The US-China AI race isn’t just about tech dominance, but control over information
AI supremacy grants immense influence over information – what is shared, how it's framed, and what is left unsaid.
The US has long dominated Big Tech, reinforcing the expectation that AI supremacy would soon follow, driven by industry giants like NVIDIA and OpenAI. But with China’s DeepSeek entering the scene in January, that assumption has been challenged.
The race for AI supremacy is now wide open. DeepSeek has exposed how fragile the US lead really is and how much is at stake: this is not just about tech dominance, but control over information itself.
ChatGPT cost around $100 million to develop. In contrast, DeepSeek was reportedly developed at a far lower cost of just $6 million. While this could be a marketing tactic, the two models fundamentally differ in efficiency. ChatGPT processes all responses before choosing one whereas DeepSeek uses smaller modules on demand, reducing hardware and energy use. This efficiency has had a major ripple effect, even contributing to a record-breaking drop in NVIDIA’s stock due to reduced demand for its high-performance AI chips.
In terms of user experience, DeepSeek is almost identical to ChatGPT, right down to the interface. The key distinction? DeepSeek is completely free, whereas access to ChatGPT’s highest-quality responses requires a subscription to ChatGPT Plus or Pro. Despite attempts to transition ChatGPT into a for-profit model, efforts Elon Musk has attempted to block, DeepSeek’s free access strategy is driving significant price pressure in the AI market. This shift brings AI closer to becoming a universal commodity, much like electricity. This means that the primary driver behind AI development is no longer profit, but rather something far more significant: power.
Whoever leads the AI race gains a substantial strategic advantage across critical sectors such as healthcare, warfare, governance, and climate management – fields where AI plays an essential role. But more importantly, AI dominance also grants immense influence over information – what is shared, how it is framed, and what is left unsaid.
ChatGPT generally aims to offer balanced responses to user queries, while DeepSeek shows a subtle but crucial difference. When asked about both the positive and negative aspects of China, DeepSeek provided an answer for the positives but, if asked "what are the negative aspects of China?", responds with: “Sorry, that’s beyond my current scope. Let’s talk about something else.” There are reports of this happening for a range of China-related questions, such as ones related to Tiananmen Square and the Umbrella Revolution.
As people increasingly rely on AI as their primary source of information, those who control AI effectively control public perception. In the hands of entities that do not uphold democratic values, this power could shape reality itself, with profound consequences for global discourse and freedom of speech.
While the US government plans to invest $500 billion in AI infrastructure, DeepSeek’s cost-effective emergence questions whether such spending is even necessary. Success may depend more on strategy than sheer investment. Meanwhile, China, with $150 billion committed, is pushing to lead global AI by 2030, starting with the support of DeepSeek.
It is in this sense that DeepSeek could represent a true Sputnik moment for the US. While the outcome is still uncertain, one thing is clear: the US is facing political choices that could undermine its own AI ambitions. Chief among them is the AI export control policy, enacted in the final days of Biden’s presidency, which bans advanced chip exports to China. Intended to slow China down, the move is instead accelerating Beijing’s push for self-sufficiency. With Huawei developing its own chips, China could gain a major edge if it reaches full AI independence.
AI development heavily depends on human-created content for training models, making creative industries an essential yet contentious part of this race. In recent weeks, protests from artists pushing back against AI’s use of copyrighted content have been growing.
In the US, Microsoft and OpenAI are currently facing lawsuits for copyright infringement. While these legal actions protect intellectual property rights – a cornerstone of democratic societies – they also slow down AI advancements. China faces fewer barriers, thanks to laws that suppress dissent and prioritise state-led tech progress over creators’ rights. This gives China a significant advantage as it can more freely amass data and content for AI training. If the US wants to maintain its competitive edge, one possible solution could be the creation of a regulated market that connects AI companies with artists.
It remains to be seen whether China’s centralised approach or the US’s diversified model will win the AI race. But we may soon see a world divided between two distinct AI ecosystems: one dominated by American technology and the other by China’s.
This polarisation would redefine global cooperation and competition, leaving other regions, particularly Europe, in a precarious position. As the old saying goes: the US innovates, China replicates, and Europe regulates. But in an AI-driven world, where technological dominance shapes global power structures, is regulation alone enough to keep Europe relevant?
DeepSeek may not have broken through in the West – yet. But it has already cracked open the illusion that this is a US-only game. The AI race is no longer just about who builds the best chatbot. It is about who shapes the infrastructure of thought, governance, and daily life.
Lynn Dascha Engelhardt is a Policy Fellow at the Pinsker Centre, a campus-based foreign policy think tank, which facilitates discussions on international affairs and free speech. The views in this article are the author’s own.