AI becomes the latest flashpoint in the US-China rivalry

If there is one phrase that profoundly defines America’s insatiable desire for global dominance, it is this statement by former Secretary of State Tony Blinken — ‘If you’re not at the table, you are on the menu.’ And recent actions lend credibility to the assumption that, if the US can’t be at the starting line, it would rather see no race at all.

This mindset has been central to Washington’s approach to containing Beijing’s rise, but it has failed before and is unlikely to produce the desired results moving forward. The latest misstep is a bipartisan proposal in the US House to ban the Chinese AI app DeepSeek from federal devices, an extension of the existing TikTok ban that turns the tech world into yet another battleground in the broader power struggle between the two rivals. But history suggests such efforts do little to halt China’s rise. Instead, they expose Washington’s waning influence, often with striking ease.

At this point, without a doubt, DeepSeek, a startup based in Hangzhou, has emerged as a formidable contender in the Artificial Intelligence race, threatening to reshape the global tech landscape. The company’s latest models, including DeepSeek-V3, are not only comparable to, but in some cases exceed, the performance of the industry-leading US models, all while being significantly more cost-effective. According to claims by DeepSeek, training its model required less than $6 million worth of Nvidia H800 chips—an amount far lower than what is typically needed by American firms. Furthermore, the company’s AI Assistant has even surpassed ChatGPT to become the top-rated free application on Apple’s App Store. This has not only arrived as a bolt from the blue for the US tech industry but has also raised serious questions, as major corporations continue to pledge billions in AI investments while shares of key players, including Nvidia, have suffered since the rise of China’s DeepSeek—a significantly cheaper entrant in the tech arena.

A wide range of experts have described the emergence of DeepSeek as a shift in the AI landscape, particularly within the context of the US-China rivalry. When OpenAI’s ChatGPT was released in late 2022, it sparked a rush among Chinese tech firms to develop their own AI models, with Baidu’s first attempt falling short of expectations. However, DeepSeek has flipped that narrative – at least for now. Its models, DeepSeek-V3 and DeepSeek-R1, have been hailed as direct contenders to OpenAI and Meta’s most recent offerings. Not only do they match US models in performance, but also DeepSeek-R1 is reportedly 20 to 50 times cheaper to use than OpenAI’s latest iteration, depending on the task. Its open-source approach has given the Chinese firm an edge over the closed, high-cost systems favoured by US tech giants—sending shockwaves through Silicon Valley and challenging its walled-garden model.

So, with US lawmakers pouncing to respond with the No DeepSeek on Government Devices Act, effectively barring federal employees from using the model—with the most common justification being national security—the reality begs to differ. It appears to be a familiar effort to undermine a rising Chinese tech competitor, reminiscent of Washington’s campaign against Huawei, which saw its 5G dominance curtailed under similar pretexts. But even in Huawei’s case, which has striking similarities to the reaction from the US against DeepSeek, the company has seen a 22% increase in revenue from 2023, the fastest growth since a 32% increase in 2016, according to CNBC calculations of publicly released figures. The gains came despite US restrictions since 2019 on Huawei’s ability to access tech from American suppliers, from advanced 5G chips to Google’s Android operating system.

At this point, there is no ambiguity about the fact that the US has long sought to limit China’s technological capabilities, implementing strict export controls on semiconductors, including those designed for AI. These efforts were intended to slow Beijing’s progress in artificial intelligence, yet DeepSeek’s rise proves that China continues to innovate despite constraints designed to curb its overall ascent. So far, Beijing’s response has been clear—rather than succumbing to Washington’s pressure, its researchers have doubled down, proving that exceptional hardware is not necessary for technological breakthroughs. While Silicon Valley depends on cutting-edge chips and massive data centres, DeepSeek has demonstrated that innovation can thrive even under tight resource limitations.

The shift has rattled the US tech industry. In a rare admission, CEO Sam Altman conceded that OpenAI ‘is on the wrong side of history’ when it comes to open-source AI, signalling a potential shift in strategy. This has sent shockwaves through the sector, sparking internal debates over whether closed-source AI can remain the dominant paradigm. On many fronts, China has already surpassed the US in high-tech sectors beyond AI. Its companies have surged ahead in industries like electric vehicles (EVs), green technology, and renewable energy. As a result, the world is witnessing a gradual but undeniable shift from US tech dominance to Chinese leadership, driven by long-term investments and state-backed research initiatives. Given China’s momentum, it’s hard to imagine it stepping aside in the future, especially in the realm of innovation and emerging technologies.

Apart from changing the tech landscape, DeepSeek has further exposed the intense cartelisation of Silicon Valley that has long been an open secret—an ecosystem where a handful of companies control access to computing resources, data, and distribution channels. The industry’s dominant players rely on planned obsolescence, where models are continuously updated and locked behind paywalls to extract maximum revenue. DeepSeek, according to experts, fundamentally challenges this approach, proving that AI can be developed through collaborative, decentralised methods rather than top-down corporate control.

China’s advantage

Considering the bevy of restrictions, China’s ability to develop world-class AI models without access to top-tier Western hardware highlights its broader strategic advantage. While Washington’s stranglehold on chips and hardware was designed to slow Chinese AI development, Beijing has outmanoeuvred Washington by prioritising efficiency and alternative pathways. The broader implications of this shift are undeniable. As China continues to lead in AI, EVs, and renewable energy, the balance of power in global technology is visibly tilting in its favour. The US can no longer rely on restrictions alone to maintain dominance—it needs to actively innovate or risk losing its position entirely. Any potential ban or restrictions on DeepSeek are unlikely to stop its progress. Just as Huawei continued its expansion despite restrictions, DeepSeek will evolve through alternative pathways, refining its models and expanding its reach.

Silicon Valley now faces a rapidly shifting landscape, where open-source collaboration and cost-efficient models threaten the dominance of high-cost, proprietary systems. The future of AI doesn’t depend on whether DeepSeek is a security threat, but on whether the US can adapt. Will Washington continue restricting competition, or will it innovate to remain competitive? The era of US tech hegemony is rapidly receding, and if Silicon Valley fails to adjust, China will expand its dominance in a space once firmly controlled by the US.