After a gap of four years, it was finally waiting for Huawei's chief financial officer Meng Wanzhou to attend the annual business financial report conference again, which was also her first public appearance after returning to China in 2021. Meng Wanzhou said in a firm, calm and gentle voice at the press conference: "The world has changed a lot in the past four years, and the motherland has also changed a lot, and I have been studying hard in the past few months of returning to China and keeping up with the pace of society." ”
Huawei pointed out at the press conference on the 28th that its revenue in 2021 was 636.8 billion yuan, down 28.6% year-on-year; net profit was 113.7 billion yuan, up 75.9% year-on-year. In 2021, R&D investment reached a record high of 142.7 billion yuan, and the cumulative R&D investment in the past decade reached 845 billion yuan.
At this conference, Chinese and foreign media invariably focused on a very hot topic: Does Huawei have the intention of building its own chip fab?
For the answer to this question, Huawei responded twice, let's see how Huawei answered:
The first question is: How does Huawei solve the chip supply chain problem? Are you planning to build your own fab? Huawei replied: From sand to chips, solving the problem of semiconductors is a complex and lengthy process that requires patience. In the current technology lockdown, investment (semiconductors) will become very commercially valuable, and we are happy to see more and more companies participate in the market, and we are more happy to see them succeed. Under the current (Huawei) advanced process technology is not available, especially the acquisition of some single technology nodes is more difficult, we will seek a system breakthrough, the future chip layout will adopt a multi-core plus software structure to inject new vitality into the chip.
Soon, foreign media continued to ask the question of "whether Huawei intends to build its own wafer production line".
The second time Huawei answered: In 2019, Huawei shipped 120 million units of 5G mobile phones, which requires 120 million baseband chips based on one mobile phone and one baseband chip. Looking at the base stations, in 2019, Huawei delivered 1 million base stations, and each base station needs a chip, a total of 1 million chips. To C's business experienced a huge decline in 2020, but To B's business continuity is guaranteed.
Huawei has been investing in research and development, and the research and development direction is moving towards theoretical reconstruction, software and system architecture reconstruction, such as exchanging area for performance, stacking for performance, which can make Huawei products very competitive even if they do not use advanced technology, and Huawei will continue to work hard in this direction.
Huawei's response to how to solve the chip supply problem is very specific, and the product will change towards structural changes and get rid of dependence on advanced process technology.
Interestingly, for the media twice asked "whether you plan to build your own fab", Huawei did not directly admit it, nor did it deny it, always answering in a different way, trying to give a better and more advanced answer. From this point, it can be seen that Huawei is a very honest and responsible company, and will not arbitrarily deny the "ongoing" style.
Huawei has repeatedly stressed the importance of R&D investment at the press conference, believing that the ability to precipitate and accumulate investment in R&D is the real value of the enterprise.
Investing in R&D with 10% of revenue per year is the Basic Law, and in the past decade it has gone from 13.2% in 2013 to a record high of 22.4% in 2021. In 2021, huawei's revenue was 636.8 billion yuan, and the investment in research and development was as high as 142.7 billion yuan, and Huawei's cumulative investment in research and development in the past decade was as high as 845 billion yuan.
As for some people questioning huawei's revenue decline in 2021, how to ensure that there are sufficient funds to invest in research and development in the future?
Huawei said that in 2021, Huawei's operating cash flow will be 59.7 billion yuan, and its net cash will be 241.2 billion yuan, which is a guarantee for future-oriented research and development. In the future, R&D investment will focus on system architecture optimization, software upgrading, and by solving technical and process problems, providing highly reliable products, and will invest in establishing a reliable and reliable supply chain to ensure the continuous supply of products.
Huawei stressed that to solve the immediate problems is not to rely on food and clothing, but to improve the performance of software, use theoretical exploration to solve the problems of the process, and establish a highly credible supply chain. Since advanced semiconductor processes are not available now, it is necessary to increase strategic investment and actively seek system breakthroughs.
For example, the demand for AI in mobile phones, whether it is taking pictures or speech recognition, is increasing, but traditional neural networks need a lot of computing power, and the price is very power-consuming, so there will be pressure to use advanced semiconductor processes.
Is there a way for AI to be more energy efficient and save chip area? Using additive neural networks instead of multiplicative calculations, reducing floating-point multiplication would be one direction. For example, under the use of additive neural networks in 16-bit AI, the power consumption can be greatly reduced by 88% compared with multiplication, and the power consumption area can be reduced by 66%, and the product is more suitable for the intelligent era.
For example, a power amplifier device with a base station whose maximum power consumption is an RF unit can use a third-generation semiconductor gallium nitride, which can save at least 20% of energy consumption.
In the strategy of smart cars, it is reiterated that Huawei does not build cars, and it is necessary to use 30 years of ICT capabilities to help car companies build good cars, and there are already more than 30 auto parts to help Huawei intelligent transformation, and car companies will also provide solutions.
Finally, Huawei pointed out that the core of solving the long-term development problem will be talent, or it will return to the strategy of attracting talent. Last year, the number of recruits was as high as 26,000, of which 300 were Huawei's "genius teenagers" program, and this year plans to recruit another 10,000 fresh graduates, only excellent talents can solve Huawei's immediate problems and promote Huawei's further development.