For Silicon Valley’s tech giants, these were times of optimism and triumph. The start of the new U.S. administration seemed tailor-made for them: an inauguration ceremony attended by the sector’s top CEOs, followed by the announcement of a multibillion-dollar investment to boost the development of artificial intelligence. Additionally, an ambitious strategy was unveiled, even compared to the Manhattan Project, to address AI with unprecedented urgency and focus. The race to dominate this technology was underway, and the rhetoric could not have been more grandiose.
However, the euphoria was short-lived. The emergence of the DeepSeek R1 artificial intelligence model, developed by a Chinese company, has shaken the foundations of the market. DeepSeek claims that its model can match or surpass the capabilities of ChatGPT but at a significantly lower development cost. Most surprisingly, the company attributes part of its success to restrictions imposed by the United States: the blockade on exporting advanced chips to China forced them to innovate using Nvidia’s H800 chips, which are less powerful than the H100s, whose access was prohibited. Nevertheless, some suggest that DeepSeek may have accessed 50,000 H100 GPUs through intermediaries, adding an ironic twist to the story.
The impact of DeepSeek has been immediate and profound. The markets reacted with volatility, and Nvidia, until then the big winner of the AI era, saw its market capitalization plummet by $589 billion in a single day. While some analysts celebrate the arrival of this new competitor, arguing that it will democratize access to AI and expand its applications, others express concern. OpenAI, one of the main players affected, accuses DeepSeek of using data generated by its systems to train its model. Interestingly, this accusation comes at a time when OpenAI is facing its own controversies over the use of data protected by copyright, as The New York Times has pointed out.
The competition for supremacy in artificial intelligence has been compared to an “arms race” reminiscent of the Cold War, with the United States and China vying for global dominance. However, this rivalry carries significant risks. Experts like Alvin Wang Grayling and Paul Triolo warn that AI developed for military purposes could threaten global stability. In contrast to this scenario, an alternative emerges: international cooperation to ensure responsible and ethical advancement of the technology, aimed at the common good. Although this idea is inspiring, it starkly contrasts with the current reality, marked by rivalry and distrust.
In short, the emergence of DeepSeek has not only disrupted the market’s balance but has also highlighted the tensions and challenges that define this new era of artificial intelligence. The question that remains is whether humanity will be able to prioritize collaboration over competition or if it will succumb to the risks of an uncontrolled race.
By: Nestor Castillo, ForAllTechNews Director