There are more companies in foundry-space (in general) than I tend to remember…
TrendForce reports that the revenue rankings of the top ten wafer foundries remain unchanged in the third quarter, with TSMC maintaining its leading position with a market share of nearly 65%. The simultaneous launch of flagship smartphone products, AI GPUs, and new PC CPUs drove an increase in TSMC’s capacity utilization and wafer shipments, resulting in a 13% QoQ revenue growth to $23.53 billion.
Samsung Foundry maintained its position as the second-largest foundry by revenue in the third quarter. Despite securing some smartphone-related orders, the company’s advanced process clients’ products are approaching the end of their life cycles. Additionally, intensified competition from Chinese peers in the mature processes led to price concessions, resulting in a 12.4% QoQ revenue decline and a decrease in market share to 9.3%.
SMIC, ranked third in revenue, saw no significant increase in wafer shipments during the third quarter. However, thanks to product mix optimization and the release of additional 12-inch capacity driving shipments, its revenue grew by 14.2% QoQ to $2.2 billion.
UMC, ranked fourth, experienced improvements in both wafer shipments and capacity utilization compared to the previous quarter, resulting in revenue growth to $1.87 billion, a 6.7% QoQ increase. GlobalFoundries, ranked fifth, benefited from inventory orders for peripheral ICs related to new smartphone and PC launches. This drove growth in wafer shipments and capacity utilization, with revenue increasing by 6.6% QoQ to $1.74 billion.
Surprising that the cyclicality of the semiconductor business is not mentioned. We continue to read that big volume commodity like memory chips are in a down cycle due to slower smart phone and personal computer sales.
Specialty chips as for AI are doing very well for TSMC but does Samsung participate.
Specialty chips are probably more profitable. And slower commodity chip demand maybe makes capacity available for mores specialty chips. That is probably behind better product mix.
And will Samsung be impacted by tariffs and possible trade conflict with China? Should investors be worried?
I think a lot of the more commodity type chips are not made on the leading edge processes. And yes, leading edge and specialty chips should be more profitable. Part of what’s killing Samsung is that their leading edge processes are stuck at around 20% yield rates on the types of chips they’re needed for. So Samsung is definitely losing business there. I wouldn’t invest in Samsung for their chip business at this point, though I think high bandwidth memory is working for them and that’s pretty profitable – SK Hynix and Micron are the only other players in that.