QBlocks AI Workstations Break Free from GPU Limitations
Enterprise-grade AI computing power without vendor lock-in. Engineered for teams of 50-100 developers.
Liquid-cooled & Air-cooled options • Open-source • Superior price-performance
QBlocks AI Workstations
Built on Tenstorrent's open-source RISC-V architecture. You get direct "access to the metal" - full control over hardware without licensing fees or vendor restrictions.
The QBlocks – Blackhole Liquid-Cooled Desktop Workstation is a great solution for developers running or testing AI models, or porting and developing libraries for HPC. QBlocks – Blackhole is equipped with four Tenstorrent Blackhole™ p150c cards, with a cumulative 480 Tensix Cores™, 64 "big RISC-V" cores, 840 MB of SRAM, and 128 GB of GDDR6.
These processors are connected with a flexible, high-bandwidth, Ethernet-based mesh topology to facilitate rapid data movement and pool resources. This empowers QBlocks – Blackhole to run single user/single models up to approximately 80 billion parameters and single/multiple user, multiple models up to approximately 20 billion parameters.
With Qlose AI Infrastructure (example) • Multi-tenant serving • Mid-size models (7B-30B) • Collaborative workspaces • API interfaces • Version control integration • Team dashboards
QBlocks – Wormhole is equipped with four Tenstorrent Wormhole™ n300 cards for a total of eight Wormhole™ Tensix Processors.These processors are connected with a flexible, Ethernet-based mesh topology that can expand to achieve a 96 GB memory pool.
QBlocks – Wormhole is supported by a fully open source software stack with multiple entry points for developers at every level.
With Qlose AI Infrastructure (Example) • Multi-tenant serving • Mid-size models (7B-30B) • Collaborative workspaces • API interfaces • Version control integration • Team dashboards
The QBlocks-LoudBox 4U/Desktop Workstation offers superior performance per dollar for developers looking to run, test, and develop AI models or port and develop libraries for HPC.
Equipped with four Tenstorrent Wormhole™ n300s cards for a total of eight Wormhole™ Tensix Processors, this flexible, Ethernet-based mesh topology can expand to achieve a 96GB memory pool. This enables QBlocks-LoudBox to run single user/single models up to approximately 80 billion parameters and single/multiple user, multiple models up to approximately 20 billion parameters.
With Qlose AI Infrastructure (example) • Multi-tenant serving • Mid-size models (7B-30B) • Collaborative workspaces • API interfaces • Version control integration • Team dashboards
Liquid-cooled (QuietBox) for offices or air cooled (LoudBox) for datacenters
Open-Source
Complete hardware access with no restrictions from vendors.
Enterprise-Ready
Qlose AI integration for supporting teams 50-100 members, boosting collaboration and productivity.
Ready to Own Your AI Data Now?
Talk to our team about your infrastructure needs – from a single deploymentto a national sovereign AI program.
Contact Us
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Or reach us directly via email – contact@qblocks.ai
Frequently Asked Questions
Can I use these workstations without Qlose AI infrastructure?
Yes. You can purchase standalone workstations with just the base operating system and Tenstorrent's open-source software stack (TT-Forge™ compiler). This gives you complete freedom to build your own infrastructure. Qlose AI is an optional add-on for teams who want turnkey enterprise AI deployment.
How do QBlocks workstations compare to NVIDIA GPU-based systems?
QBlocks workstations offer superior price-performance with no vendor lock-in. You get full hardware access through open-source RISC-V architecture, no licensing fees, and complete control over your software stack. Traditional GPU solutions restrict you to proprietary tools and lock you into a single vendor's ecosystem.
What size teams are these workstations designed for?
Our workstations are optimized for teams of 50-100 developers. For Qlose AI deployments, a typical setup uses 10-15 workstations to support this user base. Smaller teams can start with 2-4 workstations, while larger enterprises can scale beyond 100 users with additional units.
What's the difference between QuietBox (liquid-cooled) and LoudBox (air-cooled)?
QuietBox models use liquid cooling for whisper-quiet operation, making them ideal for office environments where noise matters. LoudBox is air-cooled in a 4U rack-mountable form factor, designed for datacenter deployments where density and standard rack integration are priorities. Both offer the same computing performance.
