Personally, I was not happy with the drop in the stock market, because the history of this company and that of its founder, represent an inspiring example of success based on innovation, education and meritocracy. Huang did not start with a privileged situation; It was a humble Taiwanese immigrant who studied at the University of Oregon. However, his brilliance led him to the prestigious Stanford University, where he found the appropriate ecosystem to found Nvidia in 1993 and revolutionize the architecture of microprocessors. Its trajectory is a testimony of how talent and effort can transform entire industries.
However, there is good news in this stock market collapse for all of us. The value of stock stock shares is an indicator of the expectations of long -term benefits of companies, but not necessarily the well -being of citizens. Nvidia’s actions had skyrocketed because their vector microprocessors seemed to be the key asset for the unlimited artificial intelligence business (AI), dominated by a few giants like. However, the appearance has changed the panorama completely.
Until last week, the dominant narrative about the future of AI suggested that the logic of the “winner takes everything”, characteristic of the digital economy in recent decades. It was thought that future quasimono -political in AI would be protected by huge entry barriers, associated with the immense investments required to develop advanced language models (LLM), both in hardware specialized as in the expensive algorithm training process. Deepseek has shown that these barriers are much lower than it was believed, opening the possibility of a competitive market in AI and, with it, the hope of innovations more accessible to all.
The experience of using tools such as Chatgpt or Gemini used to be summed up in three emotions: fascination, curiosity for the potential of algorithms and, ultimately, fear. We anticipated that the companies behind these technologies, which have invested billions of dollars and still register losses, would seek to monetize the AI increase their costs as they became more essential. However, the competition that Deepseek represents (which, in addition, is open source) could be a shield against this dynamic. Their performance is close to the best algorithms, which could significantly reduce the market power of these companies and force them to redesign their business strategies. The problem for the current Giants of the AI is that although they can keep market niches based on excellent benefits, the real business is in the average consumer.
Depseek’s good news is not limited to its low development cost. Although technical details are not entirely public, everything indicates that it requires less computational power and, therefore, has a lower energy cost. This is crucial at a time when many voices warned about the environmental threat that the traditional AI model supposed, due to its enormous demand for energy and water for the cooling of the processors. Deepseek not only democratizes access to AI, but also makes it more sustainable.
Finally, Depseek sends an important message for the implementation of the recently announced “Competitiveness compass”. With this strategy, Europe, which has played a marginal role in the digital economy, intends among other objectives to boost innovation and in particular not lose the AI train. Deepseek’s story teaches us how risky it is to bet on specific companies in such a disruptive sector. Industrial policy based on injecting public money into a few projects, in addition to distorting competition and putting the single market at risk, is almost a game of chance in an accelerated technological change environment. , the European battery company to which more than 900 million euros were granted in subsidies and now faces financial problems, including the bankruptcy of its subsidiary in the United States.
The role of Europe in AI will depend on its ability to create an ecosystem that foster innovation, such as the one that allowed Deepseek’s emergence. To do this, there are no shortcuts: it is necessary to bet on a highly competitive integrated market, to facilitate the financing of startups and strengthen knowledge and research centers. It is no accident that the Deepseek driver, Liang Wenfeng, is another brilliant engineer and that his team is formed by doctors from the best Chinese universities. The lesson is clear: the future of AI is not built with subsidies to a few, but with an ecosystem that rewards innovation and talent.
