Tenstorrent Blackhole p150a (32GB)

The Tenstorrent Blackhole p150a is Tenstorrent’s newest and most powerful single-card AI accelerator: an actively cooled PCIe card built around the Blackhole architecture, the successor to Wormhole. It packs more Tensix cores, more on-chip SRAM and faster memory into a single-ASIC design, backed by a fully open software stack. It suits researchers, kernel engineers and AI developers who want serious on-card compute in a desktop- or workstation-ready package.

Who it's for

• AI research teams building or fine-tuning custom model architectures
• Kernel engineers who want to hand-optimise compute graphs at a level GPU vendors do not expose
• RISC-V-focused developers who want general-purpose cores sitting beside the AI compute fabric
• Teams planning a multi-card cluster that scales bandwidth beyond standard PCIe-only multi-GPU setups

Key features

• 120 Tensix cores: Tenstorrent’s core compute unit, each built around a cluster of small RISC-V processors rather than a traditional GPU shader pipeline, for fine-grained control over how work is scheduled and executed.
• 16 “big” RISC-V cores: general-purpose processors on the die for control, orchestration and host-side logic, so latency-sensitive or highly custom pipelines make fewer round-trips to the CPU.
• 180MB of on-chip SRAM: roughly 1.5MB per Tensix core, a large local pool that keeps data close to the compute units and eases pressure on off-chip DRAM bandwidth.
• 32GB GDDR6 on a 256-bit bus: 16GT/sec for 512GB/s of memory bandwidth, enough room for meaningfully large local models and datasets.
• Up to 1.35GHz AI clock: delivers 664 TeraFLOPS at BLOCKFP8 precision, a substantial step up from Wormhole.
• 4x QSFP-DD 800G ports: passive high-speed networking that links multiple Blackhole cards to pool memory and scale compute across a multi-card system.
• PCIe 5.0 x16 host interface: the latest-generation host interconnect for maximum throughput between card and system.
• Fully open-source software stack: the open compiler and runtime approach (TT-Metalium, TT-NN) that defines Tenstorrent’s product line, with nothing hidden behind proprietary drivers.

Specifications

Good to know

• The trade-off is software maturity versus openness. TT-Metalium and TT-NN are open, actively developed and show how your workload maps to silicon.
• The ecosystem of pre-optimised model support is smaller than CUDA’s or even ROCm’s. Expect to work with Tenstorrent’s programming model directly rather than getting instant, zero-effort parity with a mainstream GPU framework.

Choose the p150a if you value an open stack you can read and modify at every layer and are willing to invest in learning it. If you want mainstream frameworks to work with no effort, a conventional GPU is the easier route.