ReadingResearchRadarInvestment frameworkAI Company Primary Sources
Sign in / Sign up
← Back to reading
NewsMaterial4 min readDisconfirmAI Editorial

Fred Sun, General Manager of Tencent Cloud Europe: Entering the European Enterprise AI Market with Open-Source Model Aggregation and Fully Local Sovereign Cloud

Fred Sun, General Manager of Tencent Cloud Europe, explained to diginomica at the 2026 Global Digital Ecosystem Summit that the company aggregates open-source large models such as DeepSeek, GLM, Kimi, and Minimax through TokenHub, allowing European enterprises to select the most suitable model for specific tasks. Simultaneously, it offers Tencent Cloud Enterprise, a private cloud solution with control panels fully deployed on-premises within the customer's local infrastructure, serving as an alternative path to US frontier laboratories.

2026-10-30 19:09 GMT+8

Fred Sun, General Manager of Tencent Cloud Europe, believes that the market landscape changes over the past year are not due to Tencent adjusting its strategy, but rather because AI itself has evolved too rapidly and European customers have moved from observation to actual adoption. Speaking to diginomica during the 2026 Shenzhen Global Digital Ecosystem Summit in China, he noted that last year most were still observing and testing, with few companies truly entering the adoption phase; however, this year almost all enterprises are advancing AI implementation. Sun emphasized that Tencent’s own strategy, execution team, and industry focus have remained stable; what has changed is external perception—he stated that Tencent was once viewed by the market as a latecomer to AI, but that is no longer the case. He mentioned that the upcoming Hy4 model will bring Tencent into the first tier of open-weight models.

Sun summarized Tencent Cloud’s value proposition in Europe as openness, transparency, and choice, arguing that customer demand for these three elements is even stronger than in the cloud computing era. The specific mechanism is the TokenHub platform: customers can not only call upon Tencent’s self-developed Hy models but also access open-source models such as DeepSeek, GLM, Kimi, and Minimax through the same API. Sun claimed that Tencent Cloud is typically among the first, if not the first, cloud provider to launch these latest open-source models on its public cloud API. His core argument is that European enterprises do not need to entrust all tasks to expensive US frontier laboratories (such as OpenAI, Anthropic, Meta, Google, etc.), but should find the most suitable model for each task—whether coding, reasoning, image generation, or agent workloads. He specifically corrected the notion to say 'not the cheapest, but the best fit.'

Sign in / Sign up
ReadingResearchRadarInvestment frameworkAI Company Primary Sources

At the sovereignty level, Sun introduced the Tencent Cloud Enterprise product, emphasizing that its control panel is also deployed within the customer’s own data center, with no connection to Tencent’s public cloud stack. He stated, 'I don’t think most hyperscale cloud vendors can achieve this—even if they offer edge solutions, the control panel usually remains on the public cloud, so it does not constitute true sovereignty or a private cloud.' This logic extends to AI model deployment: if customers possess appropriate data center environments and GPU clusters, Tencent Cloud can help them deploy DeepSeek or Tencent’s self-developed Hy models within their own clusters. Sun said customers choose local deployment for two reasons—sovereignty and data control, and cost: deploying within their own data centers means capital expenditure is fixed, and subsequent operational expenditure is highly controllable.

Regarding the drivers for European customers choosing Tencent Cloud, Sun believes it is almost always a mix of reasons, but the common point is that they look at value first—the product must provide tangible technical value, which is the foundation. The catalyst is often specific cost pressure; for example, if a company’s AI bill suddenly spikes, they must immediately find an alternative. In terms of infrastructure, Tencent Cloud opened a new data center in Frankfurt, Germany, over the past year, but there are currently no similar plans for the UK. Sun explained that the next expansion is likely to remain within the European Economic Area, as more customers require cross-region availability and disaster recovery capabilities within Europe.

However, it must be pointed out that all of the above is Sun’s account as the head of Tencent Cloud’s European region. The article itself notes that Tencent was not immune to the rogue agent problem this year—a risk that also plagues vendors such as OpenAI, Anthropic, Google, and Meta. Furthermore, Sun’s claim that the Hy4 model will place Tencent in the first tier is a vendor expectation that has not yet been verified by independent evaluation.

From a competitive landscape perspective, Tencent Cloud’s European strategy is actually leveraging the maturity of the open-source model ecosystem to position itself as a combination of a model distribution layer and a sovereign deployment layer, rather than simply competing with OpenAI or Anthropic for model capabilities themselves. Whether this path translates into verifiable market share depends on the sustained intensity of European enterprises’ sensitivity to costs and sovereignty demands, as well as whether open-source model quality can keep pace with the closed-source iterations of frontier laboratories.

Sources

  • diginomica.comdiginomica↗
Continue researching →