Transformation of AI data center into "computing power factory": MIC points out three major business keys

The global AI infrastructure continues to expand. When the AI ​​Data Center (AIDC) further transforms from a traditional infrastructure carrying IT equipment into a "computing power factory" that produces computing power, the business model is also facing a reshuffle. The Institute of Industrial Intelligence (MIC) pointed out that future AIDC planning and operations will focus on three key issues, including reassessing market demand, site development, and adjusting financial planning.

The first is whether market demand can support the continued expansion of AIDC investment. MIC pointed out that future demand for AIDC construction will still come from the continued growth of orders and various off-balance sheet commitments from ultra-large-scale cloud service providers such as Amazon, Microsoft, Alphabet, and Oracle.

According to MIC estimates, the cloud revenue of the four major players will grow by approximately 20% to 60% annually in 2026, and the annual growth in backlog of orders will reach over 80% or even double. In addition, the construction of "sovereign AI" driven by government agencies around the world requires the implementation of sovereign computing power, which will form a new support force for the AIDC market from the public sector.

The second key issue is site. As AIDC is facing challenges such as electricity, cost, public opinion and regulatory friction, the first choice for collaboration on the demand side of AIDC will be suppliers and countries that can provide "low friction sites" to accelerate site development, operation and generate revenue.

In addition, the relationship between AIDC players and energy is also changing. In terms of power planning, global policies are shifting towards requiring AIDC owners to create, finance and manage more power resources on their own, which will change the positioning of AIDC companies from mere power users to "computing power + energy developers".

Future developers must therefore prioritize sites with low friction and locked-in power, while planning ahead for behind-the-meter (BTM), energy storage and dedicated power solutions.

The third major problem is the huge demand for funds. Adjusting financial planning will be the key to whether AIDC construction can continue to expand. MIC industry consultant Zheng Zhaolun said that the revenue model of AIDC operators is different from that of traditional data centers, and the future revenue model still needs to be continuously reconstructed. On the other hand, AIDC construction requires a huge investment, and how to obtain funds has become the key to whether it can continue to expand.

MIC expects that after 2027, AIDC players will continue to work with the ecosystem to promote computing power financial tools such as equipment financing, computing power financing and computing power securitization, thereby expanding financing path options.