Collaborations between tech giants and capital giants are reshaping the global AI infrastructure competitive landscape.

According to recent reports, Alphabet's Google and BlackRock, a global asset management giant, are planning to establish an artificial intelligence cloud computing company, aiming to seize the "high ground" of next-generation computing infrastructure. The collaboration model is clear: BlackRock will invest $5 billion in equity capital and hold a majority stake in the new company; Google will provide technology - including TPU dedicated chips as hardware foundation, as well as accompanying software stacks and cloud service capabilities.

This is not a simple "capital + technology" combination, but rather a precise complement of capabilities. BlackRock excels in long-term capital allocation and heavy asset operations, while Google holds globally leading AI chip architecture and cloud-native experience. Together, they aim to build a "vertical cloud vendor" focused on high-efficiency, large-scale AI training and inference, avoiding direct confrontation with general cloud giants and instead focusing on the computing demand in the era of large models.

More notably, their computing target: the new company plans to build a 500 megawatt (MW) computing capacity by 2027, and reserve space for continuous expansion. As a reference, the single regional computing clusters of current global top cloud providers are mostly in the hundreds of megawatts. If this goal is achieved on schedule, it will directly enter the ranks of the world's top intelligent computing centers.

For the industry, this move sends a clear signal: as large models enter the second half of the competition characterized by "computing power, energy efficiency, and implementation," pure technological advantages are no longer sufficient to build a moat. The composite capability of "capital endurance, engineering efficiency, and ecological collaboration" is becoming the core variable in the next round of competition.