The Computing Power Industry is Thriving

The rapid rise of emerging fields such as scientific research, artificial intelligence, digital twins, and the metaverse is driving the rapid growth of global computing power, fostering multiple innovations in computing technologies and products. This transformation is reshaping and reconstructing industry structures, positioning computing power as a new engine for global digital economic development and a key focus of national strategic competition.

Steady Growth in Computing Power Scale

The global computing power scale is maintaining a steady and high growth trend. In the era of the digital economy characterized by ubiquitous sensing, connectivity, and intelligence, global data volume and computing power scale continue to grow at high levels. According to relevant research reports, the global data production in 2021 reached 67ZB, with an average annual growth rate of over 26% in the past three years. Research shows that in 2021, the total global computing power reached 615 EFlops, with a growth rate of 44%, including basic computing power at 369 EFlops, intelligent computing power at 232 EFlops, and supercomputing power at 14 EFlops.

It is projected that by 2030, humanity will enter the Yottabyte (YB) era, with the global computing power expected to reach 56 ZFlops, representing an annual growth rate of 65%. Specifically:

• Basic computing power will reach 3.3 ZFlops, with an average annual growth rate of 27%.

• Intelligent computing power will grow to 52.5 ZFlops, with an annual growth rate exceeding 80%.

• Supercomputing power will reach 0.2 ZFlops, with an annual growth rate exceeding 34%.

Diversified demand accelerates the diversification and upgrading of computing power. Diverse intelligent scenarios require diverse types of computing power. The rapid rise of emerging fields such as artificial intelligence, scientific research, and the metaverse is driving higher computing power demands. For example, Intel predicts that the metaverse will require a 1,000-fold increase in computing power, while NVIDIA estimates that real-time rendering for immersive experiences still lacks computing power by a factor of one million.

In terms of basic computing power, cloud computing has essentially returned to pre-pandemic growth levels. According to IDC data, the global cloud computing IaaS market grew to $91.35 billion in 2021, a 35.64% year-on-year increase. The combined IaaS + PaaS market reached $159.6 billion, growing 37.08% year-on-year. Cloud computing is expected to become the mainstream general-purpose computing model of the future, providing foundational support for technologies like big data, artificial intelligence, and 5G, and driving industry digital transformation and intelligent upgrades.

Regarding intelligent computing, there is a growing mismatch between the demand for processing massive, complex data and the supply of single-type computing power. With the rapid growth of global data, more than 80% of data is unstructured (e.g., text, images, voice, and video data). As Moore's Law and Dennard scaling slow down, chip performance improvements, represented by CPUs, are limited to less than 15% annually, making it challenging to meet the needs for processing unstructured data such as videos and images. The demand for diverse intelligent computing power is increasingly urgent.

The Frontier supercomputer at Oak Ridge National Laboratory (ORNL) in the United States achieved a performance of 1.102 EFlops (1 quintillion floating-point operations per second) in the Linpack benchmark, surpassing Japan's Fugaku and becoming the world's first publicly confirmed exascale supercomputer. This marks the official entry of global supercomputing into the Exascale computing era.