The Agentic Lab: Why Automata and NVIDIA Are Building the Next Layer of Life Sciences Powered by AI

The lab of the future is not a bigger lab. It's a smarter one.

For the last decade, lab automation meant one thing: robots doing repetitive tasks faster than humans. Liquid handlers. Plate readers. Robotic arms moving labware from point A to point B. Useful, but limited. The robot did exactly what it was told. Nothing more.

That era is ending.

A new layer is forming on top of physical automation: agentic AI. Systems that reason, plan, and adapt. That watch an experiment in progress, notice an anomaly, and decide what to do next. That turns a lab from a collection of instruments into a connected, thinking environment.

This is the moment Automata is building for. It's why we're deepening our work with NVIDIA.

Why NVIDIA, why now

NVIDIA has been building the infrastructure layer for agentic AI across nearly every major industry. Life sciences is now one of its fastest-growing priorities.

NVIDIA's BioNeMo platform is the clearest signal of where things are heading. It's an open development platform purpose-built for AI-driven biology and drug discovery, giving researchers tools to generate and process scientific data, train and optimize models, and deploy them across the full AI lifecycle. NVIDIA NIM microservices handle the path from model to production without rebuilding infrastructure from scratch.

Two NIMs are already reshaping what's possible at the bench. Boltz-2 delivers state-of-the-art biomolecular structure prediction, modeling protein and ligand interactions at speed and accuracy that would have been unthinkable a few years ago. Proteina-Complexa brings multi-chain protein complex prediction into reach for research teams that previously needed dedicated compute resources and weeks of turnaround. These aren't future capabilities. They're live now.

Layered on top is the NeMo Agent Toolkit, enabling multi-agent orchestration for lab environments. Agents that draft protocols, prioritize experiments, and run real-time quality control with far less human intervention than today's workflows require.

But intelligence at this scale needs somewhere to land. A generated protocol is only useful if something in the physical world can act on it, run it, and report back.

That's where Automata comes in.

LINQ: the open execution layer

Automata's LINQ platform is the fully integrated, AI-ready execution layer for this new stack. It integrates instruments, moves labware, runs workflows, and feeds clean, structured data back into the AI layer above it. A closed loop between cognitive computing and robotic implementation.

What makes LINQ distinct is that it's open by design. LINQ talks to any LLM. Claude from Anthropic. GPT-4 from OpenAI. Nemotron from NVIDIA. Scientists aren't locked into a single AI provider or forced to rebuild integrations every time a better model ships. LINQ maintains the physical execution layer while staying agnostic at the intelligence layer above.

That openness matters for the BioNeMo story. As NVIDIA releases new NIMs and expands the BioNeMo catalog, LINQ is positioned to consume them directly. Boltz-2 and Proteina-Complexa are the first examples. The framework is ready for what comes next.

What agentic workflows actually mean in a lab

The real unlock is what happens across the wet lab and dry lab boundary.

That boundary is one of the most expensive friction points in drug discovery. Data from the bench gets manually cleaned, reformatted, and handed off to computational teams in separate systems. The gap between physical experimentation and digital analysis is measured in days, sometimes weeks.

LINQ and the BioNeMo Agent Toolkit collapse that gap. A LINQ-run plate assay feeds structured results directly into a Boltz-2 workflow, which predicts structural implications and surfaces candidates for the next round. The NeMo Agent Toolkit sequences the handoffs. The scientist reviews recommendations rather than transcribing data. What used to take days collapses into hours.

None of this replaces the scientist. It removes the parts of the job that were never really science: babysitting equipment, transcribing data by hand, waiting days to know if a hypothesis was worth testing.

What this means for life sciences leaders

The AI investment you're already making is only as useful as the data feeding it. Siloed, manually operated lab automation is a bottleneck no amount of compute can fix.

The labs that win won't have the most robots or the most AI subscriptions. They'll be the ones that have closed the loop between wet lab execution and dry lab intelligence, turning a slow handoff into a live conversation across any model, any agent, any NIM.

This shift is already in motion. The question isn't if agentic AI reaches the bench. It's whether your infrastructure is ready when it does.

The LINQ platform is built with this future in mind. Open. Connected. Intelligent. The physical, AI-ready execution layer for the agentic systems reshaping life sciences R&D.

We're glad to be building that future alongside NVIDIA.

‍

About Automata

Automata is transforming lab automation for the AI era. The company delivers the only fully integrated, AI-ready platform that is modular, configurable, and open by design. While legacy systems fragment workflows and slow discovery, Automata modernizes lab operations, enabling scientists to program experiments, not robots.

For more information, visit www.automata.tech.