Baidu chief executive officer Robin Li said on Tuesday, February 11, that investments in data centers and cloud infrastructure are still necessary, even though the example demonstrated by DeepSeek contains reasons to doubt the cost efficiency of large artificial intelligence models.
Robin Li made the mentioned statement during the World Government Summit in Dubai. He noted that financial injections into cloud infrastructure are still very much required. According to him, in order to come up with models that are smarter than everyone else, it is necessary to use more computing.
Compute refers to hardware resources that ensure the operation of artificial intelligence models, allowing the corresponding functional systems to train on data, process information, and generate forecasts.
It is worth mentioning that the Chinese startup DeepSeek has gained global attention for the elaboration of language models that match the performance of leading products, such as OpenAI’s GPT while using significantly less computing power. Against this background, doubts began to circulate about the need for huge financial injections into the artificial intelligence infrastructure.
Baidu was one of the first Chinese companies to launch machine intelligence products after the debut of OpenAI’s ChatGPT in November 2022. The firm claims that its large language model Ernie matches GPT-4’s capabilities. It is worth noting that the mentioned model has seen limited public adoption.
Robin Li sees the prospects for the development of the Chinese artificial intelligence industry through the prism of enthusiasm and considerable optimism. At the same time, in his opinion, another company like OpenAI is unlikely to emerge in China. He is also a proponent of closed-source artificial intelligence models. In his opinion, this is the only viable path to develop machine intelligence.
On Tuesday, Robin Li acknowledged that Deepseek’s sudden emergence demonstrated the unprecedented nature of innovation. In this context, he noted that the computing power constraints have driven Chinese companies to innovate to reduce costs. Robin Li also probably softened his stance on closed-source development. He stated that open-source approaches can accelerate artificial intelligence adoption.
In the context of the question of the expediency of significant financial injections into the machine intelligence industry against the background of the success of DeepSeek, which demonstrated maximum results at minimal cost, it is worth noting that recently, investors have been gradually starting to face uncertainty about returns from the funding of digital intelligence area. Artificial intelligence as a promising technology is what can be called a long-term project. This means that the potential of AI in all its dimensions, including in terms of the economic aspect, is not a kind of issue of instant realization. At the same time, there is still no definitive understanding as to when exactly the moment of returning from financial injections to the artificial intelligence industry will come. However, AI clearly has significant potential, which is why investing in the relevant industry can lead to a result that turns out to be unimaginably good. At the same time, it is worth noting that in the artificial intelligence race, as in any other competitive process, literally every player will not become a winner and will not occupy their specialized niche.
DeepSeek spent less than $6 million on the development of its digital product. From the point of view of the average statistical indicators typical for the artificial intelligence industry, the mentioned amount is actually minimal. At the same time, the DeepSeek example is what can be called a single case, which has not yet become a general practice that has proven its viability.
As we have reported earlier, EU to Mobilize 200 Billion Euros for AI Investments.
