AI + PaaS: A New 'Variable' Emerging in China's Cloud Computing Market?

[baoshi.rao]

There are no small markets, only big opportunities yet to be discovered.

As enterprise digital transformation deepens, market demand is increasingly shifting towards PaaS and SaaS layers, making them the primary drivers of growth in the public cloud services market.

According to the latest IDC report, China's public cloud market is expected to achieve a compound annual growth rate (CAGR) of 26.9% between 2022 and 2027. Among these, PaaS (Platform as a Service) shows the fastest growth at 30.5%, followed closely by SaaS (Software as a Service) at 28.7%. With the completion of cloud computing infrastructure, China's public cloud market is transitioning from resource-driven to technology and business-driven models.

Meanwhile, the rapid development of generative AI and large models is accelerating the evolution of PaaS into the core capability of next-generation intelligent cloud platforms.

As the intermediary layer connecting the upper and lower levels, the PaaS layer must withstand the pressure from the rapid scaling of the IaaS layer while also supporting the platform-based capabilities derived from the functional accumulation of the SaaS layer. The PaaS layer will become a key component in helping enterprises comprehensively build and apply AI-native capabilities.

Looking back at the history of PaaS, it is evident that the development of PaaS has been closely intertwined with the evolution of cloud computing.

In 2006, Amazon Web Services was officially launched and commercialized, marking the beginning of cloud computing. In 2007, Salesforce released force.com, aimed at enabling third-party customers to develop, deploy, and manage applications on its platform. Subsequently, similar platforms began to proliferate.

In 2011, AWS launched the Amazon Beanstalk platform, Red Hat released OpenShift, and VMware introduced CloudFoundry. In 2015, IBM started building the Bluemix platform.

Since 2016, China's PaaS platform construction has accelerated, with tech giants like Alibaba and Huawei focusing on public cloud PaaS, and a number of startup PaaS service companies emerging in the market.

Currently, there are three main types of cloud computing service delivery models: IaaS, PaaS, and SaaS.

According to the definition by NIST (National Institute of Standards and Technology), PaaS integrates databases, development tools, and other components that support application service delivery into a platform provided as a service to developers. This allows developers to focus less on the resources required for application operation and maintenance, while users remotely utilize the application software developed by them.

From an overall architectural perspective, PaaS builds a foundational platform. Below PaaS, the platform is provided as a service to upstream SaaS and numerous developers.

PaaS providers offer services that encompass technical support for the platform, development and optimization of platform application systems, and a series of subsequent service modules. The services provided by PaaS operators are built on a strong and stable foundational operational platform, requiring the support of a professional technical team.

By transforming internet resources into APIs, they improve the development efficiency of third-party developers, reduce development costs, and enhance the agility of WEB application development.

PaaS plays a pivotal role in the cloud service industry chain. In this chain, the bottom layer (upstream of the industry chain) can provide services to any layer above it (downstream of the industry chain) or to end customers.

In the chain, the lower the layer, the higher the degree of standardization, and the more effective price competition becomes, such as with hardware devices and servers. The higher the layer, the closer it is to users, the lower the degree of standardization, and the more performance evaluation is influenced by various indicators such as performance stability, feature richness, interaction, and user experience, making simple horizontal comparisons difficult, as seen with SaaS layer services.

With the widespread adoption of cloud computing, enterprises are placing higher demands on cloud-native applications and new models of application development and management. The PaaS layer, which serves as a critical intermediary, is becoming increasingly important. PaaS targets software developers and SaaS-layer enterprises, requiring robust functionality and stability to maintain its position.

In the future, the "variable" that could reshape China's cloud computing market may well emerge from the often-overlooked PaaS layer.

First, PaaS enters the market as a platform, providing SaaS vendors with integration platforms (iPaaS) and application deployment and runtime platforms (aPaaS). From a deployment perspective, PaaS can be further subdivided into database services, application development, application infrastructure, middleware services, and more.

As a platform, PaaS offers far greater potential than SaaS and possesses the capability to transform the cloud computing market landscape.

Secondly, PaaS has ample room for growth. Naturally positioned between IaaS and SaaS, PaaS benefits from the current focus of IaaS giants on AI, which creates a niche for PaaS to flourish.

Thirdly, and most crucially, there is a strong market demand for PaaS. While IaaS dominates and SaaS thrives in the cloud computing sector, PaaS has yet to fully take off. This lag is primarily due to the previously fragmented and underdeveloped state of SaaS, which did not necessitate PaaS solutions.

However, with the explosion and maturation of the enterprise application market, SaaS applications increasingly require specialization, cross-layer integration, efficiency, collaboration, and seamless connectivity. This has significantly elevated the importance of PaaS, leading to a foreseeable surge in PaaS adoption and rapid evolution.

With the advancement of artificial intelligence large models, it is gradually becoming a reality for various industries to process big data using AI in the cloud.

According to IDC research, enterprise users' evolving demands for cloud services mainly manifest in three aspects: acquiring AI application capabilities on the cloud, obtaining AI-enhanced tools on the cloud, and achieving intelligent-driven application innovation on the cloud.

IDC believes the next-generation cloud will be "Intelligent Cloud" that adapts to enterprises' smart development needs on demand. The cloud serves as the foundation for AI implementation and growth, while AI will also propel cloud platform development. Technically, enterprises need intelligent architectures and systems to accelerate the deployment of smart infrastructure. At the business level, they should leverage cloud platforms' resource management capabilities to enhance workflow automation and achieve intelligent operations. Ecologically, enterprises must rely on next-generation cloud-based intelligent tools to improve product development efficiency and industrial collaboration capabilities, enabling intelligent innovation.

As the crucial middle layer connecting upper and lower levels of cloud platforms, the PaaS layer's market demand and functionality will continue to strengthen with the development of next-generation cloud computing.

On one hand, PaaS-layer tools and products enable enterprises to better schedule and manage the foundational resources of the IaaS (Infrastructure as a Service) layer. On the other hand, SaaS-layer applications are gradually descending to become part of the PaaS platform's functionality, providing more convenient services for enterprise users.

Data from the first half of 2023 shows this trend becoming increasingly evident, with PaaS growing at twice the rate of IaaS. In H1 2023, China's overall public cloud services market (IaaS/PaaS/SaaS) reached $19.01 billion, with the IaaS market accounting for $11.29 billion (13.2% YoY growth) and the PaaS market reaching $3.29 billion (26.3% YoY growth).

In fact, PaaS becoming the next-generation core growth driver for cloud computing is already an industry consensus. According to the China Academy of Information and Communications Technology's "Cloud Computing White Paper 2023," China's PaaS market grew by 74.49% in 2022, reaching ¥34.2 billion in total revenue, while the SaaS sector achieved ¥47.2 billion in revenue with 27.57% growth.

Enterprise users are no longer satisfied with merely using IaaS for resource cloudification but expect comprehensive cloudification of management and business systems to seize new development opportunities. Meanwhile, influenced by factors like AI large models, PaaS and SaaS will become the main battleground for cloud computing growth in the coming years.

PaaS has now been developing for over 15 years. Over the past 15 years, aPaaS has arguably been the representative of PaaS, with its model maturing steadily. However, as technologies like artificial intelligence, the Internet of Things, and big data increasingly penetrate enterprise applications, they have deepened the value of PaaS and generated new demands.

Particularly with the current surge in generative AI, AI as a foundation for reconstructing business operations has become feasible. Data-centric PaaS services, represented by AI PaaS, are emerging as the future trend in PaaS development.

AI PaaS delivers AI capabilities as a service, providing enterprises with a user-friendly platform to quickly build and scale their AI applications. Another critical point is that PaaS platforms need to shift from being closed to open.

In the era of digital intelligence, one of the core competitive advantages for enterprises is the ability to construct an open ecosystem. Such an ecosystem blurs the boundaries between industries, companies, and even competition. For example, the automotive industry now extends beyond traditional suppliers to include internet and cloud computing sectors. In this evolving landscape, PaaS platforms must also build open, boundaryless ecosystem services to meet enterprises' needs for low-cost and rapid digital intelligence transformation.

From a technical perspective, AI PaaS is a platform that integrates artificial intelligence and machine learning services for building, training, and deploying AI-driven applications. It solves the infrastructure creation and maintenance problems enterprises face when rapidly developing AI products.

AI PaaS provides enterprises with an easy-to-use platform, enabling them to quickly set up and scale their AI applications.

The key components of AI PaaS include pre-trained machine learning models and AI APIs. These components can process and analyze data, solve specific tasks, and deliver corresponding results.

Pre-trained machine learning models help enterprises rapidly develop and deploy AI solutions, while AI APIs offer a set of predefined functions and tools, making it easier for developers to build and call their own AI applications.

Using AI PaaS offers numerous benefits, starting with reduced development costs and time. AI PaaS provides pre-built infrastructure and environments, eliminating the need to start from scratch. Additionally, AI PaaS is highly scalable, allowing for rapid expansion based on demand.

Moreover, AI PaaS comes with a variety of powerful built-in tools and functions, enabling developers to build and deploy their AI applications more efficiently.

When selecting an AI PaaS service, several factors should be considered. First is data quality, as the training and performance results of AI PaaS directly depend on the quality of the training data.

Second is technical compatibility. Ensure that the AI PaaS is compatible with existing systems and tools to avoid unnecessary complications. Additionally, API availability is a crucial consideration—ensure the stability and reliability of the APIs.

In various industries, the application prospects of AI PaaS are extremely broad. From finance and healthcare to manufacturing, AI PaaS can help enterprises achieve automation and intelligence, significantly improving work efficiency and business competitiveness.

The future development trend of AI PaaS involves greater application in edge computing and the Internet of Things (IoT), along with continuous enhancements in performance and reliability to meet evolving market demands.

During the implementation of AI PaaS, some challenges may arise. Poor data quality, technical compatibility issues, and insufficient API reliability can all impact the performance and effectiveness of AI PaaS.

To overcome these challenges, enterprises should ensure data accuracy and completeness, conduct thorough testing and validation. Additionally, maintaining close collaboration with AI PaaS providers, promptly reporting issues, and seeking support are key to addressing these challenges.

AI PaaS, as a platform integrating artificial intelligence and machine learning services, will play an increasingly important role in the future. It helps enterprises quickly build AI-driven applications, improving work efficiency and business competitiveness.

With the continuous development of technology and increasing market demand, the performance and reliability of AI PaaS will be further enhanced, laying a solid foundation for enterprises' digital transformation and intelligent development.

The continuous evolution of PaaS is closely linked to the wave of digital transformation in enterprise business. Digital transformation aims at business model transformation and business process efficiency improvement, as the saying goes, "Cloud as the foundation, intelligence as the application"—that is, using cloud as the system and base, and leveraging the comprehensive application of IT technologies such as artificial intelligence, big data, and blockchain to solve systematic problems.

PaaS, especially when combined with data analysis and AI services, will undoubtedly unleash its greater potential, becoming another powerful tool to enhance the level of business digitization.