Wen 丨 Chinese businessman Tao Strategy Ze Liang
At present, there are only 4 companies in the country that can design and mass-produce commercial 6nm chips:
Huawei HiSilicon, OPPO, MediaTek and Alibaba.
"In the next three years, spend 50 billion!"
At the first OPPO Future Technology Conference in 2019, OPPO founder Chen Mingyong aimed at research and development and sounded the rallying cry of the attack. In the same year, OPPO established the Chip Technical Committee (TMG) to seriously build cores.
Three years later, at the Future Technology Conference, OPPO officially released the first imaging-specific NPU chip, MariSilicon X.
MariSilicon X focuses on imaging, using DSA architecture, TSMC 6nm top process technology, support 4K, 20bit RAW, AI, Ultra HDR top image computing, can perform 18 trillion operations per second, making pixel-level image processing a reality.
The technical barriers are extremely high, the talent requirements are extremely high, and the difficulty of research and development and productization is extremely high. Is it really worth the large-scale R&D investment?
After all, the mobile phone market has long since bid farewell to the era of rapid incremental growth.
According to IDC data, in 2020, global smartphone shipments were about 1.29 billion units, down 5.9% year-on-year; shipments in the Chinese market were about 326 million units, down 11.2% year-on-year.
Global smartphone shipments have been declining year after year, and mobile phone manufacturers have quickly shuffled. Brands that once competed with OPPO have long been classified as "others" in market share.
The surviving large factories are not comfortable, and the stock game and in-roll competition are not good. More and more mobile phone brands have begun to use the same batch of technologies, the same batch of solutions, the same batch of suppliers, the same batch of components, and the product experience is becoming more and more similar.
What to do? Can't sit still, rolled into no way out, will eventually be voted by the market with their feet!
Chips are a key part of the competitiveness of mobile phones, in the past, manufacturers like Qualcomm, in order to sell more, will try to design chips with strong versatility structure. However, the improvement of universality will inevitably sacrifice the performance in specific use scenarios. Only self-research can establish a real distinction with other manufacturers.
This road has been successfully verified by Huawei's Kirin chip and its high-end mobile phone products. In recent years, Apple has also strengthened its core competitiveness with the innovative ability of M1 series chips on tablets and laptops.
"Heavy capital, heavy talents, heavy technology", making cores is a stupid way, but it is a real shortcut. Many business stories confirm that doing the hard and right thing is the only way to create long-term value.
Since OPPO announced the start of the core-making plan, news such as "recruiting chip talents in a big way" and "annual salary of graduates of 400,000" have been rushed to the hot search for many times, accompanied by a huge degree of discussion, and the voice of doubt has been rising and falling, after all, the idea of making a core is everyone's, but there are only a few that can be made.
In fact, it is precisely because of this that OPPO has encountered far more difficulties than these doubts have thought.
The first attempt to develop a chip, OPPO chose a 6nm process chip. This chip is not only extremely difficult to design, but also very expensive, the cost of tens of millions of dollars, the cost of each test tape-out film up to 100 million yuan, once it fails, a large number of manpower and material resources will be wasted, torn down and restarted.
Moreover, there is no reference design on the market for 6nm independent NPU, the relevant IP ecology is almost zero, OPPO can only choose to start from the core IP to start all self-research, which further increases the difficulty of research and development, prolonging the research and development cycle.
Without reference, cross the river by feeling the stones. In MariSilicon X, from the core IP of MariNeuro and MariLumi to the 6nm I/O interface, all of them are from OPPO's self-research, which is the valuable place of this chip.
Adhering to the core science and technology self-research reflects the responsibility of a large factory.
OPPO this self-developed chip, not only its own pursuit of the ultimate embodiment, but also the domestic core is imperative at the moment, in the late night before dawn blowing a counter-attack horn, when countless horns sound, is the "Chinese core" rises the day, no longer by other people's chips "card neck" is not just a slogan, but the Chinese enterprises to the world of the bottom.
——END——
Welcome to pay attention to [Huashang Taoluo], recognize the heroes, and read the legend of Taoli.
All rights reserved, unauthorized reproduction is prohibited!
Some of the images come from the Internet
If infringement is involved, please contact to delete it