Is there a new trend in Xiaomi's self-developed chips?
On May 24, at Xiaomi's 2023 Q1 earnings conference, Xiaomi President Lu Weibing said that Xiaomi's determination to invest in chips has not wavered.
You know, OPPO recently closed its chip company Zheku, which has had a great impact on the domestic chip industry, and the outside world is worried about whether Xiaomi's self-developed chips will be affected.
President Lu's words seem to stabilize the military's heart, after all, the domestic self-developed chip at this stage is very difficult, and Xiaomi's road to self-developed chips is quite tortuous.
Xiaomi began to develop its own chips very early, and as early as 2014, it cooperated with Leadcore to establish Songuo Electronics, which is the third year of Xiaomi's establishment.
In February 2017, Lei Jun officially released Xiaomi's first self-developed mobile phone SoC chip surging S1, and Xiaomi became the fourth mobile phone manufacturer with chip self-research capabilities after Apple, Samsung and Huawei.
The Paper S1 is a mid-range chip mounted on Xiaomi 5C, with 8 high-performance cores and 3 high-performance cores, with a core frequency of up to 2.84GHz.
But judging by the results, The Paper S1 is undoubtedly a failure.
Subject to Qualcomm's baseband chip monopoly, The Paper S1 only supports mobile 4G, 3G, 2G and Unicom 2G networks, and does not support all networks of Unicom 3G, 4G and Telecom.
In the case that mainstream mobile phones are basically full Netcom, but also engage in non-full Netcom, in this case, the sales of Xiaomi 5C have been seriously affected.
In addition, the process of The Paper S1 is 28nm, but at that time, the 16nm chip has become the mainstream, and the Surging S1 is also weak in performance.
Non-full Netcom, weak performance, the final outcome can be imagined, the failure of The Paper S1, and even has a serious impact on the research and development progress of the Paper series chips.
From 2017 to 2018, rumors about the failure of the research and development of the surging S2 were endless, and by the end of 2018, the surging S2 tape-out failed 5 times, each time losing tens of millions of yuan.
Xiaomi, whose research and development is constantly frustrated, can only choose to change the track, and they give up the development of integrated chip SoCs and turn to other research and development.
In March 2021, Xiaomi MIX FOLD first released the self-developed image chip The Paper C1, the research and development of the surging C1 lasted for two years, with a capital investment of nearly 140 million yuan, which is also the second product launched by Xiaomi since it set foot in self-developed chips.
On December 28, 2021, Xiaomi held a new product conference and released Xiaomi's first self-developed charging chip The Paper P1, realizing a 120W single-cell charging solution for the first time.
In July 2022, Xiaomi launched the surging G1 chip for its own gaming mobile phone, which uses the 5nm process and has 8 high-performance cores and 3 high-performance cores.
Although in recent years, the launch of Xiaomi's self-developed chips is quite efficient, it can be seen that these chips only focus on a certain part of the function, not SoC chips.
It seems that Xiaomi has strategically abandoned SoC research and development, but starts from the research and development of independent chips, easy first and then difficult, and then overcomes SoC chips when there is enough technology accumulation.
Xiaomi's chip road, although tortuous, the road of other manufacturers' self-developed chips is not smooth.
Just when Xiaomi's research and development of the Pengbai S2 suffered setbacks, OPPO and vivo successively took over the baton and became the backbone of the self-developed chip army.
In August 2019, Zheku was officially established, as a self-developed chip company under OPPO, Zheku was once regarded as "HiSilicon II" by the outside world.
OPPO's chip research and development has been smooth, and in 2021, OPPO released two chips: MariSilicon X and MariSilicon Y.
MariSilicon X is OPPO's first self-designed and self-developed image-specific NPU chip.
Mariana X chip shipments reached tens of millions, successfully applied in OPPO and many of its mobile phone product lines, and became one of the main selling points of mobile phones.
MariSilicon Y is a flagship Bluetooth audio SoC chip based on TSMC's N6RF process.
OPPO's two chips use advanced process chips and are also quite distinctive in specific applications, so they have attracted special attention.
While people were still looking forward to Mariana's third chip, Zheku suddenly announced the shutdown, and the industry was shocked, because there were almost no signs before.
Why Zheku was shut down is still a matter of opinion, some people believe that OPPO directly sprinted 4nm chips, and the result was too big a step, and huge research and development was lost.
But some people believe that the reason for Zheku's shutdown may be interference from external forces, and even internal ghost leaks.
Some media speculate that the data generated by the EDA software used by Zheku will be obtained by the United States in some way, and the leakage of Zheku on the Internet is likely to be done by non-domestic EDA design software.
In addition, OPPO is not Huawei, there are other businesses to do profit support, once OPPO mobile phones encounter risks, it may be dragged down by Zheku, resulting in disastrous consequences.
Shutting down Zheku is more like a strong man's wrist breaking move, and the result is that OPPO's chip self-research can only be forced to stop, and it is difficult to say when it can be restarted.
Just when OPPO released the Mariana series of chips, vivo also released the first self-developed ISP chip V1 in September 2021.
Vivo's research and development strategy coincides with Xiaomi, and the first product chooses an image chip with a low development threshold, which is also related to vivo mobile phones' main photography.
In November 2022, at the dual-core imaging technology conference, vivo unveiled the second generation of self-developed image chip V2, which greatly upgraded the on-chip memory unit, Al computing unit, and image processing unit.
Compared with friends, vivo's core manufacturing plan announcement node came later and lower-key, and the investment and publicity in the semiconductor field were very cautious.
Why keep a low profile? It may be that Vivo's technical reserves are weak, but it may also be that it avoids the risk of "shooting the first bird" and does not want to be interfered with too much by the outside world.
Self-developed chips have never been a good road, on this road, Huawei, Xiaomi, OPPO, vivo have encountered their own problems.
Who can stand out on the road of self-developed chips? It's hard to say yet, I hope they can stick to the moment of success.
Resources:
Dark horse community: terminate R&D, dissolve the team, the most promising self-developed chip in China, gone!
Master Core: After the dissolution of OPPO Core Bureau, we have gone all over the past of Chinese mobile phone manufacturers in making cores
Web images
Editor: Chen Zhanxiang