Semiconductor shortage likely to persist for a long while

A supply response to the auto chip crunch will take its own time while demand might surge sooner, writes Neeraj Mohan, leader of EY-Parthenon in India.

Enough ink has been spilt on today’s semiconductor shortage, some of which has touched upon its severe impact on the automotive sector. The sector’s response has been a mixed bag: some parts of it have taken action to improve supplies while others have taken comfort in the view that the shortage will ease in 2022. The chip shortage, however, is likely to last longer.

The perception that the semiconductor shortage sprang out of nowhere is partially correct. The crunch is traceable to how global semiconductor manufacturing has evolved and the automotive industry’s sourcing practices.

The semiconductor industry underwent a major change in the mid-80s as chips started getting commoditized. To protect against price risks and secure heavy investments that go into fabs, chip makers started blocking capacities in return for purchase commitments at pre-agreed prices. To cover for demand variability, excess capacity was maintained.

The result has been that in a ‘normal’ year, around 80% of all production would go to pre-committed buyers, while the remainder would be offloaded to the industry’s equivalent of ‘traders’.

The auto industry’s reliance on cheap semiconductors has evolved from this pool of low-cost supplies. Freely available chips weakened the need to lock in long-term consignments. As its need for processors rose, the auto industry became a big buyer that now accounts for around 12% of global demand.

The current shortage has resulted from a convergence of multiple factors: Covid-19 led to a surge in demand for digital devices, supply-chain disruptions ran down inventory buffers, and crypto emerged as an important source of demand, as power-hungry computing applications like artificial intelligence and machine learning. As demand surged and residual supplies became intermittent, there was no respite.

The auto industry’s initial response was to treat the situation as a short-term problem. By mid-2021, the majority view was that the shortage would ease by early 2022. This median view (it wasn’t a consensus) has gradually shifted to late 2022. But a close analysis suggests that a different scenario is unfolding.

At EY-Parthenon, we have analyzed all announcements on additional capital expenditure and new capacities by fabricators. Given the lead time involved in getting lithography equipment from suppliers like ASML and in commissioning capacity, a supply assessment through 2025 can be made.

As new capacities take heavy capital expenditure, fabricators tend to add capacity based on commitments. There are equipment constraints involved as well. While there is a possibility that more capacity can be planned than what we have listed, the chances of it coming on stream in the next 1-2 years are thin at best. The world will likely have to manage within the constraints of capacity going up by about 14% through 2025.

In the near-term, the industry will scramble to cope by reverting to older systems, reducing features in vehicles, and delaying deliveries. In India, for instance, the delivery backlog currently stands at around 700,000 passenger vehicles and had hit a peak of almost a million during the festive season.

Rationalizing the number of chips per vehicle will be important. While new automakers tend to follow a ‘System on Chip’ (SoC) approach to contain their count of wafers, some luxury models use up to 100 chips per car. With ongoing work to combine multiple SoCs, the industry should see a progressive easing of the crunch.

A longer-term view

A longer-term response has also started to emerge. Tesla has a long-term deal for next-generation autonomous vehicle chips with Samsung , which has announced a new chip plant in Texas. More dramatically, both Ford and GM have said they would get into long-term design and manufacturing arrangements. Importantly, national governments have also started treating chip supplies as a strategic imperative.

It is heartening that India’s government has come out with a comprehensive strategic plan for securing long-term supplies. The government’s newly announced $10 billion incentive scheme for high-end chip-making should evoke interest among global majors and can go a long way in establishing a strong electronics manufacturing base in India.

This plan might not help alleviate the near-term shortage, but will help Indian manufacturers secure supplies and compete globally in the long-term.

Looking ahead, all the announced measures have generated the hope that the shortage will ease by late 2022. Our view is that the situation will ease only for manufacturers who have taken concrete steps to secure supplies.

While capacity is indeed being added, the more important factor is demand. If semiconductor demand slows down, the situation can ease. However, a demand surge in segments that have secured supplies would leave little residual output for those that have not. This will create risks for component suppliers, as their ability to secure long-term supplies is limited. Particularly in the case of electric vehicles, suppliers need to choose a controller and build embedded software for major systems.

In a scenario where demand increases from all quarters – consumer electronics, industrial and automotive – the risk for automakers in general and auto component companies in particular is high.