What Alibaba Cloud’s New President Reveals About the Future of Cloud Computing

About the Future of Cloud Computing

Q: You notice the focus of cloud computing competition is changing, and many still care about price?

Answer: Price is a low‑level competition; ultimately value returns to the core, reflected on both the technology side and the application side. Cloud started as infrastructure, but Alibaba’s original vision was a data‑processing platform, which was disconnected from business ten years ago because few companies were data‑driven.

Later we realized that just doing this was not enough; to succeed in China you also need solid infrastructure.

We initially tried to build data processing on open‑source systems, but found it insufficient and had to develop our own. Alibaba not only built a cloud infrastructure system but also a data‑processing system (now called MaxCompute, formerly ODPS), which is unique both domestically and internationally.

The competition at the infrastructure layer is about maturity, stability, and price, while the upper layers compete on AI and other advanced capabilities.

Q: Why does open‑source not work for you?

Answer: Open‑source software companies have never handled data at the scale we have. Hadoop and Spark made a big step beyond databases with distributed MapReduce, but their data volumes are still far behind Alibaba’s. Open‑source also cannot meet our stability requirements. We built a 5K‑node cluster, breaking the 5,000‑machine limit of open‑source at the time, allowing larger scale processing. Cost is also a factor; as data volume grows, we cannot let costs grow linearly.

We started large‑scale mixed workloads (mixing online and offline processing on the same cluster) to save resources and meet high scheduling demands, which open‑source cannot achieve.

Without real business use, it is hard to design a flexible, cost‑effective, highly stable processing platform.

Q: Hybrid cloud is a hot topic abroad; how do you see its impact on cloud computing?

Answer: Hybrid cloud remains an important direction. Many customers, especially governments and banks, have rigid requirements for controllable and isolated applications. Public‑cloud isolation is improving with hardware solutions, but hybrid cloud demand also stems from architectural considerations, so it will stay prevalent for a long time.

Q: Regarding the backendization of cloud computing, you said Alibaba Cloud does not do SaaS. Could cloud vendors become forgotten like power plants?

Answer: Being forgotten yet indispensable is a good state, like air. Cloud should be a platform, not a software system. It provides the underlying infrastructure for countless businesses, similar to electricity or water utilities.

Alibaba’s core strength lies in handling massive systems, rapid technology innovation, and foresight on technology evolution.

We have abundant business scenarios that let us outperform others, but many industries still need specialized partners for implementation, consulting, and training, which Alibaba Cloud cannot cover alone.

Q: Do you have expectations about Alibaba’s scale compared to Oracle or SAP?

Answer: A platform company with 10,000 people is already huge; Alibaba’s scale is comparable to a few thousand‑person operation.

Q: How will AI and open‑source affect cloud computing?

Answer: Open‑source will standardize cloud components like containers, making cloud development more consistent. Large companies using these components will stabilize the ecosystem, and engineers will be familiar with the stack, easing integration.

Q: Will open‑source become paid?

Answer: Some companies now charge for services, but the trend is that cloud providers will offer native cloud solutions, while big internet firms may release fully free open‑source projects.

Q: What technology are you most focused on in cloud computing?

Answer: Hybrid cloud development, security, fast‑growing cloud‑native databases (challenging Oracle replacement), and IaaS layer improvements like containers and hardware isolation (e.g., Shenlong servers).

Cloud is evolving rapidly; business models also shift, with Microsoft moving down from SaaS and AWS taking a different path.

Q: About Alibaba’s middle platform, what is its relationship with organization?

Answer: It is closely tied to organizational capability, trust, and assessment mechanisms.

Middle platform is both an organization and a concept. If a technology fits a high‑level product (e.g., intelligent robot), it belongs to the relevant department, not necessarily the middle platform.

We have a technical committee and sub‑teams for voice, video, vision, etc., solving unified problems before escalating.

Q: What KPI do you focus on?

Answer: Scale, public‑cloud market share, and ecosystem depth (thousands of partners, aiming for higher business proportion on Alibaba Cloud).

Q: Will DingTalk collaborate more closely with Alibaba Cloud?

Answer: DingTalk will enhance online work collaboration, complementing cloud’s data and AI strengths. We aim to integrate work collaboration as a key differentiator for internet companies.

Q: How do small programs (mini‑apps) fit into the larger deployment?

Answer: Mini‑apps are crucial as many applications move to mobile. They can run on major traffic entrances (e.g., WeChat, Alibaba) rather than independent apps, enabling a “one‑cloud‑many‑ends” strategy.

Q: What is the “being integrated” concept?

Answer: Alibaba Cloud aims to be a component integrated into various solutions, similar to Intel CPUs being integrated into computers.

We are shifting from centralized delivery to distributed delivery, leveraging partners for implementation.

Q: How will the ecosystem evolve under the “being integrated” strategy?

Answer: It will remain a collaborative ecosystem where partners co‑define industry solutions, with Alibaba providing core technologies as part of their solutions.

Q: How do you view Alibaba Cloud’s short‑comings?

Answer: Our B2B service capability needs ecosystem support, and we must improve productization (documentation, manuals) of our platform.

Q: How do you measure the “being integrated” initiative?

Answer: By the number and scale of partners with business, and building flagship cases with top partners like Deloitte and Accenture.

Q: How will you keep close communication with customers?

Answer: We co‑create market samples with partners, integrating our technology into their solutions.

Q: How does AI affect technology requirements for enterprises?

Answer: Enterprises should focus on industry‑specific AI applications; core AI capabilities are provided by cloud, but the business‑specific integration is what adds value.

Q: Will large IT teams continue to develop in‑house solutions?

Answer: Likely less; specialization and outsourcing will increase efficiency, allowing companies to focus on core advantages.

Q: What overlooked issues should traditional enterprises consider when moving to cloud?

Answer: Operations and maintenance become critical for system health; security is actually easier on cloud due to rapid patching and vendor support.

Q: How will cloud adoption change personnel and capabilities?

Answer: Roles like DBA become less critical; third‑party firms will handle operations, and companies will need fewer large IT teams.

Q: How does AI’s high technical demand affect enterprise technology choices?

Answer: Enterprises should leverage cloud‑provided AI services and focus on industry‑specific use cases rather than building AI from scratch.

Q: Will large enterprises continue extensive in‑house development?

Answer: No, specialization and leveraging external expertise will become the norm to achieve higher efficiency.

“More dry goods, more gains”

Not writing a line of code, complete machine‑vision algorithm development

Evolution of E‑commerce Search Technology

N Ways to Detect Anomalies, All Unpacked by Alibaba Engineers

Follow Machine Intelligence – Seize Future Possibilities