Fetch.ai and industrial automation company Festo announced a collaboration in February 2021 to explore a decentralised manufacturing marketplace. The idea was to use autonomous software agents to match changing production demand with available capacity across factories.
Manufacturers frequently face an uneven pattern: one facility has idle machinery while another cannot fulfil an urgent order. A digital market could allow equipment and organisations to discover one another, negotiate terms and coordinate work with less manual intervention.
Agents Could Represent Factory Resources
In the proposed system, autonomous economic agents acted for production equipment, buyers and service providers. An agent could advertise a machine’s capabilities and availability or search for capacity that met an order’s technical requirements.
Negotiation could consider price, timing, location and quality constraints. Once owners defined acceptable boundaries, agents could handle frequent matching decisions that would be impractical for people to manage individually.
This model differed from conventional automation inside one factory. It attempted to coordinate independent participants across a broader network, where no single company necessarily controlled every asset.
Volatile Demand Created an Opportunity
Supply chains experience seasonal demand, disruptions and unexpected shortages. Fixed production plans struggle when circumstances change quickly, while adding permanent capacity is expensive.
A marketplace could expose underused resources and make networks more adaptable. A manufacturer facing a temporary surge might purchase capacity rather than invest in machinery that would later sit idle.
The pandemic had made resilience particularly visible. Companies wanted more awareness of alternative suppliers and production routes. Autonomous discovery promised faster responses, although it could not eliminate physical constraints such as tooling, transport and material availability.
Festo Supplied Industrial Knowledge
Festo’s experience in automation and factory systems gave the project a connection to real manufacturing requirements. Industrial equipment has precise capabilities, maintenance needs and safety limits that a generic marketplace cannot infer.
Accurate digital descriptions were therefore essential. An agent needed to know not only that a machine was free, but what materials, tolerances, certifications and processes it could support.
Data quality became part of market integrity. Incorrect availability or capability information could delay an order and create costly defects. Participants needed incentives to maintain reliable records and mechanisms to resolve disputes.
Decentralisation Could Reduce Platform Control
Fetch.ai’s distributed approach aimed to let participants interact without handing every transaction and dataset to one central marketplace operator. A ledger could provide shared records and settlement, while agents retained local control over commercially sensitive information.
That structure might reduce dependence on a dominant intermediary and allow companies to join through common protocols. It could also make governance more complicated.
Rules still had to be written and changed. Identity, access, fees and software updates could shift power even without a traditional platform owner. Effective decentralisation required transparent governance, not simply a blockchain component.
Physical Production Raised the Stakes
An incorrect online recommendation may be inconvenient; an incorrect manufacturing instruction can damage equipment or produce unsafe goods. Autonomous transactions therefore needed clear limits and human approval for high-risk decisions.
Cybersecurity was fundamental. Agents capable of committing capacity or payments would be attractive targets, and connected factory systems can expose operational technology to new threats. Strong identity, segmented networks and detailed audit trails were necessary.
Legal responsibility also had to remain clear. A ledger may prove that a transaction occurred, but it does not decide who is liable when a component fails to meet specifications.
Efficiency Depended on Integration
A useful marketplace would need to connect with enterprise resource planning, scheduling, logistics and quality systems. Without those integrations, a successful match could create more manual work than it removed.
Standards for machine data and digital twins could help agents interpret resources consistently. Adoption would be easier if companies did not need to rebuild their infrastructure around one vendor’s format.
Economic incentives mattered just as much as technology. Companies may hesitate to expose spare capacity if doing so reveals strategic information or creates competition with established customer relationships.
A Practical Test of Industry 4.0
The Fetch.ai and Festo project explored a compelling vision: factories acting as participants in a responsive network rather than isolated production sites. Autonomous agents could improve utilisation and make supply chains more flexible.
The announcement was an experiment, not evidence that manufacturing could be fully delegated. Success depended on trustworthy data, security, interoperability and carefully defined authority.
Its broader lesson was that industrial AI creates value when it coordinates real constraints. The sophisticated part is not merely negotiating a price. It is translating physical capabilities, safety requirements and business commitments into a digital system that companies are willing to trust.
