Global Outlook: Asia #1
The launch of the DeepSeek R1 chatbot in early 2025 prompted a flurry of responses. In Asia, it touched off discussions about gaining ground against big tech and national AI strategies.
Asian responses to DeepSeek
The launch of the DeepSeek R1 chatbot in early 2025 prompted a flurry of responses. In Asia, it touched off discussions about gaining ground against big tech and national AI strategies. The reactions offer a window onto thinking in the region about technology policy, US-China rivalry, and international cooperation.
US-China rivalry and national security
The story of DeepSeek as a salvo in Sino-American technology competition is almost too easy. State media and ordinary people in China presented the accomplishment as a source of national pride, with some suggesting DeepSeek was a Chinese victory over US restrictions and competition. US President Donald Trump cited DeepSeek’s launch as a lesson that American firms must work harder.
Within China, state-owned enterprises, other large firms, and government agencies have picked up DeepSeek for internal use. Chinese firms are integrating DeepSeek into consumer products and services, such as mobile phones and smart home appliances. At the same time, DeepSeek is not the only domestic option for Chinese users.
On the surface, reactions in Asia to DeepSeek seemed to line up with national positions in US-China rivalry. Taiwan took immediate action on national security grounds. Personal data going through Chinese data centres was one problem for Taipei; another was the concern that DeepSeek could propagate false narratives over cross-Strait relations. In South Korea, a close US ally, a number of prohibitions from firms and government agencies came right away. Governments that maintain friendlier ties with Beijing, such as Indonesia and Cambodia, issued no warnings. In Vietnam, the chatbot has seen mass downloads. Singapore, with close security ties to the United States but growing links to China, displayed ambivalence.
A deeper look suggests these patterns do not reflect the biggest story. Rather than pressing countries to choose sides between the United States and China, DeepSeek’s launch evoked responses linked to the local – not global – concerns of governments and businesses in Asia. In particular, it fed into struggles over “big tech” and over how to build competitive technologies.
A new hope to challenge big tech
The DeepSeek R1 launch shattered the received wisdom of the past two years: you need not be home to the world’s tech giants to build a powerful AI chatbot. ChatGPT’s emergence had fueled the view deep pockets and access to top-of-the-line chips were prerequisites. That ruled out opportunities for firms in much of Asia – or anywhere else for that matter. DeepSeek changed that story. If a firm in China, on a relatively modest budget, could develop this technology, then maybe our prospects in AI are bright. This was an empowering lesson.
An AI industry figure in Vietnam stressed that DeepSeek’s disruptive example suggested “a more equitable AI landscape,” both in the sense of being less expensive to use and giving hope that others could develop their own generative AI tools. In Indonesia, the National Economic Council found inspiration in the support the Chinese government gave to promising individuals to study abroad, return home, and start technology firms. Days after DeepSeek’s launch, Indonesian officials with the Ministry of Communication and Digital Affairs called for measures to develop the same hospitable environment. DeepSeek’s example offered a lesson.
The point in highlighting in these responses is not that governmental and private actors were happy about or unsuspicious of Chinese technology, but that the focus was rather on possibilities for building their own AI industries.
Regulation and national industry promotion
Reactions to DeepSeek have fed into longer-term battles over the regulation of big tech and efforts to strengthen national technologies. South Korea’s response gives an example.
In Seoul, the reaction to DeepSeek came from the agency tasked with protecting private information. On 31 January 2025, days after the DeepSeek R1 release, South Korea’s data protection agency sent a formal enquiry to DeepSeek requesting further information. Soon after, the agency took action on the grounds that DeepSeek was obtaining private information without proper consent. DeepSeek was removed from app stores.
The backstory is that the legal basis for this response was established in a raft of legislation in 2019-21 aimed at regulating big tech. Policymakers were grappling with the same challenges about privacy, power, and market structures as counterparts around the world. GDPR was a reference point.
There is another dimension. Both the longer-term regulation of big tech and the response to DeepSeek involved concerns over how to promote domestic technology firms and strategies. South Korea, after all, has its own web and social media platforms. On 4 February this year – two weeks after DeepSeek’s launch – legislators held a hearing on “Responses to the ‘DeepSeek Shock’ and Strategy for Developing AI.” As the title indicates, the national technology competition angle loomed large.
DeepSeek later returned to app stores in South Korea, after the Chinese firm made changes. DeepSeek made adjustments regarding the data collected and where data would be stored. It also added options for users to release less information. While some South Korean government agencies, education institutions, and private firms single out DeepSeek as a prohibited tool, many organizations have no need for a DeepSeek-specific policy. SK, LG Electronics, and Samsung Electronics do not permit the use of any applications that store data externally; they also use in-house AI tools.
South Korea is fully entrenched in supply chains that serve US-based technology firms. At the same time, as a country with a strong tradition of industrial policy, another priority is to build support for technologies created and owned by domestic firms.
Policy discussions
Responses to DeepSeek turned swiftly to conversations over national AI support policies. The key policy questions relate to identifying a specific vision for a given country’s AI development and how to support the realization of that vision.
Take, for example, Vietnam. As an overseas-based Vietnamese technology researcher put it, Vietnam can both gain from large foreign technology firms and find ways to carve out space for local firms. He argues for investing in education, supporting talent, and encouraging those with skills to stay in Vietam or return. That is the lesson he took from DeepSeek. The goal should not be, he states, to create a “Vietnamese DeepSeek” but to determine which AI technologies can be developed for the local context.
Malaysia has attracted technology firms – including Nvidia and Google – to build data centres at scale. It helps that US export control rules have not prohibited Nvidia chips from moving to the country. In December 2024, Kuala Lumpur announced the establishment of a National AI Office. Then, if DeepSeek’s appearance means that massive data centres are less necessary, observers ask whether Malaysia’s investments might be at risk. Because of where Malaysia inserted itself in the AI supply, lower-cost technology could work against its approach. This example illustrates the uncertainty that countries face in designing AI strategies.
These discussions reflect a particular focus in the regional responses to DeepSeek. Outside the region, a point of interest in DeepSeek is its “openness”. DeepSeek has distinguished itself from chatbots developed by OpenAI or Meta, for example, by permitting users to copy and modify it. This feature has generated enthusiasm in some quarters, especially from those critical of the attachment of proprietary licenses on technologies. DeepSeek’s position as a quasi-open source model has garnered less attention in Asia. In this region, the critique of big tech is framed less around ideals of openness and technology, and more in terms of what space there is for new players to enter and compete.
Lessons
While the US and China might see themselves as competing via technology for influence in the region, the outcome of that competition is not the centre of attention for other Asian countries. Policymakers in the region are, like their counterparts in Europe, trying to figure out how best to regulate technology companies, to identify specific ways to expand domestic AI capabilities, and to cultivate their own firms. International cooperation, not isolation, is recognized as essential. The features of this international cooperation remain to be defined.
/Hyejin Kim and Erik Mobrand for the Royal Swedish Academy of Engineering Sciences (IVA)
Hyejin Kim is a researcher in the Department of Economic History, Lund University School of Economics and Management, and an Associated Senior Lecturer in Political Science Department and Global Studies Programme, National University of Singapore. The author of five books, her research addresses education, globalization, and science policy.
About Global Outlook: Asia
Technological development and innovation in Asia are extensive and ultra-fast, and impact the entire world. To gain better insights, IVA now offers quarterly thematic status reports from Asia, authored by expert researchers with extensive regional experience, under the title 'Global Outlook: Asia'.
