Autonomous software systems running on Tencent Cloud AI agents are creating more than 20,000 TDSQL-C database instances every day, the Chinese cloud provider has reported. The figure marks an eightyfold increase in automated provisioning, taking the total count of active agent-managed databases on the network past one million.
Key points
- Tencent Cloud reported that autonomous agents now provision more than 20,000 TDSQL-C databases daily.
- The current creation pace represents an eightyfold increase, lifting active instances past one million.
- Individual users are running tens of agents concurrently, with some targeting hundreds within a year.
Surge in automated infrastructure
Tencent Cloud disclosed the milestone as software developers shift from single-prompt assistants toward independent agents capable of managing technical resources. TDSQL-C, the provider’s cloud-native relational database service, is increasingly being spun up directly by autonomous software rather than human system administrators.
Because autonomous agents frequently require dedicated environments to store task states, process records and maintain memory across interactions, programmatic database deployment has grown alongside agent autonomy.
The rapid uptake mirrors broader developer trends across the industry. While AI coding agents gain direct access to internal programming tools and codebases, auxiliary agents are being granted permission to provision backend services without manual intervention.
Scaling agent fleets
The operational data highlights how single operators are assembling fleets of autonomous software. Li Ning, a developer using the agent platform WorkBuddy, currently operates approximately 50 agents simultaneously. Li said he plans to scale that number to 200 agents within the next 12 months.
These setups allow individual engineers to delegate distinct operations, such as data scraping, verification, workflow tracking and customer communications, to separate software entities. Each independent process often requires its own persistent storage layer, which explains why database creation figures have outpaced raw user growth.
Similar patterns are emerging in enterprise deployments worldwide. Investors have poured heavy funding into autonomous operational software, highlighted when Serval raises $150 million to expand its enterprise AI agents across corporate systems. As developers expand these multi-agent networks, cloud platforms face a different profile of computing demand characterised by rapid provisioning, short-lived workloads and automated teardowns.
Impact on cloud database architecture
Traditional cloud architecture relies on human planning: administrators estimate traffic, choose compute tiers, set up failovers and run databases for months or years. Agent-driven consumption breaks that model by creating and configuring storage clusters on demand in seconds.
To support this volume, cloud providers have adjusted serverless database engines to handle frequent spin-up cycles without running into provisioning delays. Other startups are building dedicated infrastructure around worker behaviour, including projects where enterprise AI agents startup Crux Labs attempts to capture human judgment and operational decisions for automation.
Tencent Cloud has not disclosed how long the average agent-created TDSQL-C database remains active or what proportion are decommissioned after completing specific tasks. The company is expected to release further technical specifications for agentic infrastructure at its upcoming developer events.
