The availability of DeepSeek, the Chinese Large Language Model, on the market under an open source intellectual property regime – that is, without proprietary claims in terms of copyright and patents – has shaken up the AI sector and challenged some of the dogmas that have become prevalent in the sector. by Andrea Monti – Initially published in Italian by Formiche.net
DeepSeek and the AI narrative
The first dogma concerns the cost and time required to produce an LLM. If it is true that the development of DeepSeek was not supported by Beijing in terms of privileged access to the hardware and energy needed to achieve the result, then it is no longer true that billion-dollar investments are necessary to compete in the market. This would make it possible to increase the number of potential players in the sector and create a genuine system of international competition.
The second belief challenged by DeepSeek is that the AI sector is firmly in the hands of the US, which controls both software and hardware technologies, from GPUs to the technologies needed to build ever-thinner chips and ever-more powerful processors. Aside from the problems of circumventing the ban on exports of GPUs and semiconductor manufacturing equipment to China, the technology embargo imposed by Washington has forced China to develop alternative approaches to theoretical research and the production of AI hardware, such as that produced by Huawei.
The impact on the stock market
The public availability of DeepSeek, in the form of an app to be downloaded on smartphones and as a platform, had an impact on the financial market, damaging the market value of nVIDIA, the almost monopolistic producer of GPUs and AI software development environments. The fluctuation didn’t last long, and nVIDIA’s share price recovered almost immediately, but it was a clear sign of what could happen in the stock market sector, where volatility in value is strongly influenced by the dissemination of information, or rather the way in which investors perceive the information that is disseminated.
The impact of AI on industry
If the news of DeepSeek’s greater affordability has affected the stock market, the Chinese start-up’s decision to release the model as ‘open source’, i.e. allowing anyone to use it without demanding payment of fees or royalties, has affected the real market. Perplexity, one of the most active US companies in the AI sector, recently announced that it will use DeepSeek to provide some services. It is true that everything ‘runs’ on American systems, no data is sent to China and no one except Perplexity has access to the model. But that doesn’t change the fact that a company has been able to improve its services without having to pay royalties to competitors developing similar models. If other companies follow Perplexity’s lead, the big tech companies in the sector will inevitably face home-grown competition that will be able to take market share and disrupt the timetable for public release of new technologies.
Is DeepSeek the new TikTok?
Taken on its own, the spread of DeepSeek in the West could be seen as part of the ‘normal’ competitive dialectic between companies in the sector; but given the strategic nature of AI, it’s clear that this interpretation would be very simplistic. If, after the first technical tests, the mechanism for the proliferation of Chinese LLMs were to become fully operational, we would be faced – whether we like it or not – with something that closely resembles an act of economic and technological warfare.
However, the announced introduction of legislation to ban the use of DeepSeek on government devices is not based on these concerns, but on the fear that the app installed on smartphones and tablets could provide information about users to the Chinese government.
While the two cases are not entirely comparable, it is clear that the request to ban the application is based on the same assumptions that led to the forced sale of TikTok.
The key difference between the two cases, however, is that in the case of DeepSeek, the need to protect national security is invoked in an essentially pre-emptive way. In other words, whereas the US took action against TikTok after the application had been in use for years, DeepSeek appears to be a case of applying the precautionary principle and banning its use before any critical issues have arisen. It should also be noted that, as clearly stated in the rulings of the courts dealing with the case, the reasons for the forced sale of TikTok were based on government statements that the judges did not consider challenging, while in the case of DeepSeek there are no similar positions.
The new national security paradigm
It is not yet known whether DeepSeek will be banned, or to what extent, but the mere fact that the hypothesis has been raised is in line with similar decisions already taken by the US government in relation to Huawei, DJI and TikTok.
In fact, all of these decisions have in common the tendency to consider the control of a technology by a foreign state as a possible threat to domestic survival, regardless of the material use of a product or service that technology provides.
The irreconcilable opposition between security and global techno-economic development
If, in the context of national security, it is better to be safe than sorry, the application of precautionary measures based on risk hypotheses rather than on verifiable evidence raises questions not only about international law and competition, but also about the future of global technological governance.
If the precautionary principle becomes a tool for systematically excluding innovation from certain countries, there is a risk of an increasingly fragmented technology sector, with the creation of separate and incompatible digital ecosystems. Such a scenario would not only hinder scientific progress and international cooperation, but could also prove counterproductive for US companies themselves, who would lose access to innovative models and solutions developed outside their borders.