In the field of advanced chip manufacturing, looking at the world, only TSMC, Intel and Samsung are left. At present, the competition between TSMC and Samsung under 7nm has attracted much attention. According to a report by DigiTimes, Samsung will directly skip the 4nm advanced process and move to mass production of the 3nm process. This move may make Samsung lead TSMC, but this is also full of risks.
Samsung skipped the 4nm advanced process it originally planned to invest in and completely abandoned its investment plan. This decision may eventually cause its domain to lose a lot of customer orders, but at the same time it may be because of skipping a generation of chip technology, the development of other technologies More resource support.
Why Samsung decided to skip the 4nm process
The competition in the chip industry is extremely fierce, especially in the field of advanced processes. Samsung is one of the very few companies that has the potential to challenge TSMC, the chip foundry leader. In fact, in 2019, Samsung is preparing to defeat TSMC in the 5nm process, but in the end it has not been successful, but the competition between Samsung and TSMC is still continuing. This plan is likely to be part of the competition strategy.
As Samsung's largest competitor, TSMC's process technology is more than a year ahead of Samsung, and it has increasingly become the first choice for foundries of chip manufacturers. Process orders below 7nm are almost full. Although TSMC may lose its second largest customer, Huawei HiSilicon, recently MediaTek, Qualcomm and other companies have added orders to TSMC, and TSMC's normal revenue may not fluctuate significantly. It is estimated that in 2020, TSMC's 7nm and 5nm combined revenue will reach 40%, plus strong demand for 16nm process orders will account for about 20%, and advanced processes will account for about 60% of revenue.
In addition, TSMC is also preparing for mass production of 4nm and 3nm, and plans to mass produce 4nm advanced process chips from 2022 or 2023. The time when TSMC can mass produce 4nm or 3nm is unclear, and Samsung is facing many difficulties.
Samsung's process of advancing advanced processes is intermittent. It used to skip the 7nm process to the 7nm LPP EUV process for the first mover. Although its 7nm EUV is known for full-scale production, due to limited yields and technical bottlenecks, except for Qualcomm, other chip manufacturers are not daring to place orders, which makes Samsung's EUV process technology more difficult. Affected by the COVID-19 epidemic in 2020, EUV and other important equipment and technologies from abroad will be completely blocked, and the 5nm mass production time will also slow down.
Perhaps considering the gap with TSMC, Samsung hopes to invest more capital resources in the 3nm process, preemptively overtake, and overtake TSMC in advanced processes.
Samsung skips the 4nm process with many risks
Abandoning investment in the 4nm process, Samsung is expected to obtain customer orders other than Qualcomm on the 3nm advanced process, leading TSMC, but this move is still risky for Samsung.
First, gradually reduce the chip size to improve efficiency. This process has been stable. At least in the past few generations, companies such as Samsung and TSMC have largely relied on chip processes that have been mass-produced to develop next-generation processes. Skip the previously planned 4nm advanced process directly into the 3nm process, which may prevent Samsung from solving the problems that it would have encountered in 4nm, which means that it will automatically give up the opportunity to lay a solid foundation for 3nm.
In addition, Samsung also faces risks in raising funds. Although Samsung has secured some orders to mass-produce 5nm chips for Qualcomm's Snapdragon X60 5G modem and is expected to maintain its 5nm production in 2021, its potential partners are about to launch products based on the 4nm process, but Samsung jumped directly to 3nm, this undoubtedly means that a large number of potential orders may be lost. This will result in more customer orders for TSMC, which is about to launch 4nm mass production.
Samsung may offset these losses by cutting the costs of other OEMs, making its solution more attractive, but this does not guarantee the effectiveness of this initiative. Mainstream chip companies that hope to achieve the best performance will undoubtedly turn to more efficient 4nm chip.
Whether Samsung can directly overtake TSMC in this transition can not be concluded at present, but generally speaking, the cooperation between chip design manufacturers and wafer foundries is close, and it will not be replaced easily and frequently. If Samsung’s strategy of skipping the 4nm process is successful, it can help Samsung maintain its relationship with existing customers and increase customer stickiness, but there are still many challenges to truly overtake TSMC.