Because of certain well-known reasons, has been doing its own processor Huawei, the past two years had to stop.
Some time ago Huawei took a 5G model to the shelves for a long time - Huawei Mate40E Pro.
More attention is paid to its processor, Kirin 9000L, than to the phone.
Some of you may spit: not that there is no chip? How to pull out a new one?
In fact, if we take a closer look at the parameters, we can easily find that the Kirin 9000L is somehow a residual version of the Kirin 9000E.
Therefore, some users speculate that
Huawei is indeed out of cores and has no choice but to pull out the "used chips" from the 9000E production line that should have been eliminated to use.
But Huawei still has no chance to revive first, it is estimated that there are many partners and I have the same question:.
In addition to Huawei, there are no other domestic manufacturers can make the core?
OPPO will soon enter the manufacture of core
Some time ago, a foreign media said that OPPO's IC design subsidiary, Shanghai Zheku, has already started the development of AP (application processor) chips and SoC chips.
SoC is probably understood by everyone, but what is AP?
In simple terms, SoCs can be divided into three broad segments.
AP for local data processing, BP (baseband, antenna) for processing communication signals, and other co-processors.
The CPU, GPU, NPU and other modules that we often talk about are all part of the AP.
According to the news, OPPO intends to launch its own AP next year (2023), which is expected to use TSMC's 6nm process.
In another year (2024), the SoC mobile platform with integrated AP and 5G communication module will be launched, and the process will be upgraded to TSMC 4nm.
Some industry insiders have previously made similar statements that have some credibility.
The planning is still pretty good, but will the execution be smooth?
The road to self-research is destined to be rocky
Last year Xiaomi, OPPO, vivo three major manufacturers, have launched their own self-research chip.
What OPPO has come up with is the MariSilicon X based on TSMC's 6nm.
This NPU chip integrates two important boards of AI (artificial intelligence) and ISP (image processing).
As a chip with an AI algorithm of up to 18TOPS, MariSilicon X also beat the iPhone 13 Pro Max (15.8TOPS) at the launch.
But if you look back into the home Android camp, the Snapdragon 888 and 888 Plus have 26TOPS and 32TOPS of computing power respectively.
In this way, the MariSilicon X is not weak, but it does not rank at the top either.
In addition, like Xiaomi's Surge P1 (charging chip), Surge C1 (imaging chip), and Vivo's V1 (imaging chip), MariSilicon X mainly assists in optimizing a single function of the phone.
Compared with the most core CPU, GPU, etc. of the phone, it is still a bit distant.
Again, the debut of MariSilicon X did not play to its full strength.
(MariSilicon X) played about 30% of the role," confessed Jie Li, president of Find product line.
Mostly, Mariana is very capable and needs time to consistently polish and get better at it.
As a newcomer in the field of chips, OPPO's performance is still a bit green.
Facing the NPU chip, OPPO has to deal with so many technical challenges.
Then when dealing with the more complex AP chips, OPPO has even more to deal with.
In addition, OPPO two years later, if you enter the SoC chip, hard just flagship market may be a bit hanging.
This year's Snapdragon 8 Gen1 and Tiangui 9000 are both already on the 4nm process.
If then OPPO's own SoC is really 4nm as the news says, theoretically it is already 2 years behind.
Of course, if the chip industry is stagnant in these 2 years without updates, then OPPO still has a chance.
But that's basically impossible.
So if OPPO's own processor is first placed on the mid-range or sub-flagship models, it may be more in line with the actual situation.
There is no overnight success in making chips
The first year is for NPU, the second year is for AP, and the third year is for SoC.
Such planning looks smooth, but the execution is not easy at all.
Take Huawei's road to self-developed chips as an example.
Not to mention the initial unpopular K3V2 products, from Kirin 910 to the real high-end Kirin 980, it took several generations to figure out.
Although the conditions for making cores today are much better than before, but the road that should be taken is not less.
Strong as Apple, in the baseband problem was also stuck for a long time.
In order to get rid of the dependence on Qualcomm, the iPhone had a period of time to choose Intel's baseband.
However, Intel is really inexperienced in this area of mobile communications.
This also makes some models of iPhone, on top of the already poor signal, even worse.
Ultimately, for the sake of communication experience and to keep up with the 5G era, the iPhone 12 series can only return to Qualcomm baseband.
Today's Apple is also trying hard to develop its own baseband.
Supply chain sources say Apple will start producing 5G baseband chips for the iPhone in 2023.
Although the front is a bit of cold water, but I still believe that OPPO has enough strength to do self-research SoC.
On the one hand, there is no doubt about OPPO's self-research capability.
OPPO has always had a wide range of patent layouts at home and abroad.
Last year, for the third year, the company was ranked among the top 10 global applicants of PCT international patents.
The appearance of MariSilicon X is also a good proof that OPPO has the ability to build cores.
On the other hand, OPPO's outstanding market performance in the past two years has also made its financial strength grow a lot.
After all, if you want to build a core, you can't help but smash money.
I believe that as long as OPPO can hold their breath and spend a few years figuring it out, they will be able to create a domestic core that makes the country proud.