In a significant development for AI infrastructure, Tiger Data has launched Ghost, a managed PostgreSQL database service designed specifically for AI agents. This launch comes as traditional database systems increasingly struggle to meet the demands of software run by autonomous agents.
The New York-based company, operating as Timescale Inc., highlights that Ghost addresses the infrastructure needs for large-scale experimentation among coding and workflow agents. As these agents gain traction, their operational complexities require a reevaluation of traditional database architecture.
"The idea is that all new applications are being built with coding agents," said Ajay Kulkarni, co-founder and Chief Executive of Tiger Data. He further asserted, "We believe the agent is the future packaging of software. Instead of an application, users want outcomes." This viewpoint reflects a broader shift in software development, with agents taking a central role.
Yet, the shift to agent-driven development presents new challenges. Agents often engage in experimental processes, creating temporary environments and testing various methods, which can lead to resource-intensive database operations. "They don’t always do it safely," Kulkarni noted, pointing out the potential for rising costs and complications when agents have unrestricted access to databases.
Ghost aims to reduce these risks by allowing agents to create unlimited databases through a feature called "fast forking." This capability enables users to generate independent copies of datasets in seconds, significantly cutting down the time and resources typically needed for database duplication. In a shift from traditional licensing models, Tiger Data's pricing structure is based on active compute usage instead of per-database fees, making it more cost-effective for users to manage multiple databases.
"Databases are free," Kulkarni stated, highlighting the flexibility of Ghost. Users can set up a database for each agent or task almost instantly, fostering a more dynamic and adaptable environment for AI-driven applications. This innovation is supported by Tiger Data’s proprietary Fluid Storage technology, which uses a copy-on-write model, allowing multiple database instances to share core data blocks. As a result, users are only charged for data that changes, optimizing costs and improving efficiency.
As organizations increasingly adopt AI agents for software development, the launch of Ghost responds to the changing landscape of database requirements. With a client base of 3,000 customers, Tiger Data's focus on enhancing the infrastructure for AI agents positions it as a significant player in the AI token economy and decentralized AI space.
Looking ahead, the implications of this service go beyond convenience. The ability to quickly provision databases is likely to speed up innovation in AI-driven applications, reshaping how businesses approach software development in an agent-centric world. As the demand for agile and responsive infrastructure grows, Tiger Data's Ghost may set a new standard in the database sector.
The stories that move AI & crypto markets — before the market reacts.
Free. 7am ET. Five stories. 62,400 readers.
