Say goodbye to hybrid cloud, Alibaba Cloud closes its formation and fights a tough battle

Mondo Technology Updated on 2024-02-01

A few days ago, there was news that Alibaba Cloud adjusted its core sales business and renamed the original hybrid cloud division to the government and enterprise division. At the same time, the report said that the sales and service teams of the proprietary cloud were merged into the government and enterprise business department and became the first-level department of the government and enterprise business department, with the aim of improving the standardization and service capabilities of the proprietary cloud product solutions and better integrating them with partners.

However, at the end of November last year, when the new public cloud, hybrid cloud and international business departments were established, the business lines related to private cloud have been included in the hybrid cloud, so this time Alibaba Cloud may have further raised the priority of private cloud standardization services.

In practice, because government and enterprise customers have a high overlap ratio with hybrid cloud customers, the signal released by Ali's adjustment this time is also very clear. Alibaba Cloud will still strive to meet the needs of government and enterprise customers, but it is mainly responsible for the private cloud part, as for how to integrate different private clouds, public clouds, and internal IT resources into the hybrid cloud architecture that customers ultimately want, Alibaba Cloud may not want to mix too much.

Obviously, this adjustment is in line with Wu Yongming's "AI-driven, public cloud-first" development strategy. According to his plan, in the future, Alibaba Cloud will make trade-offs between all products and business models, reduce project-based sales orders, and increase investment in core public cloud products.

However, some people close to Alibaba Cloud commented, "Hybrid cloud has been engaged in two or three years, and people at the grassroots level know that this is a problematic business, and the senior management has only reacted now, indicating that Alibaba Cloud's organization is indeed flawed." ”

Probably no year in Alibaba Cloud's history has experienced as much turmoil as last year.

At the end of the year and the beginning of the year, when Zhang Jianfeng left his job due to a crash in the Hong Kong computer room, no one expected that Daniel Zhang, who made this decision and succeeded him, would leave even more completely than him a few months later. Before Zhang Jianfeng left office, Alibaba Cloud spent 14 years to increase its revenue from zero to 100 billion yuan, and there were only two transfers of the highest power in the process; In the past year or so, Daniel Zhang has replaced Zhang Jianfeng, and Wu Yongming has replaced Daniel Zhang twice, but the recent quarterly growth rate has dropped to 2%.

If the previous handover was a spillover of Alibaba Cloud's internal influence, the latter one reflects the fate of Alibaba Cloud that is difficult to escape under the macro adjustment of Alibaba Group.

The top management is more or less transmitted to the business level. Less than a year after the last Hong Kong data center went down, Alibaba Cloud failed again on November 12 last year, affecting not only a number of internal businesses including Taotian, Xianyu, DingTalk, etc., but also many external customers such as Luckin, Huya, and Mixue Bingcheng.

Therefore, after Daniel Zhang emphasized that "stability and security are the most basic responsibilities for customers for cloud computing", Alibaba Cloud decided to set up a "Product Architecture and Stability Department" (internally known as Risk Control Information System) at the end of that month, reporting directly to CTO Zhou Jingren. However, the effect of these measures remains to be seen, after all, the leader prefers the best reward, and the good fighter has no outstanding achievements, and how to fully motivate the team responsible for the stability of the structure is a problem.

In addition to the rotation of the top leaders, Cai Yinghua's departure is also a very important turning point, to a certain extent, marking that the strategy of growing to big B customers has not worked.

This process began in June 2021, when Zhang Jianfeng initiated a major overhaul of the sales line, setting up a sales interface and delivery system consisting of 18 industry departments and 16 regional lines. In the second year, Cai Yinghua took office, and he further clarified the organizational structure of "industry main construction and regional main battle", and continued to streamline and optimize the sales business line. The last wave of adjustments is in July 2023, when Cai Yinghua announced that the sales line will no longer be divided by China and international, but will be merged into 7 regions, including Beijing-Tianjin-Hebei, Guangdong, Shanghai, Zhejiang, and Suzhou-Anhui, and report directly to him.

Tsai's departure is not a bad thing for Alibaba Cloud. Alibaba Cloud first approached him in the hope that he could bring Huawei's systematic approach to government and enterprise customers. But what is Huawei's style of play? In fact, it is very detailed and highly customized. For example, 36kr once mentioned a detail in a manuscript that HUAWEI CLOUD can set up a company with dozens or even hundreds of people for a large customer. Obviously, this kind of practice is difficult to calculate in terms of ROI, and it is essentially contradictory to the general trend of standardization and integration in the cloud computing industry.

Of course, the bigger and more direct impact on Alibaba Cloud probably came from announcing the split and listing and then deciding to stop. According to the original assumption, Alibaba will dispose of Alibaba Cloud in a similar way to Tencent's disposal of JD.com and Meituan**, which means that Alibaba and Alibaba Cloud will be completely independent companies from now on.

But then Ali said that "restrictions on the export of advanced computing chips have brought uncertainty to the prospects of Cloud Intelligence Group." We believe that a complete spin-off of the Group may not enhance shareholder value as originally envisaged, so we have decided not to move forward with the full spin-off of the Group, but to focus on building a model for sustainable growth of the Group in the face of an uncertain environment."

However, the export restrictions on advanced computing chips were already "elephants in the room" before Alibaba Cloud decided to spin off the nodes. Since it was already considering going public at that time, it was impossible for the management not to be aware of the risks in this regard. Therefore, the outside world seems to be more willing to believe the latter sentence, Alibaba Cloud's spin-off "cannot improve shareholder value as originally envisaged", accompanied by rumors such as the initial contact with potential investors who do not recognize Alibaba Cloud's high valuation.

For Alibaba Cloud, it can be expected that the next year should not be as "thrilling" as 2023. After redefining the binding relationship with the group, Alibaba Cloud will have more room for long-term business strategic choices.

In fact, if Alibaba Cloud had indeed achieved a complete spin-off from the group, it would have brought more uncertainty than certainty. Because so far, whether it is Amazon's AWS abroad, or Microsoft's Azure, or the industry status quo of domestic counterparts, it shows that cloud computing vendors still need the business collaboration of their parent companies, and independent cloud vendors are usually struggling.

While chips are still a real constraint, the opportunities that advances in AI technology bring to cloud computing are certain. In his speech at last year's Apsara Conference, Wang Jian regarded the combination of artificial intelligence and cloud computing as the third wave. His analogy is that the relationship between cloud computing and GPT is the relationship between electricity and motors, and the computing power of cloud computing in the future will be consumed by motors, that is, models in these intelligent eras.

He also mentioned the previous two waves, namely cloud nativeization in the Internet industry and the adoption of cloud computing by traditional enterprises.

Frankly speaking, from the perspective of "New Position", the first wave was as thorough and swift as the autumn wind swept the leaves. In today's Internet industry, no startups will build their own IT rooms, and all their businesses have been running on the cloud from the beginning. But objectively speaking, the progress of the second wave mentioned by Wang Jian is clearly not satisfactory. The enthusiasm of traditional industries to go to the cloud has proved to be not so high, and the most typical evidence is the extremely weak performance of domestic cloud vendors in the past few years.

Nonetheless, for the third wave mentioned by Wang Jian, New Position believes that there are more reasons to be optimistic. Because the progress of the large model represented by ChatGPT is a kind of progress in general artificial intelligence. Versatility means that the technology itself is more widely applicable, and people need to do less customization, which is exactly what cloud vendors are most eager for.

Compared with its domestic counterparts, Alibaba Cloud's current layout in this area has certain advantages. Alibaba Cloud was the first in China to put forward the concept of Model as a Service (MaaS), so it has taken a step forward in terms of ecology and technical reserves, including the Moda community. We are more likely to focus on the bottlenecks caused by the number of AI training cards in our day-to-day discussions, and may overlook the difficulty of integrating these infrastructures into a single training system.

In fact, in the AI revolution brought about by this large model, the progress of cloud technology and model technology is strongly aligned with each other. The training of large models requires hundreds or thousands of servers to be connected, which involves various aspects such as distributed systems, training stability, and network congestion. Zhou Jingren said in an interview before that behind the large model is the upgrade of the entire cloud system and the competition of the technical system. Although the growth rate of Alibaba Cloud has not been good recently, the foundation laid from "Feitian" can still come in handy.

In addition, another positive factor for Alibaba Cloud is that it has formed a stable and healthy revenue cost structure. Alibaba Cloud's adjusted EBITA profit in the latest fiscal quarter increased from 38.7 billion yuan, a significant increase to 140.9 billion yuan. Other domestic peers have either not achieved a turnaround, or have only achieved quarterly profits. At a time when business health is particularly important, it may also make Alibaba Cloud more comfortable in dealing with AI changes.

Although Alibaba Cloud is still the "number one player" in China's cloud computing industry in terms of market share and technical level, it has also experienced the biggest impact in the past year. The continuous adjustment of the three CEOs, the complete split of the group and then the cancellation, and even the dramatic downtime have cast a shadow on its goal of becoming Alibaba's second growth curve, so that some grassroots employees once sighed "I don't know what to do next".

Wu Yongming, who is the CEO of Alibaba Group and the chairman and CEO of Taotian Group, is also in charge of Alibaba Cloud, which on the one hand confirms the importance of Alibaba Cloud's business to the group after the separation is stopped, which in turn helps stabilize the hearts and minds of the grassroots people, but on the other hand, it also raises doubts whether Wu Yongming's energy is enough to lead the transformation of Alibaba Cloud, especially in the context of AI change.

In any case, I hope that Alibaba Cloud has really adjusted its state.

The title picture and the accompanying pictures in the text** are on the Internet.

Related Pages