Back to News
IndustryAI Understanding briefing

Metacognition AI raises $10 million for robotics operating system

Metacognition AI secured $10 million from Main Sequence to develop a voice-controlled operating system that allows non-experts to train industrial robots and AI agents.

4 min readRead the primary source
Source-provided image accompanying Metacognition AI raises $10 million for robotics operating system
Source referenceSource recorded
Publisher
startupdaily.net
Source link
startupdaily.nethttps://www.startupdaily.net/topic/funding/ai-startup-raises-10-million-to-build-robotics-operating-system/
Source type
Linked source — primary-source status has not been established.

Story last revised

ContextUnderstand this in 60 seconds

Start here

Key terms

Natural Language Processing (NLP)
The branch of AI focused on understanding and generating human language.
Machine Learning (ML)
Methods that allow systems to learn patterns from data and improve over time.
Test yourselfAI Agents Quiz

What happened

Metacognition AI, a startup cofounded by Anton van den Hengel, Stephen Gould, and Paul Dalby, has raised $10 million in funding from Main Sequence, a deep tech venture capital firm backed by CSIRO. The company is developing a 'Windows for robots' operating system designed to sit beneath large language models, enabling users to train robotic hardware and AI agents using simple verbal instructions and feedback. The startup is currently targeting existing industrial systems, with simulations already built for directing open-cut mines and managing electricity grids.

Metacognition AI has closed a $10 million funding round led by Main Sequence, a deep tech venture capital firm with backing from CSIRO. The raise was initially reported by Capital Brief. The startup was cofounded this year by Anton van den Hengel, the chief scientist of the Australian Institute for Machine Learning, along with Stephen Gould and Paul Dalby.

The company is building an operating system specifically for AI agents and robotics, which it describes as a 'Windows for robots.' This software layer sits underneath large language models to control how they interact with physical machinery and other systems. The core functionality allows users to train robots to perform useful tasks using simple verbal instructions and feedback, rather than requiring complex coding or expert programming.

Metacognition states that its platform will enable robots to work alongside humans, remembering instructions and improving performance as they interact with their environment. The company claims this approach will lead to more capable and safer AI agents that are more personalized to their users by providing the missing software platform for non-expert training.

The startup is initially targeting existing industrial systems. It has developed simulations where operators can direct an open-cut mine or manage an electricity grid by speaking to the system. The team currently consists of eight members and is expanding by hiring additional AI researchers.

Source details: startupdaily.net

Why it matters

This funding supports the development of a middleware layer that aims to lower the barrier to entry for industrial automation. By allowing non-experts to program complex machinery through voice, the technology could significantly reduce the specialized skills required to operate advanced robotics in sectors like mining and energy. The approach focuses on making AI agents more personalized and safer by grounding their actions in specific, user-provided verbal feedback rather than relying solely on pre-trained model capabilities.

The development addresses a significant gap in industrial automation: the complexity of programming robotic hardware. By shifting the interface to natural language, Metacognition aims to make advanced robotics accessible to a broader workforce, potentially reducing the reliance on specialized engineers for routine task configuration.

The positioning of the software as a layer beneath large language models suggests a focus on orchestration and safety controls. This architecture could help mitigate risks associated with autonomous agents by grounding their actions in specific, user-verified instructions, which is a critical concern in high-stakes environments like mining and energy management.

The involvement of Anton van den Hengel, a prominent AI researcher, lends technical credibility to the project. The backing from CSIRO-linked investors indicates a focus on deep tech applications with practical industrial utility rather than purely consumer-facing AI products.

What to watch next

Monitor the hiring of additional AI researchers and the expansion of the eight-person team. Watch for the release of public demos or pilot programs in the mining and electricity grid sectors. Track whether the platform achieves interoperability with major existing industrial hardware standards.

The next phase of hiring will indicate the company's technical priorities, particularly in areas of natural language processing and robotics integration. The expansion of the eight-person team is a key indicator of development pace.

Public demonstrations or case studies from the mining and electricity grid simulations will be crucial for validating the effectiveness of voice-controlled industrial operations. Independent verification of these simulations' accuracy and safety protocols will be necessary.

Partnerships with major industrial hardware manufacturers or system integrators will determine the platform's ability to scale beyond simulations into live operational environments.

Related guides & quizzes

AI AgentsWhat is AI?Future of AITest what you know — try a free AI quizLook up an AI term in our glossary
Found this useful?