Acurast and Fortytwo to Explore Swarm Inference on Edge Devices
Acurast and Fortytwo Explore Swarm Inference on Edge Devices
We’re thrilled to announce an exploratory partnership between Acurast and Fortytwo. The fastest-growing DePIN compute network and the visionary decentralized AI network will join forces to explore how Fortytwo’s decentralized inference can be achieved on a swarm of thousands of edge devices from Acurast.
The unique value of both networks: Fortytwo uses a network of smaller, specialized AI models that work together to assess and improve responses, delivering high-quality, scalable results without relying on costly infrastructure or centralized management. Additionally, due to their smaller size, these models can be run on Acurast, which in addition is able to provide confidential compute through their TEEs (A Trusted Execution Environment), allowing for usage with sensitive data and preventing the compute being tampered within the process.
By leveraging Acurast’s decentralized architecture of 100,000+ smartphones, spread across 140+ countries, Fortytwo can scale horizontally becoming smarter as both networks scale and paving the path for production-ready confidential swarm inference at the edge.
How Fortytwo works
In Fortytwo’s approach, computation is distributed so no single node handles an entire task. Each node runs smaller, specialized models that generate results reflecting its unique expertise. While individual outputs may be partial or approximate, nodes then review and refine one another’s work, merging domain-specific insights into a collective answer that surpasses any single contribution. The emphasis is on shared intelligence rather than raw compute power.
Swarm inference follows distinct stages. First, a query is split across multiple specialized nodes, each offering its own perspective. These initial responses act like expert hypotheses. Next, nodes evaluate and rank each other’s outputs in an open peer-review process, discarding weaker answers and elevating stronger ones. The top-ranked outputs are then iteratively refined and combined through repeated feedback cycles, steadily improving accuracy and depth. The process concludes with a consensus answer, verified and documented at every step, ensuring both quality and trust.
Where Acurast comes in
Fortytwo’s approach is built to thrive on everyday devices, and it pairs exceptionally well with Acurast’s secure, distributed infrastructure. Rather than requiring expensive data-center hardware, Fortytwo fragments workloads into many small, specialized models that run locally. When those models execute inside Acurast’s Trusted Execution Environments (TEEs), they do so confidentially. That means private fine-tuning and inference can occur on a device without exposing raw data or model internals to device operators or the network, enabling truly confidential AI and secure compute at the edge.
Within Acurast, each device can host multiple light-weight models inside isolated TEEs. The result is a privacy-preserving collaboration: nodes enrich the swarm with specialized knowledge, while sensitive data never leaves the protected execution layer.
Crucially, Acurast’s strength comes from scale and distribution. Acurast’s network leverages the massive, global footprint of smartphones, creating horizontal scaling that grows not just in raw compute but in diversity of expertise. Every additional smartphone that joins the network brings incremental compute capacity plus unique, local knowledge. And because those contributions happen inside TEEs, users keep control of their data while still improving the collective model.
The combination of Fortytwo’s model-fragmentation approach and Acurast’s confidential execution and ubiquitous device reach creates a virtuous cycle: more devices means more private, specialized knowledge leading to stronger, more capable swarm intelligence. In short, Fortytwo and Acurast make decentralized, confidential AI practical and accessible to everyday users worldwide.
In a first step, the two teams will explore how to deploy Fortytwo’s swarm inference on Acurast.
Stay tuned for an update about this cutting-edge partnership!
About Acurast
Acurast is a decentralized physical infrastructure network (DePIN) and a global fleet of smartphones to provide distributed, secure edge compute. Acurast redefines decentralized compute with smartphones — battle-tested devices trusted by billions. With Acurast, AI/ML inference workloads can be securely deployed to a global network of attested edge devices, enabling privacy-preserving tasks close to data sources. Read more on https://acurast.com/
About Fortytwo
Fortytwo is a decentralized AI network that introduces a new type of inference using swarms of small language models running on consumer-grade hardware to achieve scale and reasoning capabilities that exceed centralized AI. The network becomes more powerful and efficient with each new node, operating in a permissionless, transparent, and censorship-resistant manner. Fortytwo allows anyone to contribute by running a node, with node ratings recorded on the Monad blockchain for evaluation and security. Developers can access frontier AI reasoning via API or on-chain, while node operators are rewarded for their contributions. Read more on https://fortytwo.network/