Smart Stuff (public number: zhidxcom)
Author | Li Shuiqing
Editor | Heart Edge
The era of high-speed growth in the Internet industry is far away, and it has become an indisputable fact. The company's website has been updated with the latest information on the Internet. The Internet giants like Ali, Tencent, and Baidu need to think about the second growth curve more than ever before.
Cloud intelligence services, is a bet on the future of the track by each eye.
The results are already visible - Ali Cloud Q3 revenue exceeded 20 billion yuan for the first time in four consecutive quarters of profitability; Baidu 2 billion a month to engage in research and development, intelligent cloud revenue growth of more than 70%; Tencent cloud this side, but also for the first time to put forward the "cloud intelligence" strategy, three years to realize the team to achieve the "cloud intelligence" strategy, and the "cloud intelligence" strategy. Affected by the U.S. sanctions, Huawei's consumer business and carrier business suffered a sharp setback, while the cloud intelligence business has become its new lifeblood of growth.
A cloud intelligence battle with BATH (Baidu, Ali, Tencent, and Huawei) as the protagonist to enter thousands of industries has been ready to start.
Huawei Ren Zhengfei personally supervised the "five legions", Ali Cloud Intelligence will establish 18 industry business units, Tencent and Baidu also carried out intensive structural adjustments or strategic upgrades, focusing on the digitalization of the market of the government and enterprises, "sharpening the knife to the cattle and sheep! ".
Zhang Jianfeng, President of Ali Cloud Intelligence, Zhang Pingan, President of Huawei Cloud, Tang Daosheng, President of Tencent CSIG (Cloud and Intelligent Industry Business Group), and Wang Haifeng, CTO of Baidu, as the helmsmen of this cloud intelligence war, are making important choices in their careers at this critical time of economic recovery and global reshuffling in the aftermath of the 2021 epidemic.
Despite the different problems they face, they are all united in leading their teams into the "deep water" of AI.
Anyone who has met Tang Daosheng is likely to be impressed by his programmer's style, which is not very fluent in Hong Kong Mandarin, and seems to be very meticulous and not very articulate. This master's degree in electrical engineering from Stanford in 1997 became Tencent QQ's vice president of research and development and general manager of the QQ Show product center at the age of 34, and became the helmsman of Tencent's CSIG in 2018.
In 2021, more and more internal Tencent employees will see Tang Daosheng in and out of purely business occasions, including cooperation with Tianyi Group in the construction industry, cooperation with China Merchants Ports in the logistics and transportation industry, and signing a contract with Guangzhou Automobile Group in the travel industry....... In order to lead Tencent to venture into the field of the previously unfamiliar industrial Internet, do The technical background of Tang Daosheng is becoming more and more like a "sales" representative.
Tang Daosheng, President of Tencent CSIG
"Tencent does not have a To B gene" - this is a common question from the outside world in September 2018 when Tencent CSIG was established. Tencent CSIG integrates the company's cloud computing, AI, smart industry and other multiple technical capabilities, Tencent's many scientists come from here, but how to land the technology in the industry, for Tang Daosheng and these scientists there is not a small chasm to cross.
Even today, three years later, Tencent CSIG has expanded from more than 1,000 people three years ago to more than 10,000 people, Tang Daosheng still frankly said in a recent interview to do the To B point of strain - "To B engineering cycle is longer, the requirements of financial management and project management higher, as well as the introduction of talent from traditional industries also needs to rub off."
Deeper business and larger organizations also mean facing new problems. In the AI-driven industrial Internet project, large manufacturers have platform potential, but it is difficult to be as flexible as the AI start-ups to play "guerrilla warfare", so in many AI bidding programs may not be able to exceed the small manufacturers.
In this regard, Tang Daosheng in May this year, has started from the organizational structure of the first adjustment in three years - announced the establishment of Tang Daosheng led by the Technology Committee, to "top-down" approach, to reduce the code duplication of internal development, open source, and improve the efficiency of project promotion. Improve the efficiency of project promotion. For example, in the recent launch of Tencent's Digital Intelligence Platform and other projects, we have seen signs of the results of the integration of cross-CSIG and AI Lab, WeChat, and Sogou's multi-departmental technology.
Deeper is the team culture change. Tang Daosheng himself personally into the "sales" of the move is consistent with the third anniversary of the CSIG in September this year, Tang Daosheng unprecedentedly upgraded the values of the business group out of the "dare to work hard and dare to speak! The requirement of "dare to fight and dare to speak". He pointed out sharply that in recent years, the introduction of a lot of employees from different companies, colleagues are more constrained, do not dare to speak out the problem has affected the efficiency of the enterprise, so the team should be as entrepreneurial "dare to fight", "dare to speak".
At present, the domestic AI technology landing climax has a trend of arrival, however, previously active in the primary market AI start-ups "money pool" has bottomed out, have turned to the secondary market to seek blood transfusion supply. For Tencent CSIG, such as the AI players with big factory pockets, a market harvesting war seems to have just begun.
At the Tencent Eco-Conference held in November this year, Tang Daosheng proposed "Tencent Cloud Intelligence" for the first time, announcing a set of "global intelligence" and a complete cloud intelligence architecture and the latest self-developed AI chips, AI development platforms, and intelligent industrial solutions, marking the beginning of Tencent's industrial cloud development.
Tencent Cloud Intelligence Strategy Architecture
Let's take a look at the industry landing, according to our statistics, Tencent Cloud has taken a lot of big orders in the large-scale smart city projects that have been announced this year. In April this year, Guangzhou's Huangpu smart city project was worth 210 million yuan; in May this year, Jiangmen's 360 million yuan smart city project; in August this year, Xiangyang's 110 million yuan Telematics Pilot Zone project; and in September this year, Shanghai's Xuhui District's 120 million yuan "One Network, One Management" project.... ...Tencent Cloud has acted as the main or only successful bidder. Official data show that the Tencent Cloud AI daily global 1.2 billion users and 200W + customers, processing more than 5 billion images per day, processing voice more than a million hours, processing natural language more than 100 billion.
"Further down the road, I think it will enter the 'deep water' and reach out to these industries that are further away from Tencent's original strengths." Tang Daosheng talked in a recent interview.
Unlike Tencent, which came up with "cloud intelligence" later, Baidu, though smaller in size in BATH, was almost the first to put forward "intelligent cloud" in BATH, and the most determined to "All in" artificial intelligence. "
If you want to say that you're not going to be able to get a good deal on your own, you're going to have to do it.
If you want to say that AI creators are most afraid of being replaced by a big company, it is undoubtedly Baidu. With a domestic PaddlePaddle (Flying Paddle) that rivals the top international deep learning frameworks, Apollo, which has been awarded 411 self-driving test licenses, and the Kunlun AI chip, which has a mass production of more than 20,000 chips,...... Baidu's founder, Robin Li, can continue to invest 2 billion yuan per month in research and development in the years to follow, and the AI founders don't necessarily have the courage to do so. not necessarily have this bottom.
Then look at the helm of the Baidu Intelligent Cloud Baidu CTO Wang Haifeng, he and many founders of AI startups, is a technical, but also "a person without news. Graduated from Tsinghua Physics, his father educated him from childhood to become a technical talent, he also wish to be elected to the IEEE Fellow in 2021, and become the first person in mainland China who is also an ACL Fellow, Baidu internal many people say he is: "Weak, but at the same time will not be easily moved by others."
Baidu CTO Wang Haifeng
It is such a seemingly gentle and elegant technology faction, driving the Baidu intelligent cloud department in the experience of organizational changes in the unstable situation, turned in 2021 to achieve business in 2021 continuous quarterly growth of 70%, the second growth curve of the Baidu company.
Baidu Intelligent Cloud just went through a turbulent year in 2020. in early 2020, Robin Li integrated ACG (Baidu Cloud), AIG (AI technology platform system), and TG (basic technology system) into the "Baidu Artificial Intelligence System" (AIG), and let Baidu Cloud General Manager Yin Shiming report to the new CTO Wang Haifeng commander-in-chief report. However, only three months later, Yin Shiming left.
At that time, in the domestic cloud market, Aliyun and Tencent cloud were in the top two, and the later Huawei cloud growth rate of 220%, Baidu Intelligent Cloud in the rear may be thrown farther and farther away. Bloomberg commented at the time, over the past decade, from natural language processing to the field of voice interaction, Baidu has invested billions of dollars, but this effort because of Baidu's artificial intelligence business head Andrew Ng (Andrew Ng) and other key executives left the company and fell into a more problematic situation. The market expressed skepticism about the potential return on the company's R&D investment, and the stock price then dropped to freezing.
Heading into 2021, Wang Haifeng was in total control, rarely coming out to make impressive speeches but quietly leading Baidu's smart cloud to rapid growth this year.
According to Baidu's financial report for the third quarter (Q3) of 2021, which showed that it achieved revenue of 31.9 billion yuan in Q3, the trend of weak growth in the advertising business has been reflected. However, Baidu Intelligent Cloud revenue grew 73% year-on-year in the quarter, becoming an important driver of Baidu's core revenue growth. In Baidu's first quarter fiscal year 2021 earnings call, Robin Li bluntly said, "AI and cloud, the next three years will account for more than half of Baidu's revenue."
From the business point of view, in the past six months, Baidu will Baidu Brain, a "soft and hard AI production platform" upgraded to version 7.0, the intelligent transportation, smart cities and other programs to land in a number of cities. For example, in September this year, Baidu has taken three consecutive billion projects, including 117 million yuan Shaoxing intelligent expressway project, 168 million yuan Cangzhou Economic Development Zone automatic driving and vehicle road coordination project, 179 million yuan Tongxiang industrial Internet project.
With Baidu CTO Wang Haifeng stabilized the Baidu Intelligent Cloud team, he further announced the new strategy of Baidu Intelligent Cloud at this year's Baidu Cloud Intelligence Summit: "Promote the digital transformation of enterprises based on the cloud, and accelerate the intelligent upgrading of industries with AI as the engine, and the cloud and intelligence as a whole, empowering thousands of lines and hundreds of industries, and promoting the high-quality development of the economy. "
Comparing to 2020, Baidu's statement on the intelligent cloud strategy is still "based on cloud computing, with AI as the gripper, focusing on important tracks." It can be seen that Baidu Intelligent Cloud emphasizes more on AI technology landing entity industry, that is to say, Baidu AI is also going to enter the deep water.
With Baidu's AI technology landing more and more on the road, it will undoubtedly become one of the biggest rivals of a group of AI startups rushing to IPO.
Also experiencing an organizational restructuring is Huawei, and it looks like Huawei Cloud has an even rougher road this year.
At the Huawei Developer Conference in April this year, Huawei's consumer division name Yu Chengdong, then president of Huawei Cloud, was still cheering and shouting for the landing of Huawei Cloud AI, saying that AI development should say goodbye to the workshop style and move towards industrialization, and bringing updates to Huawei Cloud Pangu big model, MindSpore and other software.
But just five months later, at Huawei's Fully Connected Conference in September, the speaker had changed to new CEO Pingan Zhang. Zhang Pingan joined Huawei in 1996, served as COO of Huawei's enterprise business BG, president of the software product line, Huawei is proficient in software and market veterans.
Zhang Pingan, President of Huawei Cloud
In fact, this year, Huawei Cloud has made several organizational restructuring. First in January, and then in March this year, the Huawei cloud and computing BG in the original server, storage, etc. out of the "network products and solutions", the BG in the cloud independent of Cloud BU, which is just one year after the merger.
In this case, the cloud and AI technology-driven To B business has undoubtedly become Huawei's new growth lifeline.
As a general of Huawei Cloud on the verge of taking office, Pingan Zhang talked at Huawei's Fully Connected Conference in September, "All infrastructure and applications should be cloud-based in the future, and we need to actively embrace AI so that data can play an important role in business decisions."
Like other Huawei leaders such as Yu Chengdong, Zhang Pingan's words and actions often show Huawei's pragmatic, wolfish, and unambiguous. He mentioned in many speeches that the biggest problem facing AI landing is the high cost. In order to reduce the cost of AI landing, Pingan Zhang led the team to update a series of AI root technologies such as the Cloud Sky Chip AI solver and Huawei Cloud Kaitian aPaaS. It can be seen that, on the basis of Huawei's complete set of ascendant AI full-stack technology that includes hardware, AI arithmetic development, AI model development, and AI application development, Huawei Cloud is increasing its investment in the PaaS layer from the cloud infrastructure layer to the upper layer.
On the one hand, the technology is iterating towards the extreme level, and on the other, it is promoting the landing of cloud and AI technology in the industry from the whole group level.
In 2021, Huawei successively set up five major industry corps - the Coal Corps, the Customs and Ports Corps, the Intelligent Highway Corps, the Data Center Energy Corps, and the Intelligent Photovoltaic Corps - marking the official opening of Huawei's curtain for the comprehensive landing of cloud, AI, and network technologies for thousands of industries.
Huawei's five industry legions are personally supervised by Ren Zhengfei, and the most elite elite soldiers are extracted from each BG and independently parallel to them, integrating cloud, network, intelligence and other vertical capabilities to strike the industry horizontally and accurately. This is the first time that Huawei has clarified the legion combat approach in its organizational structure.
Ren Zhengfei, founder of Huawei, presided over the Huawei Legion Formation Conference
Ren Zhengfei has mentioned the expectations of the Legion - for example, can the technology be matured in 2-3 years, and then provide mine intelligent services to the world? From the point of view of the existing progress, Huawei "mine Hong" mine operating system has been successfully landed in the Shendong four mines, from the point to cover the AI real-time risk early warning identification and other functions, to the launch of the mine Hong operating system to make the entire mine become connected, verified the agility and efficiency of the legion of combat methods.
Driven by a strong need for survival, and helmed by founder Ren Zhengfei himself and strongly promoted by fierce general Zhang Pingan, Huawei's cloud and smart technologies have perhaps become the fiercest pole of landing in BATH.
At the beginning of the establishment of Huawei Cloud in 2017, Huawei made a social commitment "not to do applications, not to touch the data, and not to do equity investment". In 2020, Huawei Cloud upgraded this mission to "empower applications, enable data, and be the black land of the intelligent world."
Faced with internal and external survival pressures, the question of whether Huawei Cloud can grasp the boundaries of eco-cooperation and keep its promises is of concern to many partners.
As we enter 2021, many people have mentioned that AliCloud is getting more and more like Huawei with its hands in the mud.
At the AliCloud Summit in May this year, Zhang Jianfeng, President of AliCloud Intelligence Business Group, revealed that AliCloud has recently completed a major organizational upgrade, one of which is the establishment of 18 industry departments, led by the industry general manager of the industry digital innovation. Similar to Tencent Tang Daosheng, Zhang Jianfeng is a technical architect who fought for Ali's fist product "Taobao", and led the team to solve the "double 11" cloud computing problems. 2019, Zhang Jianfeng stepped down as CTO of Ali Group, focusing on Ali Cloud. Intelligent business group and Dharma Academy research and development.
Zhang Jianfeng, President of Ali Cloud Intelligence
Since Zhang Jianfeng took the helm, Ali Cloud has expanded the landing industry from three in 2019 to seven in 2020, and now it has expanded to 18 and risen to the sector. This speed of expansion confirms Zhang Jianfeng's ability to "fight".
AliCloud Intelligence covers the expansion of the industry (according to public information, if there is any error, welcome to correct)
Then look at the regional layout horizontally, AliCloud Intelligence has set up 16 regional departments, appointing general managers of branch offices to connect with local customers. It is important to know that this is not in the history of Ali, in the past has been Huawei, Haikang, such as ICT manufacturers are keen to do, has now become an important organizational strategy of Ali cloud.
Considering that May is also a time when Huawei's coal army led by Ren Zhengfei is riding high, AliCloud's big move in May to set up 18 industry sectors and 16 regions to power the digitization of government and enterprises can easily be seen as a counterpoint to Huawei. The two companies, China's largest cloud giant in the smart city sector and the largest communications giant, seem to be entering a phase of close combat in cloud intelligence technology.
However, Zhang Jianfeng dismissed the adjustments as competition-related. He talked about how it was mainly due to major internal conflicts within Aliyun itself, and that there was a big difference between Aliyun's services and what customers wanted, which prompted Aliyun to make the shift.
Whether subjective benchmarking or not, but Ali is jumping out of the previous habits of mind of doing e-commerce consumption, through a more grounded approach to the government and enterprise digital customers, to realize the technology cash.
The change in posture is accompanied by the upturn of AliCloud in financial indicators. According to the Q3 financial report of AliCloud released on November 18, AliCloud revenue of 20.007 billion yuan, an increase of 33% year-on-year, adjusted EBITA profit (EBITA) of 396 million yuan, for the first time to achieve profitability for four consecutive quarters.
A large share of AliCloud's revenue comes from infrastructure, SaaS, and other revenues, but AI technology has become increasingly integrated into its cloud services.
A brief review, this year, Ali Dharma Academy open source pre-trained language modeling system AliceMind, which has topped six authoritative lists of NLP, and launched a visual intelligence open platform with more than 100 visual AI capabilities; AliCloud Intelligence has launched a big data AI platform that allows enterprises to use it "out of the box. --Ali Lingjie", training the world's first breakthrough 10 trillion parameters of the AI model; chip, Pingtou brother's self-developed cloud native processor chip Yitian 710 debut, XuanTie CPU has shipped more than 2.5 billion, becoming the largest domestic application scale of the Domestic CPU ......
Zhang Jianfeng talked about in an interview with Xinhua, 12 years ago, Alibaba was the first to invest in the exploration of cloud computing in China, and became one of the three clouds in the world. In the next decade, Alibaba will unswervingly continue to increase investment in basic and cutting-edge technologies, and conduct scientific research in the fields of artificial intelligence, chips, quantum computing, blockchain and other areas.
It is no wonder that Alibaba Cloud is becoming more and more like Huawei, although Zhang Jianfeng is not willing to admit, but Huawei is good at "rooted, breaking through the sky", Alibaba Cloud Intelligence has been striding in this direction this year, and Huawei to compete with the digitalization of the market of government and enterprises.
It can be seen, whether in the Internet development inflection point looking for new momentum BAT, or in the U.S. sanctions to find a new granary of Huawei, they are cloud intelligence as a more important lifeline in the operation, trying to switch to a new growth momentum.
As the helmsman of each cloud intelligence business, the objects mentioned in this article include both excellent product managers from the technical school, as well as management talents who have experienced great winds and waves in the past battlefield, and now, their gazes are coincidentally cast from the technology to the industry, ready to roll up their sleeves, hands into the mud, and make a big effort.
The giant's actions in the field of AI "aboriginal" AI companies may bring a small impact. On the one hand, the giant's high fight can bring the progress of the overall industry ecology, so as to create a more suitable environment for AI landing; on the other hand, their own more irreplaceable competitiveness is what? What is the most suitable business model? Will become a non-giant AI enterprises need to think about things.
When the AI landing into the deep water, all kinds of AI players are pushing the technology chariot to each industry, each region of the industrial digital transformation of the field, a showdown.