mistakes made going into the pandemic have resulted in a semiconductor shortage cutting across several markets, from computers to automobiles and, most recently, appliances. These shortages will continue to worsen as back-to-school and holiday buying seasons hit, mainly before these mistakes can be undone. Shortages generally lead to a sharp uptick in prices as demand significantly exceeds supply, suggesting that you can either buy now (before prices have a chance to peak) or, better yet, wait until mid-2022, when you should get far better prices when supply should again exceed demand.

I would also expect some vendors to move in plans to shift to other architectures to take advantage of parts that aren’t constrained and to strengthen their supply lines to better weather future storms like this pandemic. Finally, if you aren’t yet doing advanced market modeling and simulation, you should consider bringing those projects forward because the anticipated disruptions (IoT, AI, robotics) due this decade may cause more dramatic changes in products and supplies than the pandemic did.