Samsung Foundry Has a Problem Every Chipmaker Wants

SammyGuru is reader-supported. We have affiliate and sponsored partnerships, so we may earn a commission when you buy through links on our site — at no extra cost to you. Learn more.
Samsung Foundry may finally be in a position many contract chipmakers would love to reach. According to industry sources, the company has started applying allocation for some manufacturing processes. In simple terms, Samsung is no longer accepting every chip order that comes its way and is becoming more selective.
Samsung Foundry is becoming more selective with new chip orders
The change is largely being driven by strong demand for AI chips. Samsung is reportedly giving priority to existing customers and only accepting new projects that meet its requirements. The surge in AI chip orders has filled up production lines, leaving less room for every customer that wants manufacturing capacity.
Samsung already manufactures autonomous driving chips for Tesla and inference chips for Groq. It has also expanded its work with Nvidia and Google. Reports claim these customers are enough to keep Samsung’s 4nm process nearly fully booked through next year. Some 8nm production lines are also said to be operating close to full capacity.
This approach isn’t anything new in the chip industry. Foundries often prefer working on a smaller number of large-volume orders instead of taking on many smaller projects. It’s the same approach TSMC has followed for years, helping it keep its factories busy while focusing on major customers.
Samsung could also be seeing the benefits of TSMC’s limited production capacity. Some companies that left Samsung in the past because of yield concerns are reportedly coming back. Some are considering Samsung as their main manufacturing partner again, while others simply want a second supplier to avoid putting all their eggs in one basket.
An industry source told Chosun Biz that allocation for Samsung Foundry processes began this year. Instead of accepting every order, Samsung is reportedly choosing projects that offer better long-term value. If AI chip demand continues at its current pace, this selective approach may stick around for some time.
















