The Pulse
Waters Launches AI Link and Joins Anthropic’s Hardware Program
Waters launched AI Link, which connects AI assistants to its waters_connect laboratory software through natural-language requests. The company also joined Anthropic’s Model Hardware Standard research program, aimed at letting AI agents oper
AI.info Team · Written with AI ·
From weeks of integration to hours
Waters says Anthropic’s Model Hardware Standard could cut the time needed to connect laboratory instruments from weeks or months to hours or minutes. The company announced on October 13 that it has joined the standard’s research program and launched Waters AI Link, software that connects AI assistants to its waters_connect laboratory platform.
The two efforts have different roles. AI Link brings Anthropic’s Model Context Protocol to waters_connect, allowing users to request supported software actions in natural language. The hardware program is intended to let agents operate physical equipment directly; Waters says the combined capabilities will support that work, but the announcement describes its participation in the program as beginning immediately, not as a completed deployment of autonomous instrument control.
AI Link works through defined software tools
Waters describes AI Link as an MCP server that exposes selected waters_connect functions as governed tools. Its current capabilities cover quantitative acquisition and processing tasks, including creating or editing sample lists, working with acquisition methods, assigning instruments, accessing results and generating reports.
“AI can generate hypotheses faster than any lab can test them, so the bottleneck has moved from computation to the bench,” said Dave Leitham, vice president of informatics at Waters Corporation. He said each hypothesis can become a physical sample that must be separated and measured, increasing demand on instruments and columns.
Waters says the software does not grant an AI model unrestricted access to the platform. Each tool is defined and constrained, and actions remain subject to waters_connect authentication, permissions and access controls. Users can perform through AI only the actions they are authorized to take in the software.
Physical instrument control is a separate research track
Anthropic introduced the Model Hardware Standard as a shared specification for agents to operate programmable devices through a common interface. Its August research preview describes use with equipment such as liquid handlers and robotic arms, and says the standard is designed to reduce bespoke device-integration work. Waters’ announcement places its analytical instruments within that research effort.
The distinction matters: a conversational interface to laboratory software is not, by itself, proof that an agent can safely execute a physical experiment. Waters says the intended hardware workflows could let agents run experiments, adjust parameters as results arrive and respond to instrument errors. Those are capabilities the announcement associates with the program, rather than reported outcomes from a Waters deployment.
Early access starts before November availability
Waters AI Link is available now in early access, with general availability planned for November for waters_connect customers. The release says the software can connect to Claude and other large language models and agents, and that agents can create methods, run acquisitions and process results without custom integration.
Waters also frames the product as a bridge between research workflows and regulated development. The company says work can move from waters_connect into Empower software, which it describes as the compliant system of record for regulated settings. Its announcement presents this continuity as a way to carry research and development workflows toward quality-control use; it does not provide customer performance data or detail how specific instrument-control actions will be validated.