Adapted by:
Javier Márquez
WriterI've been in media for over a decade, but I've been marveling at the possibilities that technology brings us much longer. I believe we live in a world where the digital revolution is changing everything and that Xataka is the best place to write about it.
Alba Mora
WriterAn established tech journalist, I entered the world of consumer tech by chance in 2018. In my writing and translating career, I've also covered a diverse range of topics, including entertainment, travel, science, and the economy.
When ChatGPT emerged in November 2022, it ignited one of the most intense technology races in recent years. Companies and governments scrambled to position themselves in the artificial intelligence boom. Central to this reaction were data centers. In the end, they’re essential infrastructures that enable the training of language models, which underpin chatbots and other AI-based applications.
U.S. giants such as Microsoft, Google, Amazon, and Meta announced plans to expand their infrastructure internationally, launching multimillion-dollar projects that ignited a veritable frenzy for these facilities. The AI movement also reverberated in China. The Chinese government declared AI development a national priority and promoted the establishment of new infrastructure to support it.
The Data Center Boom Is Showing Signs of Faltering
According to data from the analyst firm IDC, more than 200 projects related to AI-focused data centers were initiated between 2022 and 2024, spanning 28 provinces and 81 cities in China. This represents a significant surge in growth compared to previous years, fueled by a wave of “smart computing” initiatives. These initiatives not only aimed to enhance the country’s digital infrastructure but also promised to stimulate local economies.
Several prominent companies are involved in this effort, including Alibaba, ByteDance (TikTok’s parent company), Tencent, Baidu, and DeepSeek. All of them are heavily investing in the AI sector. The goal was clear: China couldn’t afford to fall behind the U.S. in a future shaped by AI. To maintain its standing in this race, the Asian nation needed to act swiftly.
Each new project was launched with the understanding that the technology wasn’t yet mature and the business model was still not fully defined. As is often the case, the expectation was that these projects would eventually generate significant economic value–either directly or indirectly–for companies and governments promoting them.
Despite millions of dollars invested in new data centers, enthusiasm for large-scale language models in China is declining. According to MIT Technology Review, over half of the recently built computing resources remain unused. This situation is exacerbated by the technical and market inexperience of many companies investing in this type of infrastructure.
As a result, numerous smaller data centers are struggling to find customers willing to pay for their services. Although China is a vast market, demand isn’t meeting expectations. Major tech companies in the country focus on their own infrastructures. Meanwhile, smaller businesses opt for pay-as-you-go solutions instead of training their own models at these facilities.
Additionally, many data centers built in recent years were designed for pre-training workloads. These long and demanding processes require vast amounts of data. However, the current demand is primarily for inference, which means running already-trained models to provide real-time answers. Many of these infrastructures aren’t adequately prepared in this area.
A Phenomenon That Extends Beyond China
According to analysts from investment bank TD Cowen, Microsoft recently canceled new data center projects in the U.S. and Europe. The company hasn’t made any official announcements regarding which facilities are affected. However, experts attribute the cancellations to reduced commitments to OpenAI, the AI startup where Microsoft has invested billions of dollars.
For years, OpenAI relied solely on Microsoft’s cloud infrastructure. However, it’s recently begun utilizing other computing providers. Despite this shift, Microsoft plans to invest $80 billion in data centers during its current fiscal year, which ends in June. Unexpectedly, analysts predict that the pace of investment may slow.
Images | Scott Rodgerson | DC Studio
Log in to leave a comment