Entering　2022,　AIGC,　represented　by　foundation　model　such　as　ChatGPT,　GPT4,　Sora,　and　GPT-4o,　is　breaking　through　the　wall　rapidly.　Generative　artificial　intelligence　foundation　model　technology　is　rapidly　iterating　and　continuously　evolving,　becoming　a　revolutionary　tool　for　content　generation,　knowledge　production,　and　human-computer　interaction.　With　the　increasing　number　of　parameters　in　foundation　model　and　the　increasing　complexity　of　deep　learning　algorithms,　the　demand　for　computing　power　in　foundation　model　is　further　increasing.　Massive　data　and　large-scale　parameters　result　in　extremely　large　computational　loads,　limited　storage　on　a　single　computer　server,　and　limited　computer　capabilities.　Training　a　foundation　model　with　billions　of　parameters　requires　tens　of　thousands　of　GPU　cards　for　synchronous　computation,　and　high-performance　computing　power　networks　have　become　the　main　method　and　means　to　meet　the　demand　for　large　computing　power.This　paper　summarizes　the　development　process　of　intelligent　computing,　explores　the　demand　for　computing　power　under　the　background　of　large　models,　and　analyzes　computing　power,　computing　power　networks,　and　technology　ecology　based　on　this,　and　analyzes　their　related　technologies.　©　2024　IEEE.