Skip to main content

NIST programmable formulation and NSF shared-access direction

Geography: United States

Verified: NIST's Autonomous Formulation Lab links AI-directed experimental planning to robotic sample preparation and in-situ scattering, rheology and spectroscopy. The US National Science Foundation is also developing a test bed for remotely accessible programmable laboratories.

Design question: Could shared instruments expose secure, validated interfaces so an experiment can move from design to preparation, measurement, analysis and human review without losing provenance?

Sources and scientific attribution

NIST Autonomous Formulation Lab

Scope: Official NIST project description; not a medicines-manufacturing approval.

Source: https://www.nist.gov/programs-projects/autonomous-formulation-lab

Air moves when the room boundary moves

A laboratory benchmark needs to include what happens at doors and handovers. This published airflow simulation gives prospective occupiers a concrete reason to ask about containment conditions, door operation and verification, instead of relying on a room label or a single design specification.

Published large-eddy simulation of air exchange between a hospital isolation room and an anteroom during single-hinged-door operation. This is an external simulation, not a NEUVIOR campus ventilation model.

Saarinen P, Kalliomäki P, Tang J, Koskela H · Air moves when the room boundary movesCC BY 4.0 | Original research and methods

Source and image notes

H.264 viewing copy; complete frame retained with proportional downscaling where required; no colour enhancement, crop, speed change or added frames. Audio omitted; original source remains linked. Original declared 749samples but decoded748frames; two terminal presentations at29.92s and29.88s excluded because first is outside declared end and second repeats an earlier timestamp. Kept746frames, original timestamps0..29.88s including original0.12s acquisition gap. No frames invented or duplicated.

Follow the changing boundary

A door movement changes the flow problem; a static room specification does not describe every operating state.

Ask how it is verified

Simulation settings, pressure conditions, measurements and the relevant use case must accompany a containment argument.

NSF Programmable Cloud Laboratories Test Bed

Scope: Funding programme and infrastructure direction; not a completed national network.

Source: https://www.nsf.gov/funding/opportunities/pcl-test-bed-test-bed-toward-network-programmable-cloud-laboratories

USTC end-to-end robotic chemistry demonstration

Geography: China

Verified: A University of Science and Technology of China research system combined literature mining, a computational planning layer, mobile robotics and fourteen workstations for synthesis, characterisation and performance testing.

Design question: Which bounded chemical workflows could be safely closed-loop, and where must a scientist approve the next experiment, exception or escalation?

Sources and scientific attribution

National Science Review: all-round AI chemist

Scope: Original research demonstration; not evidence that all Chinese laboratories are autonomous.

Source: https://academic.oup.com/nsr/article/9/10/nwac190/6694008

White Andrew+ liquid-handling robot with electronic pipettes, multichannel tips and laboratory holders.

Andrew+ liquid-handling robot by Andrew Alliance, photographed by Pocar19 in 2019. External equipment benchmark for discussing laboratory automation; no installation, ownership, access, endorsement or performance result at NEUVIOR is represented.

Pocar19 · Andrew+ robot.jpgCC BY-SA 4.0

Chinese Academy of Sciences: robotic chemist Luke

Scope: Institutional report on a defined research platform.

Source: https://english.cas.cn/newsroom/news--archives/2024/cas-in-media/202411/t20241108_1130330.shtml

RIKEN biological protocol optimisation

Geography: Japan

Verified: RIKEN reported a human-robot-AI system that evaluated 143 cell-culture conditions over 111 days from a much larger search space, improving a defined retinal-cell culture metric. The result concerned experimental optimisation, not a clinical treatment.

Design question: How should long-duration cell workflows combine robotic precision, image-based quality control, incubator continuity and accountable human decisions?

Learning from research

Which part of a laboratory breakthrough can travel?

A published automation result can inspire a programme without proving that its operating model transfers. Inspect what sits between the demonstrated method and a dependable service: instrument interfaces, review, recovery and the intended use. The comparison helps researchers and operators decide what deserves a bounded feasibility study.

Researchers and prospective laboratory operators

Editorial decision model, not measured NEUVIOR results.

The transferable insight may be a method rather than a machine.

A scientist should establish what made the published experiment informative and which assumptions need to be retested for a new material or question. That approach produces a researchable transfer task while preserving attribution and avoiding borrowed claims about speed, throughput or discovery.

Compare the original experimental boundary with the proposed one.

In focus

Demonstrated method

A benchmark is useful when its experimental task and limits remain attached. NIST’s formulation work illustrates connected preparation and measurement; another institution’s successful result is a starting point for questions rather than evidence that Constellation can reproduce it.

  • Preserve the method
Evidence to keep in view

Original method, material range and stated research purpose.

A useful next step

Identify the mechanism worth investigating rather than copying the performance claim.

A strong benchmark becomes a testable operating question, with its original result still attributable to its authors.

Sources and scientific attribution

eLife: robot and AI optimisation of stem-cell differentiation

Scope: Original research supporting the reported 143 conditions and 111-day workflow; not a clinical outcome.

Source: https://elifesciences.org/articles/77007

RIKEN: human-robot-AI teamwork

Scope: Research-process optimisation; no clinical outcome or NEUVIOR performance claim.

Source: https://www.riken.jp/en/news_pubs/research_news/pr/2022/20220628_2/index.html

RIKEN Center for Biosystems Dynamics Research

Scope: Current institutional context for AI biology and laboratory robotics.

Source: https://www.riken.jp/en/research/labs/bdr/

Liverpool mobile robotics and UK medicines translation

Geography: United Kingdom

Verified: The University of Liverpool demonstrated a mobile robotic chemist working with ordinary human-scale laboratory equipment. In 2026, Innovate UK offered up to £7.5 million for digital, automated and robotic medicines process-development and manufacturing technologies.

Design question: Can adaptable robotic workcells and shared translation space complement existing British facilities without duplicating specialist capability already available elsewhere?

Sources and scientific attribution

University of Liverpool mobile robot scientist

Scope: University research demonstration using a defined workflow.

Source: https://news.liverpool.ac.uk/2020/07/08/experiment-too-big-hire-a-mobile-robot-scientist/

UKRI: medicines manufacturing labs of the future

Scope: Closed funding competition; not a NEUVIOR award or participation claim.

Source: https://www.ukri.org/opportunity/medicines-manufacturing-labs-of-the-future/

Telangana policy and IIT Hyderabad shared measurement

Geography: India / Hyderabad

Verified: Telangana's 2026 to 2030 policy describes a shift from volume-led growth towards value-driven innovation across advanced biomanufacturing, biologics, precision medicine and digital health. IIT Hyderabad's SATHI centre provides shared advanced characterisation capability.

Design question: Could the Constellation brief connect automated discovery, advanced characterisation, quality-by-design and pilot translation while adding to, rather than erasing, the region's established medicines strengths?

Sources and scientific attribution

India DST SATHI shared infrastructure

Scope: Official shared-facility description, including IIT Hyderabad.

Source: https://dst.gov.in/sophisticated-analytical-technical-help-institutes-sathi

What the benchmarks can inform

Transfer the operating pattern

Shared access, modular workcells, interoperable instruments, traceable samples, machine-readable methods and explicit human approval gates can inform the brief.

Do not transfer another institution's result

A published discovery, throughput figure, clinical result or facility status belongs to its source. It is not evidence of NEUVIOR performance, demand or readiness.

Design for exception handling

Autonomy still requires calibration, maintenance, loading, safe states, deviation management and qualified people able to stop or recover the system.

Keep location claims specific

Potential host regions already have scientific assets. Constellation should establish the additional access or translation problem it could solve, compare access to existing facilities and avoid treating another institution’s equipment or results as its own.

NEWSPerspectives & features