In January 2025, the Chinese startup Deepseek, founded by Liang Wenfeng and based in Hangzhou, released its first AI assistant, Deepseek-R1 —an innovation that would soon influence the geopolitics of Generative AI.
This marked a critical shift in the Geopolitics of Generative AI (GenAI), where technological breakthroughs are now entwined with national strategic ambitions.
This release not only stirred controversy in the tech domain but also had a strategic impact. The AI assistant was not merely an addition to existing Generative AI (GenAI) artefacts; it challenged established players like US-based OpenAI, Google, Microsoft, etc..
This marked a critical shift in the geopolitics of Generative AI, where technological breakthroughs are now entwined with national strategic ambitions.
Challenging the U.S. GenAI Supremacy
Thus, it is confronting the US prowess in the realm of GenAI.
Openai released the first GPT model in September 2022. Subsequently, other tech giants like Google, Microsoft, etc., participated in GenAI development, paving the way for US dominance in this domain.
Power-Hungry Language Models
Large language models are the building blocks of the GenAI revolution. These models are built on extensive data, so developing GenAI assistants requires enormous computing power.
It is perceived that around 10,000 NVIDIA Graphic Processing Units (GPUs) were used for the first version of ChatGPT. These GPUs are mainly responsible for performing complex tasks related to neural networks on a massive amount of data.
OpenAI vs. Deepseek
The latest Openai’s Chatgpt-4 version was trained using over 20,000 NVIDIA GPUS. Similarly, AI models from Google and Microsoft were created using vast processing power, which requires enormous resources.
Thus, the computing hardware industry is rising. NVIDIA proved to be the primary beneficiary due to its efficient hardware, which is perfectly aligned with GenAI’s core technologies of neural networks.
Deepseek’s Disruptive Efficiency
Contrary to this enormous spending by US companies, Deepseek came forward with an economised version of an AI assistant that used fewer computing resources and efficient data processing with improved algorithms.
It is estimated that only 2000 GPUS were used to develop DeepSeek-R1. Interestingly, the R1 model was released on January 20, 2025, the day of President Trump’s second inauguration.
DeepSeek has cemented its reputation as the top frontier AI research lab in China and caused assumptions about the landscape of global AI competition to be reassessed.
By January 27, DeepSeek’s iPhone app had overtaken Openai’s Chatgpt as the most downloaded free app on Apple’s U.S. App Store.
The stock prices of some U.S. tech companies tumbled, including the AI chip designer Nvidia, which lost more than $600 billion off its valuation in a single day.
Resultantly, investor interest in Chinese tech companies has grown significantly.
The Deep Learning Advantage
Initially, the Generative AI realm was dominated by the US. The reason for this dominance was advancements in deep learning, made mainly by US companies.
Deep learning is a subdomain of Artificial intelligence. These advancements resulted from enormous computing power and capabilities to process vast amounts of data.
A significant achievement in 2017 was when the scientists at Google presented their seminal work on transformer architecture, known as the attention mechanism.
It’s a sequence-to-sequence model capable of converting sequences from one domain into sequences in another. For example, translating French sentences to English sentences.
The original transformer architecture consists of an encoder and a decoder. The encoder converts the input text (e.g., a French sentence) into a representation, which is then passed to the decoder.
The decoder uses this representation to generate the output text (e.g., an English translation) autoregressively.
GenAI: From Research Labs to Everyday Life
GenAI is also known as democratic AI due to its acceptance and usage by the common man.
Unlike traditional AI, which is mainly used in research labs, GenAI leverages an ordinary person with creativity and generates synthetic data.
Further advancements in GenAI are transforming it into more impactful Agentic AI, which is bound to influence and disrupt traditional working methods.
In his recent interview, Bill Gates, the founder of Microsoft, predicted that within the next ten years, many professions, including teaching, healthcare and hospitality, would be taken over by AI agents.
These AI agents would be more efficient and economically cost-effective.
GenAI as a New Frontier of Geopolitical Competition
China’s entry into GenAI is based on the international relations theory of competition between great powers. This emergence underscores how the geopolitics of Generative AI is now shaping national policies, economic security, and global influence.
Scholars of international relations suggest that great powers compete to maximize their influence and elevate their position in the global order.
Post-Cold War, a unipolar world order prevailed for two decades, ending with China’s peaceful rise as a global competitor to challenge the US-led unipolar world order.
The prevalent world order is not merely competition in the military domain; great powers compete in economic, digital, and climate change domains to project their influence and maximize their gains.
China faced chip export control restrictions from successive US regimes. These restrictions were aimed at curtailing its capability to compete in the emerging domain of Artificial intelligence.
However, China has progressed remarkably.
In February this year, at a meeting between Chinese Communist Party (CCP) Chairman Xi Jinping and Chinese technology executives (including DeepSeek CEO Liang Wenfeng), Huawei founder Ren Zhengfei told Xi they are ready to achieve self-sufficiency in the semiconductor value chain. T
his means that competition between great powers in the AI domain will likely be accentuated.
Final Thoughts
Progress in the AI domain is a human pinnacle: training the machine for its benefit and sustenance. It is envisaged that the great powers’ competition in the AI domain will positively impact human society as artefacts like Deepseek continue to emerge.
Wg Cdr Jamal Abdul Nasir (R) is an IT professional with more than 25 years of experience in software engineering, cyber security and AI He is also a graduate of the National Defence University. His areas of interest include Defence & strategic studies, with special linkage to technological advancements.
