The rapid evolution of AI big models has led to an increased demand for intelligent computing centers, reflecting a broader shift in the technological landscapeAs industries worldwide continue to push for advancements in AI and its applications, such centers are becoming essential for training foundational and domain-specific modelsDespite some fluctuations in demand, the continuous iteration of AI technologies is propelling these centers toward a new phase of development.

Recent data from IDC indicates that the market for intelligent computing services in China will grow by an impressive 79.6% year-on-year in the first half of 2024, reaching a market scale of approximately 146.1 billion yuanNotably, the intelligent computing integration service sector is expected to surge by 168.4%, with a market size of 57.0 billion yuan, while the GenAI IaaS market is predicted to expand by a staggering 203.6%, hitting 52.0 billion yuan

These statistics underscore the significance and urgency of constructing such centers in response to the burgeoning need for computation capacity.

According to Guo Rensheng, Senior Vice President at Prologis China, the demand for computing power has been robust over the past yearCompanies are increasingly requesting that servers be immediately operational upon arrival, allowing them to quickly set up networks and provide services internally or externallyHe notes that the time from data center infrastructure delivery to operating services has significantly improved, illustrating the industry's responsiveness to client needs.

Despite this growth, the intelligent computing industry is in a phase of ongoing technological iterationFor long-term sustainability, it is essential for the industry chain to collaboratively explore and develop shared frameworks and methodologies to accommodate the accelerating advancements in computational technology.

The significance of intelligent computing centers extends beyond mere infrastructure; they function as vital components for the burgeoning digital economy

• New Challenges from Technology to Industry Chain
• Why Apple Continues to Reject Nvidia
• Digital Currencies Reshape Finance
• Healthy Plus Group Undergoes Major Transformation
• Gold and Silver Rise to Meet Expectations

An industry expert highlighted that the combined digital economic output from the United States, China, Germany, Japan, and South Korea exceeded $33 trillion in 2023, demonstrating an 8% year-on-year growth, with the digital economy now accounting for approximately 60% of the GDP—a notable increase from 2019 figuresIn China, for instance, the digital economy constituted 42.8% of the GDP in 2023 and contributed significantly to overall GDP growth.

As opposed to the previous cycle that centered around CPU-based general-purpose computing centers, the current paradigm shift towards GPU-centric intelligent computing centers brings with it a new set of complex technical requirementsGuo emphasizes that as the scale of these centers increases along with power density, there is a greater demand for robust physical infrastructure to support their operations, including advanced cooling systems and electricity distribution capabilities.

For larger projects, such as those involving thousands of computing servers, power requirements can reach 30 megawatts at the outset, with plans to eventually scale up to 100 megawatts, indicating a radical shift in capacity planning

The industry's trend is moving away from traditional single-unit data centers to larger base-style centers equipped to meet extensive demands.

Current technological trends indicate a rapid evolution within the intelligent computing supply chain, especially a movement towards integrated systems that combine power, refrigeration, and liquid cooling capabilities within cabinet designsGuo notes that Prologis is focusing on making infrastructure layers more resilient and intelligent while collaborating closely with customers to explore various cooling setups that align with different operational needs.

As Prologis continues to invest heavily in data center operations, they are witnessing substantial growth within this sector, which has become the company's second-largest business streamRecent reports indicate a 63% increase in sales within the Prologis data center division, reflecting the rapid expansion of intelligent computing services driven by major clients' investments in building their intelligent computing infrastructures.

Looking toward the future, IDC's research manager Yang Yang predicts that the intelligent computing service market will maintain its significant growth trajectory over the next five years

This is largely due to the increasing demand from leading clients for advanced training capabilitiesFurthermore, the industry is still in the early stages of the GenAI transition, with substantial room for growth as expenditure on inference-side computing is expected to rise dramatically.

As organizations increasingly move toward digital transformation and AI applications, the role of intelligent computing centers will evolveGuo highlights “centralization” as a clear trend in future computing infrastructure and services, indicating that the focus will intensify around green energy, adaptable architecture, and aligning operational advantages with customer needs for heightened computational capabilities.

In the realm of cooling technologies, liquid cooling is gaining tractionPrologis has been collaborating with clients to develop liquid cooling test environments, examining the balance between operational costs and energy savings with the intent to further integrate this technology into their facilities

This focus on emerging cooling methodologies aligns with Prologis's goal of facilitating effective temperature management, which is crucial for the high-density environments of intelligent computing centers.

Both the integration of new storage technologies and reducing supply chain costs are factors set to benefit future data center developments, making storage solutions more financially viableYang emphasizes that as construction accelerates, there will be a sharper focus on professional service providers and the commercialization of operations to ensure efficient deployment and maintenance of large-scale computing systems.

Amidst rapid advancements, Guo acknowledges the challenges posed by high-density computing environments that necessitate innovative operational management and intelligent monitoring systemsTraditional server units, which can maintain cooling for around 15 minutes during a power interruption, face serious risks in high-density settings, where temperature can rise alarmingly within mere minutes