Sigmatic Sciences rolls out agentic life-science R&D workflow built on NVIDIA BioNeMo

Sigmatic Sciences said it has expanded its SigmaticOS platform, an “agentic operating system” aimed at life sciences research and development, by connecting it to NVIDIA’s newest BioNeMo agentic capabilities. The announcement frames the update as a way to move beyond single-purpose AI tools toward coordinated systems that can handle parts of the scientific workflow more autonomously.

In the company’s description of the release, Sigmatic Sciences also said it is already running NVIDIA BioNeMo in production. BioNeMo is NVIDIA’s application and tooling effort built around generative AI for domains including biology, where teams seek to use AI models to support discovery and analysis tasks.

Sigmatic Sciences’ central claim is that the SigmaticOS update can connect AI agents, scientific applications, and workflows in a single environment. The company characterizes the system as “already running” and emphasizes the addition of NVIDIA’s “latest agentic innovations,” suggesting the platform is being updated to reflect ongoing changes in NVIDIA’s agent-building approach for AI.

The announcement was carried by Yahoo Finance, which highlighted the pairing of SigmaticOS with NVIDIA BioNeMo and positioned the result as a step toward more operationalized AI in life-science R&D. While the headline focuses on NVIDIA’s agentic capabilities, the practical takeaway in the release is that Sigmatic is building an orchestration layer intended to manage how AI agents are used across research tasks.

For NVIDIA, the update adds another example of BioNeMo moving from model development into operational workflows. For the life-sciences sector, it underscores a broader shift in AI adoption, where researchers increasingly want systems that can coordinate multiple steps, incorporate domain context, and integrate with existing tooling rather than generating answers in isolation.

It is not clear from the published post alone how SigmaticOS routes specific tasks, what agent configurations are used, or what performance benchmarks the company is targeting. The release also does not provide disclosed details on customer deployments, timelines, pricing, or measurable outcomes such as reductions in experimental cycle time, cost per study, or improvements in discovery yield.

That uncertainty matters because “agentic” systems can vary widely. Two platforms can both claim agentic orchestration but differ substantially in autonomy level, safety controls, traceability, and how reliably they handle uncertainty in scientific data.

Still, investors and customers in AI infrastructure often look for indicators that major model and agent vendors are being incorporated into real workflows. What to watch next is whether Sigmatic Sciences and NVIDIA provide follow-on technical documentation, case studies from production use, or public performance metrics tied to the BioNeMo-linked update.