According to the analysis of Gavekal Research and Tilly Zhang, DS achievements must be contextualized. Although it has optimized the use of old chips with the «mixture of experts», its development has required a much greater investment than it is advertised. Nevertheless, His software innovation could represent an important competitive advantage.
Ben Thomson’s «Deepseek FAQ» raises key questions: is DS leading? Despite its advances, Openai continues to show models with better reasoning. So why the stir? The surprise comes from China’s ability to challenge US laboratories, low training costs and technology development despite chips export restrictions.
The impact of this news is significant for Big Tech, particularly for semiconductor manufacturers such as NVIDIA. This company maintains a key advantage with its CUDA language, exclusive to its chips, and its ability to integrate multiple GPUS (the data centers graphics processing units are accelerators that enable and improve technologies such as virtual AI) to enhance AI. However, the advance of DS could affect their domain.
From the business point of view, Three factors continue to favor NVIDIA:
1. Although DS has found more efficient ways to use computing, this does not eliminate the need for greater processing capacity.
2. A lower inference cost can increase the use of AI, as Satya Nadella, CEO of Microsoft, pointed out, referring to the «Jevons paradox»1 .
3. Advanced models such as R1 and O1 depend largely on more computing power to improve their performance.
Among the great technology, Microsoft would benefit from reduce costs in data centers and gpus, while Apple could take advantage of the reduction of memory requirements. Goal would also win, since lower inference costs would facilitate its leadership strategy in AI. On the other hand, Google seems to be the most vulnerable, since lower costs could facilitate the appearance of products that challenge their search engine, affecting their dominant position.
In conclusion, the future of AI and its impact on the industry remains uncertain. In the short term, the irruption of DS generates concern and calls into question the competitive advantages of American technology. Nevertheless, History shows that technological evolution tends to improve productivity and reduce costsbenefiting the economy and users. AI, far from being an isolated phenomenon, seems to be only part of a broader process, in which other innovations, such as quantum computing, could play a crucial role in the future. Of course, technology still has their space in investment portfolios, although everything that shines is not gold, analyzing the fundamentals, the moment of entry and active management are more necessary than ever in this context, the winners are not so clear in this technological career and will come more surprises due to the vertiginous innovation of the sector.
As technological improvement increases, the efficiency with which a resource is used, an increase in the consumption of said resource is more likely than a decrease.