At the 2026 World Intelligent Industry Expo, Li Jun, Chairman of Sugon, articulated that AI for Science (AI4S) is propelling humanity swiftly into the "Age of Scientific Intelligence." He pointed out that while traditional supercomputing has long depended on mathematical models, the advent of large-scale data models in the current era serves as a complementary force. Looking ahead, AI computing is set to transition into a form characterized by "hyper-intelligence integration," which will seamlessly merge high-precision scientific computing with low-precision intelligent computing. Presently, we are witnessing an accelerated integration across various scenarios, including scientific computing, model training, and intelligent reasoning. The competitive landscape in AI infrastructure has evolved from a mere comparison of computational prowess to a more holistic contest that encompasses technology, industry, and ecosystems. In light of these developments, comprehensive coverage of computational power across all precision levels and robust system integration are poised to become the defining features of next-generation AI computing.