Tencent Cloud and Elastic have announced an expanded strategic collaboration in Shenzhen to help modern businesses establish dedicated context infrastructure for the AI era. The deep partnership focuses on technical and product collaboration across enterprise data, AI search, and agentic applications.

As part of this effort, the two companies jointly introduced Tencent Cloud Elasticsearch Service (ES) Enterprise Edition. This platform combines Elastic’s enterprise-grade capabilities with custom in-house engineering from Tencent Cloud, which has been offering cloud solutions based on Elasticsearch since 2016. In specific AI search scenarios, Tencent Cloud’s optimizations have improved performance by up to 5x while reducing memory consumption by over 50% and cutting hybrid search latency by 60%.

Search has become a critical piece of modern AI systems because large language models do not inherently understand a company’s internal data. To provide accurate answers and perform complex tasks, these models depend on search to retrieve real-time context from secure databases.

Patrick Dixon, Senior Director of Channels & Alliances for APJ and Vice President for China at Elastic, said that as large language model capabilities advance, context is becoming increasingly important to enterprise AI applications. For agents, the ability to locate accurate and reliable enterprise data when it is needed directly affects the quality of subsequent responses, decisions, and execution.

The newly launched service allows enterprises to adopt AI search capabilities without altering their current data architecture. It works seamlessly alongside the Tencent Cloud TCRay inference platform and Tencent Hunyuan large language models. Furthermore, customers who are running open-source editions can easily access technical migration support.

Several organizations are already using these features to enhance their operational workflows. For instance, Tencent ima uses ES as its underlying retrieval layer. Following performance tuning for high numbers of concurrent users, retrieval speed improved by 58% and memory consumption dropped by 71%. Automotive manufacturer NIO utilizes the platform to process hundreds of billions of security records each quarter. During a real-world security incident, analysts reconstructed an attack chain in just two minutes, a process that previously took most of a day to investigate manually.

Cheng Bin, General Manager of Data Platform Products at Tencent Cloud Data Analytics, said, “AI search is evolving from a standalone retrieval capability into infrastructure that connects enterprise data, models, and agentic applications. Tencent Cloud will combine Elastic’s enterprise-grade capabilities with its own experience in operating large-scale systems and its AI ecosystem, helping more enterprises establish context infrastructure for the AI era.”

Han Xiao, Vice President of AI at Elastic, noted that search in 2026 is about test-time compute. He added, “When search quality falls short, teams should first consider whether additional inference compute can address the problem before adding data or switching to a larger model.”

Tencent Cloud runs 20,000 clusters and 100,000 nodes for its ES platform, maintaining a reliable foundation for major global search traffic peaks.