The emergence and continuous evolution of large-scale models are sparking an AI industry revolution worldwide. Countries are accelerating their AI strategies, and China's artificial intelligence industry development has been elevated to a national priority. The 2024 Government Work Report specifically emphasizes deepening research and application of big data, artificial intelligence, and launching the "AI+" initiative to create internationally competitive digital industry clusters.
Coincidentally, just before the Two Sessions, the State-owned Assets Supervision and Administration Commission of the State Council (SASAC) convened a special meeting on "AI Empowering Industrial Renewal" for central enterprises, deploying efforts to "accelerate the development of artificial intelligence." This clearly signals the entry of the "national team" into seizing the AI dividend. The meeting focused on how central enterprises can leverage their advantages and take proactive actions, emphasizing the importance of building AI industrial clusters and leading the charge in integrating AI into traditional industries.
While this isn't SASAC's first mention of artificial intelligence, it marks the first time a dedicated meeting has been held to discuss how to promote AI development within central enterprises, underscoring its significance. Amidst the rapidly evolving global landscape and the strategic opportunities presented by AI, particularly large-scale models, central enterprises have an obligation to embrace the profound changes brought about by AI and prioritize the development of next-generation AI.
Empowering businesses and building platforms, several central enterprises have already taken action, riding the wave of large-scale models to enhance operations and create platforms, setting examples and achieving initial success.
China Railway Group, leveraging Huawei Cloud's Railway Large-scale Model and TFDS system, upgraded manual train inspections to an "AI inspection" system, intelligently identifying over 380 types of key faults primarily related to freight trains. This has resulted in higher efficiency and reliability, with the system gradually being piloted from Zhengzhou North Vehicle Depot to various railway bureaus nationwide. State Grid Corporation utilized Baidu's large-scale model technology to predict electricity peak and off-peak periods, providing recommendations for grid scheduling and using intelligent inspections to reduce safety hazards, ensuring stable power supply. Shandong Energy Group, in collaboration with Huawei Cloud, released the Pan Gu Mining Large-scale Model and Best Practices White Paper, applying it to over 40 application scenarios across 9 specialties such as coal mining, excavation, transportation, and safety monitoring. This has led to continuous improvements in working conditions, reduction of safety incidents, and enhanced production quality.
In addition to integrating AI into business operations, central enterprises are exploring platform-building to empower industrial ecosystems. Ordos Innovation Investment Group and Huawei jointly developed the first industrial Internet platform centered around the Pan Gu large-scale model, featuring capabilities such as intelligent computing power, mining large-scale models, multiple pre-installed algorithms, and application development environments. This platform is open and inclusive to industry partners, addressing challenges such as high AI development thresholds and long cycles.
From individual initiatives to collective endeavors, central state-owned enterprises (SOEs) are ramping up their efforts in AI, focusing on three key areas.
Deep Integration of AI with Business Scenarios: This serves as the cornerstone for accelerating AI development within SOEs. Given their involvement in sectors such as manufacturing, electricity, petroleum, coal mining, and transportation, SOEs possess vast scale and diverse application scenarios, demanding a high level of industry-specific expertise in AI. It's evident that generic large-scale models have limited applicability, prompting a shift towards specialized models tailored to specific industries. Examples include Baidu's "Wenxin Yiyuan" and Alibaba Cloud's self-developed model "Tongyi Qianwen," which quickly gained traction across various sectors like e-commerce, healthcare, and finance.
Solidifying Computing Power Infrastructure: Adequate computing power forms the bedrock of AI development within SOEs. With the escalating demand for resources related to training large models, ensuring a steady supply of computational resources becomes paramount. Recognizing this, the State-Owned Assets Supervision and Administration Commission (SASAC) emphasizes the need to expedite the establishment of intelligent computing centers and deepen cooperation to leverage cross-enterprise innovation platforms. This initiative not only addresses existing shortages but also presents new business opportunities within the domestic computing industry.
Upholding High-Quality Development with Security and Compliance: Ensuring secure and compliant AI deployment is critical for SOEs, given their pivotal role in national development. Embracing indigenous technological solutions for infrastructure development while safeguarding data integrity and adhering to regulatory requirements is imperative. Consequently, deploying large-scale models through hybrid cloud solutions becomes an inevitable choice for SOEs. Huawei Cloud Stack, as the industry's first large-scale model hybrid cloud, offers end-to-end innovation from hardware infrastructure to AI training platforms, catering to SOEs' needs for secure and compliant AI deployment.
By prioritizing the deep integration of AI with business operations, fortifying computing power infrastructure, and upholding stringent security and compliance standards, central SOEs are poised to leverage AI effectively, contributing to their sustainable growth and technological advancement.
The flourishing development of the artificial intelligence industry requires collaboration across the entire industry chain, including central and state-owned enterprises. No single entity can monopolize success; instead, all stakeholders must work together for mutual benefit. Central enterprises are opening up application scenarios, while technology companies provide technical empowerment, jointly building China's AI industry ecosystem to unleash AI productivity and reshape various industries. It's worth mentioning that during the recent State-owned Assets Supervision and Administration Commission (SASAC) conference, 10 central enterprises signed a pledge to actively open up AI application scenarios to society. This is a good start and an accumulation of quantitative changes. With the continuous participation of these powerful central enterprises, China's AI industry will embrace a new look and lead the way in this wave of AI transformation.
From March 14th to 15th, the "Huawei China Partner Conference 2024" with the theme "Born from Convergence, Empowered by Intelligence" will be held in Shenzhen. What releases and insights related to government-enterprise intelligence will Huawei bring to this conference? It's definitely something to look forward to.
