AGM Group and HashBeaver target home Bitcoin mining via DePIN nodes
AGM Group Holdings (NASDAQ: AGMH) and digital infrastructure company HashBeaver have announced a joint development project to produce a desktop decentralised physical infrastructure network (DePIN) node built around a purpose-built ASIC chip for Bitcoin mining. The device is designed to sit in homes and offices, connecting individual users to distributed computing networks rather than requiring the industrial-scale facilities that have come to dominate Bitcoin mining over the past decade.
The announcement is a small-company product launch, but it points toward a structurally significant question: can the economics of distributed physical computing infrastructure shift meaningfully toward consumer and enterprise edge devices, and if so, what does that do to the data centre-centric model that currently anchors cloud and AI infrastructure spending?
From mining rig to infrastructure node
The jointly developed product is described as comprising three functional layers. A Bitcoin computing layer provides SHA-256 hashing power via a purpose-built ASIC chip designed by HashBeaver, with a focus on power efficiency and uptime stability. A DePIN connectivity layer assigns each device a unique digital identity, recording verifiable operational data including hashrate, online time and power consumption. A third layer provides AI-assisted device management, analysing ambient temperature, electricity pricing and network conditions to optimise scheduling and power consumption.
AGM Group's role draws on its existing server engineering, thermal management and global supply chain capabilities, applying what the company describes as server-class systems engineering to a compact desktop form factor. HashBeaver contributes chip development, firmware optimisation and mining-pool connectivity, as well as planned integration with third-party DePIN networks.
Dr Bo Zhu, CEO of AGM Group, said the collaboration is intended to position the product at the intersection of Bitcoin mining, DePIN connectivity and AI-assisted device management. Chovy Ni, CEO of HashBeaver, framed it in broader structural terms, noting that "DePIN applies similar distributed principles to a broader range of physical infrastructure sectors, including computing, storage, communications, energy and data."
Both companies are careful to hedge: the product is in pilot production, no pricing or delivery schedule has been confirmed, and the release includes an explicit disclaimer that no Bitcoin mining output, node rewards or economic return is guaranteed. The product's ASIC chip is not designed for general-purpose AI model training or inference, which limits the scope of the "AI" billing.
The convergence angle: edge devices and the data centre duopoly
The broader thesis here sits within a live debate about the architecture of AI and distributed computing infrastructure. The dominant model, heavily backed by sovereign wealth, hyperscaler balance sheets and infrastructure private equity, is the large-scale data centre: GPU clusters concentrated in specific geographies, drawing on dedicated power infrastructure.
DePIN represents an architectural counter-proposal, using blockchain-based incentive mechanisms to coordinate individually owned edge devices into aggregate infrastructure networks. The concept has attracted genuine venture interest, with projects such as Helium (wireless), Filecoin (storage) and Hivemapper (mapping data) establishing proof points across different verticals. Bitcoin mining itself is the oldest and largest DePIN network by computing value, though the term is rarely applied to it retroactively.
For cross-sector investors, the question is whether a wave of consumer-grade DePIN hardware can achieve the verifiable uptime and reliability that enterprise and AI workloads require, or whether the economics of centralised infrastructure are too entrenched. The capital flowing into hyperscale data centre construction, including significant GCC sovereign commitments, suggests the institutional consensus still favours concentration. A desktop Bitcoin miner with AI scheduling software does not yet challenge that thesis. What it does indicate is that hardware manufacturers are positioning for a world in which the edge layer of distributed infrastructure becomes commercially legible to ordinary users, with blockchain providing the coordination and incentive rails.
Whether that market develops at scale, and on what timeline, will depend on Bitcoin network difficulty trends, regulatory treatment of node reward mechanisms across jurisdictions, and whether DePIN networks beyond Bitcoin can demonstrate sufficient demand for verifiable edge compute. All three remain genuinely open questions.