Amid a chip supply crunch, 28nm could end up underutilized
To analyse Amid the semiconductor crisis, an interesting cliff is forming at 28nm.
As demand for other process nodes outstrips supply, the tech world’s need for this mid-tier node may fall below available manufacturing capacity, if it hasn’t already.
Early indications of a potential oversupply at 28nm emerged during an earnings call with UMC, a major contract chipmaker headquartered in Taiwan.
“On the supply side, based on the announced capacity expansion plan, we expect the 28nm oversupply situation to occur beyond 2023, not before 2023,” said Jason Wong, chairman of UMC, to analysts on the call.
Manufacturers, such as those owned by UMC, TSMC and Intel, generally cannot produce chips fast enough to meet demand. Microchip makers are adjusting their manufacturing capacity accordingly and bringing new factories online, increasing the possibility of chips flooding the market, with supply eventually matching or exceeding demand.
Analyst firms believe chip shortages will end in 2023, and IDC predicted the emergence of an oversupply scenario.
This oversupply problem may particularly affect the 28nm node, which is used by UMC to manufacture microcontrollers as well as integrated circuits for wireless communications, sensors, and mixed-signal applications. UMC also manufactures chips on a 22nm node.
Financial analysts on a TSMC earnings call last month probed the industry over the 28nm oversupply that has plagued the Taiwanese giant in previous years. There are fears that TSMC and its peers could end up with overcapacity after competing to establish several rival factories.
TSMC built the world’s first 28nm factory in 2011 and last year invested in a new 22/28nm factory in Japan, and is expanding capacity in Nanjing, China. Rival chipmakers are also increasing their 28nm capability; SMIC began building a 28nm plant last year that will go into service in 2022.
During his call with analysts, TSMC CEO CC Wei admitted that 28nm was problematic before the pandemic: his manufacturing lines for this node had a utilization rate of just over 80% in 2018 and 2019. He said he hopes it will happen again. situation can be avoided as chip supplies normalize.
NAND flash controller company Silicon Motion Technology said last week that its supplier TSMC was running out of capacity on many nodes, but not for 28nm. And if Silicon Motion can’t get enough dies on non-28nm, it will rethink how it can use 28nm to get the product out there, he said.
“We are in severe shortage in some technology nodes, including 55nm and all advanced technology nodes, 16nm and 12nm. However, I think we are comfortable with the supply of 28nm wafers,” said Wallace Kou, CEO of Silicon Motion Technology, on an earnings call.
“If we can’t get additional wafer supply from TSMC, it all depends on how we can fully utilize 28nm with a better product line.”
Microchip demand is expected to generate $680.6 billion in revenue for semiconductor companies in 2022, trending to $1 trillion in the future as cars electrify and companies integrate more electronics in the products.
The secular model
Chipmakers, especially memory makers, have faced volatility for decades, and volatility is seen as part of the demand and supply dynamics of semiconductors.
The boom and bust cycle of chipmakers is exacerbated by macroeconomic events, such as economic crises, natural disasters, and over or under investment. Historically, the semiconductor market crashed after events such as the 2007-2008 financial crisis halted computer purchases, drying up demand for chips.
But the COVID-19 pandemic has worked in favor of chipmakers. Demand for chips exploded as people worked and learned from home, leading to PC upgrades and cloud players purchasing equipment to power collaboration and communication at distance. This has come up against increasing semiconductor content in electric vehicles and already high demand for smartphones, laptops, custom silicon and servers.
Chip companies have taken in record revenues and committed billions to expand manufacturing through facility upgrades or new factories. Most investments are in state-of-the-art nodes for compute-intensive components such as graphics accelerators and smartphone system-on-chips, and older nodes for cost-optimized chips, from microcontrollers to circuits. built-in analogs.
The 28nm node may be snubbed or ignored by fab customers in favor of other nodes that offer more performance or are more cost effective, which could create an oversupply scenario, said Akshara Bassi, analyst at Counterpoint Research. “So 28nm could be stuck in this precarious transition cycle, where some products transition to more than 28nm more advanced than expected,” Bassi said.
Some products, such as memory controllers and I/O arrays, might also be in a phase of transitioning from late nodes to “matured nodes that are a bit more advanced for better cost savings and to meet commitments.” environmental, social and governance issues of OEMs,” added Bassi. .
Microcontroller chips and other embedded components are moving toward the 14nm node, STMicroelectronics CEO Jean-Marc Chery said on an earnings call last week. He pointed out that 28nm is at an uncomfortable threshold for foundries investing billions in factories.
For chip designers, going from 28nm to a node like 22nm is a reasonable leap that on the one hand opens up the possibility of getting more dies from fabs and thus increasing cost efficiency, said Brian Matas , vice president of market research at IC Insights, a semiconductor company that crunches the numbers.
Scaling down a design to a 22nm node should get more dies on a wafer and ship more finished components, for example.
“You’re able to lower prices and pass them on to some of your customers, and at the same time you’re still making more money on that smaller process node,” Matas said.
UMC and TSMC have acknowledged that a correction will occur as chip momentum slows, and that the negative impact on 28nm will be a temporary glitch as applications emerge to make the process relevant again.
“We observe that our long-term structural demand at 28nm was to be well supported by several specialized technologies such as CMOS image sensor for multi-camera trend and better application of non-volatile memory and other specialized technologies” , said TSMC’s Wei. “Enriching the silicon content of many end devices that have developed in recent years has helped meet this demand.”
UMC’s Wong said the oversupply on 28nm in 2023 will be light on his business, and ultimately the node will emerge as a sweet spot for many applications in the future. “So expect demand to continue to migrate to 28nm and demand for 28nm to continue to grow,” he explained. ®