How accelerating AI is the foundation for industry intelligence in Asia-Pacific

Like technologies, such as steam engines, computers and the Internet in history, digital technology has changed human society at an unprecedented speed and scope over the past 20 years. It has greatly promoted socioeconomic creativity and growth.

Facing these transformative opportunities, the Asia-Pacific region has entered the golden era of the digital economy. In 2022, the GDP growth rate in many Asian countries was over 5%. This region enjoys a great population and resource advantage and the rise of the younger generation has brought new potential to its emerging economies. By 2030, ASEAN is expected to become the world’s fourth-largest economy and the regional internet economy market should reach $1 trillion.

Nearly all Asia-Pacific countries are taking proactive steps to drive a national digital masterplan. Indonesia, for example, pays special attention to new infrastructure in its Vision 2045. The Smart Bangladesh Blueprint and the Korean Digital New Deal have shown a strong investment commitment to digital government. And, Thailand has released its thirteenth Five-Year plan to increase the proportion of its digital economy in GDP from 12% to 30% by 2027.

In this favourable policy environment, various industries are utilizing ICT technologies to improve efficiency sustainably. In a digital government, traditional services, such as tax, customs and law enforcement, are efficiently processed on online platforms. In airports and highways, the application of 5G, cloud and AI optimizes travel experiences and relieves traffic congestion. Some universities are also building ‘smart campuses’ and bringing digital resources to more young people.

At the same time, digitalization and de-carbonization have become two mutually strengthening forces. For example, the large-scale deployment of rooftop photovoltaics in public hospitals and energy storage facilities represents a sustainable future that is enabled by technology.

As the Asia-Pacific region moves from the digital era to the intelligent era, computing power has become the decisive factor for AI innovation and the core productivity of the digital economy.

Computing power is the foundation of the intelligent era and its most critical infrastructure is the data centre. Specifically, the computing power scale determines productivity and the computing power industry leads to competitiveness, while computing power applications become the driving force.

According to an IDC survey of 15 countries, a one-point increase in the computing power index will lead to a national digital economy growth of 3.6‰. Take Indonesia as an example, its data centre market will grow at an annual growth rate of 18% from 2023 to 2029. In China, the data centre scale has achieved an average annual growth rate of nearly 30% in the past five years.

As we can foresee, as foundation model training deepens, AI development crosses the inflexion and singularity points. It is evolving far beyond being used only for entertainment purposes and is being integrated into the industry and production system. The resulting super-large-scale AI models and massive data have made computing power one of the most scarce and strongest resources.

The development of computing power drives AI to upgrade to the third generation. The knowledge-driven first-generation AI uses knowledge, algorithms and computing power to build AI. The data-driven second-generation AI adds data as the fourth dimension. And the third generation of AI is the convergence of the previous two and facilitates to secure, trusted, reliable and scalable AI innovation.

With the release of ChatGPT and the large language model, Asian governments are accelerating the deployment of AI. For example, Hong Kong plans to invest $3.8 billion to develop large language models for healthcare, law and finance. In addition to investment, governments are working more closely with stakeholders to promote AI governance. An open, collaborative and sustainable governance platform is needed to achieve international consensus.

Moving from a small model to a large industry model, the main theme of AI development will be accelerating industry intelligence. A good example is the Thailand Meteorological Department’s exploration of the Huawei Pangu-Weather model, an advanced AI model for weather forecasting. According to a Nature paper, this Pangu model is the first AI prediction model to demonstrate higher precision than the traditional numerical forecast method, allowing a 10,000x improvement in prediction speed and reducing prediction time to just seconds.

Computing infrastructure itself has become one of the major sources of carbon emissions. For example, Singapore has more than 60 data centres, accounting for 7% of the country’s power consumption. Therefore, its movements toward green computing are particularly important. Traditionally, the data centre computing architecture focuses on single-node computing power. Driven by the low-carbon trend, the new date centre computing architecture will develop towards computing power diversity, efficient cooling systems and green energy storage devices for low power supply efficiency.

According to a McKinsey report, in China only, the AI talent gap is expected to reach 4 million by 2030. Behind this number is an urgent requirement to upgrade the current workforce and ecosystem. To partially address this gap, Huawei has announced that it could work with Thailand’s ministries, startup associations and computing schools to develop 20,000 local cloud developers in the next three years.

The new era of AI represented by ChatGPT and industry large industry models has arrived. We can boldly predict that all personal and industrial applications will be rewritten by this new AI as a core productive force.

For Asia Pacific to achieve fast-tracked growth, a comprehensive framework with four key factors is crucial. First, a favourable policy environment with national investments and a governance framework in AI infrastructure. Second, continuous innovation and R&D to make large language models at all levels easier and faster to deploy. Third, open and scalable industry applications. And, finally, an upskilled talent ecosystem that bridges government, industry and academia.