Global Computing Power Expected to Grow in Next Five Years, Emergence of AI-Native Applications Drives AI Computing Opportunities
-
Native Applications Proliferate, AI Computing Opportunities Highlighted
On November 21, Baidu released its third-quarter financial report. Baidu's revenue for the third quarter reached 34.447 billion yuan, a 6% year-on-year increase; net profit attributable to Baidu (non-GAAP) was 7.3 billion yuan, up 23% year-on-year. Baidu's R&D expenses in the third quarter increased by 6% year-on-year. "We will continue to prioritize investments in artificial intelligence, particularly generative AI and foundational models," said Baidu CFO Luo Rong. Industry analysts believe that the continuous emergence of generative AI applications is attracting sustained investment. Computing infrastructure is a critical component of AI investment, with accelerated vitality across the industrial chain. Sectors such as memory chips, optical modules, IDC data centers, liquid cooling, fiber optic cables, and connectors are poised for explosive growth opportunities.
Domestic AI Computing Development Expected to Accelerate
From the Q3 2023 reports, many computing-related stocks saw significant earnings growth. Specifically, five companies, including HiSoft and Saite Intelligent, reported year-on-year net profit growth exceeding 100% in the first three quarters. In Q3 2023, nearly 10 companies, such as Jialitu and Lianhua Health, doubled their net profits year-on-year.
Galaxy Securities analyst Zhao Liangbi noted in a research report that, based on Q3 results, the PE levels of relevant communication sub-sectors showed overall recovery after September, driven by sustained catalysts in computing leasing and scheduling.
Globally, with the continuous release of AI computing demand, Nvidia, a core chip manufacturer in the commercial computing sector, saw explosive growth. Its Q2 FY2024 revenue hit a record high of $13.51 billion, up 101% year-on-year, with data center revenue reaching $10.3 billion, a 171% increase. Net profit surged 843% to $6.188 billion. Demand from China accounted for 20% to 25% of Nvidia's data center revenue. Nvidia forecasts Q3 revenue of $16 billion, up approximately 170% year-on-year. Nvidia founder and CEO Jensen Huang stated, "A new computing era has begun. Companies worldwide are transitioning from general-purpose computing to accelerated computing and generative AI."
"Global cloud providers are intensifying AI investments, and the performance boost from the computing sector will be reflected quarter by quarter. The computing market is becoming more refined, with emerging technologies and business models evolving rapidly, potentially giving rise to 'dark horse' players," noted Guosheng Securities. Kaiyuan Securities analyst Fang Guangzhao also believes that large model training and applications may drive demand for GPUs, AI servers, computing leasing, and IDCs, bringing incremental earnings to companies involved in computing-related businesses.
Abundant computing power is the prerequisite for the successful commercialization of AI large models. Many institutions believe that with the rapid development of AI applications, the localization of AI computing power will also accelerate. "Considering factors such as comprehensive costs, cost-effectiveness, technological iteration, and product support, it is highly likely that domestic manufacturers will increasingly embrace domestic computing power chips over foreign ones in the future," said Fang Wenqian, an analyst at Huajin Securities. Domestic manufacturers are rapidly catching up in hardware performance and software ecosystems, and domestic computing power chip manufacturers will continue to benefit. He further pointed out that domestic manufacturers are accelerating the upgrade and iteration of AI chips and have already achieved commercialization in various application fields such as telecommunications, internet, finance, and energy.
With the continuous development of new technologies like AI, the market's demand for computing power is enormous. Against the backdrop of supply shortages, the computing power market has entered a strong "price increase" mode. Many institutions believe that due to factors such as the accelerated iteration of AI applications, the demand for computing power will continue to grow, potentially leading to a medium-to-long-term price surge. The performance boost from the computing power sector will be reflected quarter by quarter, and the localization of AI computing power will also accelerate.
China's computing power industry is massive, reaching a scale of trillions of yuan. According to calculations by the China Academy of Information and Communications Technology, the core industry scale of computing power in China reached 1.8 trillion yuan in 2022. For every 1 yuan invested in computing power, it drives 3 to 4 yuan of GDP economic growth.
With strong policy support, the construction of computing power infrastructure is advancing steadily, and development momentum continues to strengthen. Currently, China has become the world's second-largest computing power country, with a total computing power scale of 180 EFLOPS and a total data center rack scale exceeding 6.5 million, with an annual growth rate of nearly 30%. At the same time, the layout of computing power is continuously optimized, and industrial innovation capabilities are steadily enhanced. The proportion of computing power facilities in central and western China has increased to 39%, and high-performance chips for large model training and inference are undergoing continuous iteration and upgrades. The competitiveness of domestic brand servers continues to grow, with production volumes of servers, computers, smartphones, and other computing products ranking first globally.
With the rapid development of AI this year, tech giants have sparked a "hundred-model battle." Data shows that as of October this year, 238 AI large models have been released domestically.
Many institutions believe that due to factors such as the accelerated iteration of AI applications, the demand for computing power will continue to grow, potentially leading to a medium-to-long-term price surge. The performance boost from the computing power sector will be reflected quarter by quarter, and the localization of AI computing power will also accelerate.
Recently, the Ministry of Industry and Information Technology (MIIT) and five other departments jointly issued the "Action Plan for High-Quality Development of Computing Power Infrastructure" (referred to as the "Action Plan"), proposing quantitative development targets for the next three years in four aspects: computing power, transmission capacity, storage capacity, and application empowerment. Industry experts stated that the release of the Action Plan is expected to accelerate the development of China's computing power infrastructure, boost industry prosperity, and create new growth opportunities for the entire computing power industry chain.
Computing Power as the New Productivity in the Digital Economy Era
As the new productivity of the digital economy era, computing power has become a crucial support for high-quality economic and social development. China places great emphasis on computing power infrastructure construction, with multiple policies and plans guiding the industry's development. The MIIT has issued several policy documents, including the "14th Five-Year Plan for the Information and Communication Industry" and the "Three-Year Action Plan for New Data Center Development (2021-2023)," actively promoting the rapid and healthy growth of the computing power industry. Meanwhile, various stakeholders in the industry have collaborated to advance the construction and application of computing power infrastructure through multiple measures, striving to build a high-quality computing power supply system and continuously enhance the comprehensive capabilities of computing power infrastructure, achieving positive results.
Projected Global Computing Power Scale Analysis for the Next Five Years
According to calculations by the China Academy of Information and Communications Technology (CAICT), the total global computing power of computing devices reached 906 EFlops in 2022, with a growth rate of 47%. It is projected that the global computing power scale will grow at a rate of over 50% in the next five years, exceeding 3 ZFlops by 2025 and surpassing 20 ZFlops by 2030.
Huaxin Securities Research Report stated that the opportunities presented by the "East Data, West Computing" project, combined with the wave of large models, will sustain the growth momentum of domestic computing power demand. Therefore, driven by the dual growth of the underlying computing power layer and the top application layer, the transformative opportunities brought by AI large models to the field of artificial intelligence remain highly promising.