Has the semiconductor market fundamentally changed?
One of the topics of discussion over the New Year was of course the chip shortage and the huge amounts of cash being committed to new semiconductor capacity. The largest chip makers have committed over $450bn to new capacity, just under the value of the entire industry in 2020 at $464bn.
Will this lead to over-capacity and a collapse in the semiconductor market as we have seen in previous industry cycles and predicted by IDC?
Not so, say the optimists, there is a new era. Some forecasts have the semiconductor market reaching $800bn by 2028 and $1 trillion by 2032.
They point to past cycles that have been driven by single killer apps such as the PC In the early 1980’s, mobile phone, internet infrastructure, tablets and smartphones. When these markets became saturated the supply vs demand changed dramatically and the market slumped, leading to the dips.
Today’s cycle is different they say as there is no one killer app acting as a catalyst but instead we have multiple market growth events occurring simultaneously, many interdependent upon each other and all contributing to a rapid increase in TAM growth, they say.
These are unlikely to dip at the same time, smoothing out the growth in the semiconductor market.
“I really do not believe that we will predict the future semiconductor TAM by studying the cycles of the past, we are in a unique market, not just because of the pandemic but for the ubiquitous spread of semiconductors into all aspects of life and end products and I can only visualise this increasing rather than decreasing in the future,” said one highly knowledgeable commentator.
- Record high for semiconductor content in electronic equipment
- Chip market to rise 10.8 percent in 2022
- TSMC’s $100bn fab plans to expand capacity
The pandemic has drastically accelerated the medical and telecoms markets through people working from home but hit the automotive sector in the short term not because of the semiconductors themselves but for management decisions made in the automotive sector when the pandemic struck.
The optimists do acknowledge it may not be plain sailing, and the industry may experience some blips in the growth rate and curve during the next decade, but the underlying trend for the semiconductor industry during the next 10 years will be substantial growth of 5 percent a year
“I do not think we can continue to use the historic trends of our industry to predict the future, we are entering an entirely new era and I think it would be wise for us all to open our minds to all kinds of new and exciting opportunities and possibilities,” says one distributor.
这也许是由增加percen备份tage of electronics in electronic equipment, rising to a record 33 per cent in 2021, and a record $680bn market in 2021.
However we have heard this before. The rise in multimedia and digital TVs led market researchers to similarly claim a new game-changing era, just before the semiconductor market collapse in 2000.
Next: A rocky road for the semiconductor market
It is those ‘dips’ and ‘blips’ that matter. No one is arguing that the semiconductor market will collapse as it is a fundamental part of the global economy, the ‘new oil’.
But the balance of capacity and demands has never been accurately aligned, and the chances of that happening are vanishingly small. Much of the investment is in large 300mm fabs at 5nm and 3nm process technologies costing $20bn a time, and these make capacity a step function, not a steady state. These fabs need to be kept full and in production, and increasing inefficiencies with fabs around the globe may well lead to duplication and significant over-capacity.
也有政治uncertainties, with various CHIPS Acts looking to provide subsidies for new ‘sovereign’ capacity in the US and Europe, potentially distorting the market and creating inefficiencies.
- Fragmented chip manufacturing brings big risks, says TSMC
- Opinion: Politicising semiconductor manufacturing is a global failure
- Future of TSMC becomes pawn in China-US tensions
The extra capacity will drive a fall in average selling prices (ASPs), agrees industry veteran Malcolm Penn, founder of Future Horizons which accurately predicted the current ‘supercycle’.Even though we will eventually need the capacity that is being committed today for that industry growth, “it’s a case of too much too late triggering a near-term oversupply,” he said. He also points out that the demand and capacity will be imbalanced, with some demand happening in areas where there is not the capacity investment.
“We were the only analyst to correctly predict 2021’s super-cycle and 25 per cent growth, plus the reasons why, with our IFS2021 forecast of +18 per cent with an upside of +24 per cent,” he said. “All other analysts at the time were half our forecast and we were dismissed as ‘ever-optimistic’ by SemiWiki in their industry outlook report. As it transpired, our forecast proved absolutely spot on, based on our 55 plus years of direct industry experience, longer than any other analyst and most industry execs. We are never afraid take a contrarian view, backed up by data and sound analytical process, which is why our industry forecasts have consistently proved both accurate and insightful, second to no-one.”
Banking on a new era of steady growth for the semiconductor industry is not optimism, but failing to learn from the past. A trillion dollar market is certainly possible with average growth over the next decade. The variations in the market will make that a rough ride.
Future Horizons is hosting its2022 forecast seminaron January 18th which will give more detail on the market predictions.
www.futurehorizons.com;www.icinsights.com
Other related articles
- Record sales for TSMC and UMC in 2021
- Galway in line for Intel megafab
- Samsung choses Austin for $17bn 3nm fab
- The fastest route out of the chip shortage
- Telecom chip shortage to continue through 2023
Other articles on eeNews Europe
