TECH WARS
BEYOND THE NUCLEAR ARMS RACE AND INTO AI
Sunera Bandara examines the race for AI global domination between the US and China – where espionage and strategic deals are shaping a new era
Historically, it has never been a wise decision to hand over technological secrets to one’s adversaries. At the height of the Cold War, when the US refused to share nuclear secrets with the USSR, the emerging communist superpower found another way to achieve its objective.
On 19 July 1953, two Americans within a wider ring of Soviet led espionage dedicated to uncovering, collecting and transmitting nuclear secrets to the USSR were executed. Julius and Ethel Rosenberg were dedicated communists who worked within the US industrial sphere and formed part of a network of spies working to access top-secret nuclear blueprints – knowledge that would eventually catapult the Soviet Union to superpower status.
Among them were spies such as Klaus Fuchs, who was embedded within the highly classified Manhattan Project – a programme that has recently garnered attention, following the critically acclaimed Christopher Nolan film Oppenheimer.
Fast forward to March 2024, when a Chinese national by the name of Linwei Ding was indicted by a United States federal grand jury. His alleged crime was stealing secrets associated with Google AI.
He was also found to be associated with artificial intelligence companies based in the People’s Republic of China. In August last year, two Chinese nationals were arrested for ‘knowingly and wilfully’ shipping tens of millions of dollars’ worth of NVIDIA H100 AI chips to China.
A month earlier, Chinese hackers had reportedly targeted the Taiwanese semiconductor industry amid US imposed restrictions on advanced chips being exported to China. Taiwan is critical to NVIDIA’s AI strategy and the NVIDIA H100 AI chips are at the peak of AI technology in the world market.
By the tail end of 2025, President Donald Trump had agreed to a deal with China involving NVIDIA’s second most powerful AI chips – the H200. Not only did it grant similar technology that China had attempted to steal for several years but also enabled Beijing to accelerate its position within the international artificial intelligence race.
Executives at DeepSeek AI have complained that their computers don’t have the power to keep pace. These same chatbots infamously refused to answer questions related to the Tiananmen Square massacre of 1989.
Access to H200 chips could provide Chinese AI firms the ability to improve and potentially outcompete American counterparts. One is reminded of the words of Vladimir Putin: “Whoever leads in AI will rule the world.”
China’s ambition of AI domination has been an initiative since the late 2010s. Billions have been poured into infrastructure and research. Calling it a pillar of national rejuvenation, the Asian superpower in its 2017 Next Generation Artificial Intelligence Development Plan set lofty goals – including AI global leadership by 2030.
Yet, despite this ambition and the massive investments, China has failed to secure global dominance – and it may have mismanaged its capital resources in attempting to do so.
Artificial intelligence in the Chinese context is deeply interwoven with state policy. SenseTime, a Hong Kong based Chinese AI Firm, received hundreds of millions of dollars in funding for AI ventures. Its valuation in 2020 was US$ 7.5 billion and deployed AI technologies across more than 150 cities. It was also the origin of China’s notorious nationwide facial recognition software.
However, SenseTime’s stock price has since plummeted following short-term gains. The company’s development of artificial intelligence was part of Chinese state policy and had reportedly been unable to keep pace with the radical advances in the field.
China’s funding did not directly translate into AI global leadership.
As the Atlas Institute for International Affairs observes, “unlike Silicon Valley, where legal frameworks provide predictable boundaries for intervention, China’s private AI firms operate in an environment where political priorities can override market dynamics at any moment.”
And the Council on Foreign Relations notes that the United States tends to define it in terms of the race towards Artificial General Intelligence (AGI) – self-improving AI which surpasses the cognitive power of human beings and is capable of executing real world knowledge work tasks.
Trump’s AI and cryptocurrency czar David Sacks warns that “China is not years and years behind us in AI. Maybe they’re three to six months.”
We are faced with the issue of two diametrically opposed worldviews pitted against each other on ground zero of AI.
If China succeeds and even outcompetes the US, one should wonder if the principle of the Thucydides Trap has finally been triggered. In 2023, the United States released a declaration lobbying for the international normalisation of AI within military contexts.
Meanwhile, Taiwan alleges that China uses AI for disinformation in a divide and conquer tactic within its borders – an especially significant point, due to Beijing’s ambitions to capture Taiwan.
One may attempt to visualise an ideal world shaped by Chinese AI leadership. Yet, the past is littered with examples of many totalitarian regimes. If history teaches us anything, it is this: technology and tyranny make for strange if not terrifying bedfellows.
