Tether Unites 15 Miners for MDK AI Automation Drive

Tether builds industry support for MDK

Tether has secured commitments from 15 Bitcoin mining companies, hardware makers and mining pools to support its open-source Mining Development Kit (MDK). The move signals a push to standardize mining software interfaces and enable AI-driven automation across diverse mining fleets. Participants include major names such as Bitmain, Bitdeer, Canaan, MicroBT, Foundry Digital and F2Pool, among others.

What the MDK collaboration means for Bitcoin mining

Tether introduced MDK earlier this year as an open development framework to let operators and developers monitor and control mining equipment without being locked into proprietary vendor software. Under the new agreement, each of the 15 participants will build and maintain a software integration — called a Worker — that translates its machines' native telemetry and control signals into a common MDK format.

The approach aims to reduce integration complexity for large-scale mining facilities that run thousands of machines from multiple manufacturers, simplifying monitoring, automation and maintenance workflows. Operators could run a single dashboard and automation layer across Bitmain, MicroBT and Canaan hardware, or switch mining pools and management services with fewer custom integrations.

Participants and industry scope

Companies publicly committed to supporting MDK include: Bitmain, Bitdeer, Canaan, MicroBT, Foundry Digital, F2Pool, Luxor Technology and OCEAN. Additional contributors listed are ACME Swisstech, Elektron Energy, EMCD, Heatbit, Proto, SpiderPool and Stratum V2. That mix covers ASIC manufacturers, hosting operators and pool operators — bringing broad representation across the Bitcoin mining stack.

Tether clarified that the announcement reflects commitments to develop integrations; it did not state that every participant has completed their Worker or deployed MDK across live operations.

How MDK works: Workers, interoperability and control

Large mining sites typically rely on separate management tools for each equipment brand, cooling system or pool. MDK's modular architecture addresses that fragmentation using Worker components. Each Worker is responsible for converting vendor-specific telemetry and commands into standardized messages that applications and automation agents can understand.

That layer enables developers to build monitoring tools, dashboards and automation apps that operate across hardware lines without rewriting code for each vendor. Importantly, MDK supports gradual, non-disruptive adoption: operators keep existing firmware and control planes and choose which capabilities to expose through the shared interface.

Tether's CEO explained that scaling industrial mining requires machines to 'speak the same language' at a software level; without a common interface, software-driven optimization and AI control are difficult to implement across heterogeneous fleets.

Open-source history: Mining OS and developer toolkit

MDK complements Tether's earlier Mining OS release, which the company made open source to provide baseline tools for managing mining hardware, energy use and operational metrics. The April MDK release also introduced a JavaScript development kit and a React-based UI library, enabling custom dashboards, controls and integrations that run on Windows, macOS and Linux.

Developers can already access pre-release versions of the framework to experiment and contribute. Tether says MDK can be deployed for use cases ranging from home mining rigs to multi-megawatt industrial facilities.

AI automation: QVAC and agent-driven operations

A central goal of MDK is to enable AI agents to manage mining infrastructure. By exposing standardized telemetry and control endpoints, MDK makes it feasible for AI software to detect equipment faults, optimize power consumption and make real-time adjustments to maximize uptime and hash rate efficiency.

Tether's QVAC mining agent is presented as an early example of an AI application running on MDK. The company reports QVAC is operational on the framework, illustrating how a single agent can interact with multiple vendor systems without separate adapters for each one. However, Tether did not publish independent performance data showing electricity savings, higher mining output or reduced maintenance costs from QVAC.

Expectations and release timeline

Tether said MDK reached version 0.11 during the Plan ₿ Forum in Lugano and is targeting a stable 1.0 release before the end of 2026. The company continues to invite additional hardware manufacturers, mining pools, hosting operators and software developers to join the project and contribute Workers or applications.

The roadmap positions MDK as a foundational layer for future mining innovation, particularly for AI-driven operational tooling and third-party monitoring services.

Wider context: Tether, AI infrastructure and investments

The MDK initiative is part of a broader strategy by Tether to expand into AI infrastructure and edge software. Recently Tether participated in Tenstorrent's funding round — a $1.5 billion raise that valued the chipmaker at $6.2 billion — as the company seeks to support AI compute development. Tether's AI work also includes efforts to run models on personal devices, reducing reliance on centralized cloud services.

While MDK's cross-vendor framework and AI ambitions could lower operational friction for Bitcoin miners and accelerate automation, adoption will depend on how quickly major manufacturers and pool operators complete their integrations and whether independent benchmarks validate the expected efficiency gains.

Bottom line

Tether's coordination with 15 mining industry participants marks a notable attempt to standardize mining software interfaces and enable AI-powered automation through an open-source MDK. If widely adopted, the framework could simplify operations, accelerate third-party tooling and open the door for AI agents to optimize mining at scale — but measurable benefits will require more deployments and independent performance data before operators can quantify cost or efficiency improvements.