This article is original, please do not copy and move, violators will be investigated
What is the limit of the chip process? People who adhere to Moore's Law theory will definitely feel that there is no limit, or that it can continue to develop. But I have to admit that below the 5nm chip process, every little bit of progress is very difficult.
Even if it has achieved 4nm, the improvement of process yield still plagues TSMC and Samsung. The next few years may still be able to move forward, but how to develop in the more distant future? TSMC has sent new news, planning to start with chip packaging and stacking technology, can it redefine the chip industry?
TSMC explores chip stacking technology
Chip manufacturing giants TSMC and Samsung have maintained high-end chip manufacturing advantages for a long time, and both have achieved mass production of 4nm chip processes. But at this point, it is very difficult to make a major breakthrough at once.
The construction of a high-end chip production line is tens of billions of dollars invested. If it is only a resource investment, TSMC and Samsung have strong financial resources and can still afford to bear these capital expenditures.
The key is that investing money always has to be effective, otherwise there is no difference between it and wasted effort. Regarding chips, TSMC has officially sent new news that it will explore chip packaging and stacking technologies.
In detail, TSMC and Samsung, Intel and other giants have formed industry alliances for chip packaging technology to create standards for related small chip packaging and stacking technology, and promote the continuous progress of the electronic chip era.
This alliance is expected to enable the interconnection of small chip products between different manufacturers, and achieve more planned performance characteristics under a unified standard. This is a forward concept of chip packaging, from the independent operation to the interconnection, if the final interoperability industry is realized, it may bring different changes to the chip market.
However, compared with the exploration of chip packaging technology, chip stacking technology is more eye-catching. Moore's Law is difficult to maintain the development of the chip technology industry in the next few decades, even if the 1nm chip is conquered, can there be more advanced process nodes further down?
It is possible that the breakthrough cost of this process technology is also very large, such as difficult to control power consumption and excessive performance. So is there a technology that can achieve a corresponding improvement effect without relying too much on process breakthroughs? Perhaps the chip stacking technology explored by TSMC can become a development direction.
If other companies were exploring the technology, it might not have caused much repercussions. However, TSMC is the world's number one chip manufacturer, and even TSMC is exploring the layout of the technology, indicating that the technical field of chip stacking is valued and recognized by TSMC.
Can chip stacking redefine the chip industry?
Theoretically, chip stacking technology is to stack chips with different functional characteristics into a system-level chip that can be coordinated in the form of advanced packaging technology.
TSMC has the world's leading 3D chip packaging technology, which itself has advantages in the manufacture of chip stacking processes. If successful, it may create a new era of electronic chips, opening up a new path in the case of Moore's Law gradually facing physical limits, so that chip performance can continue to extend.
So can this technology redefine the chip industry? This may depend on the actual application performance. At present, most of the action plans carried out by giants such as TSMC and Intel are on the theoretical basis. Theoretically, two chips are stacked together to create a more powerful system of chips, then break the shackles of the process and move into the next golden decade of semiconductors.
But in fact, whether it can smoothly stack chips with different characteristics together, and can also play an effect of one plus one equals two, or even greater than two, there is no practical evidence.
Assuming that chip stacking technology can be successful, so that different chips between different chip manufacturers can be combined with each other and stacked into SOC processors, then the chip industry can certainly be redefined. Because this not only proposes a solution to the limits of Moore's Law, but also opens up a new direction of technological development, redefining the chip industry will become a reality.
Huawei chose the right one
No one knows how long Moore's Law will last, but it must be prepared for the future development of the chip manufacturing industry. Once chip stacking technology is taken as a clear direction, it may also prove that Huawei has chosen the right one.
Huawei's HiSilicon Semiconductor once published a chip superposition technology patent diagram, from the patent diagram can be seen that a main chip has a chip 1 and chip 2 composition, chip 1 and chip 2 between the continuous synchronous signal transmission, and then the chip is transmitted to the main chip.
This patent once caused a heated discussion from the outside world, some people think that this technology is too early, not very realistic, two chips superimposed together can control power consumption? There are also views that this is a good innovative technology and is worth a try.
Because there are no relevant practical cases, it remains to be seen whether this technology is feasible. But since TSMC has explored chip stacking technology, it shows that Huawei has chosen the right one, and this direction is not necessarily wrong. If we can make real progress, it is naturally a good thing to use the domestic chip manufacturing process to achieve greater performance breakthroughs.
summary
Regarding chips, TSMC has a new direction of exploration, using chip packaging technology to achieve chip stacking, and Huawei has already had relevant layouts before this. If this road is feasible in the future, it may be a turnaround. But practice is the only criterion for testing truth, and it is advisable to observe more before the final breakthrough is completed.
Do you think chip stacking technology is feasible? Feel free to leave a message below to share.
Technology is fun and takes you to know new things about technology