Tenstorrent Blackhole P150a AI Accelerator – 32GB GDDR6 | 664 TFLOPS | 120 Tensix Cores | 512GB/s | PCIe 5.0

Next-Generation High-Performance AI Computing with Multi-Card Scalability

Experience powerful artificial intelligence acceleration with the Tenstorrent Blackhole P150a, an advanced PCIe 5.0 AI accelerator engineered for AI developers, machine learning researchers, data scientists, and high-performance computing professionals.

Powered by Tenstorrent’s Blackhole architecture, the P150a combines 120 Tensix AI cores, 16 SiFive X280 RISC-V processor cores, 32GB GDDR6 memory, 512GB/s memory bandwidth, and up to 664 TFLOPS BLOCKFP8 computing performance.

Unlike conventional AI accelerator cards designed exclusively for standalone operation, the Blackhole P150a features four 800Gbps QSFP-DD interconnect ports, enabling multiple compatible Tenstorrent Blackhole accelerators to communicate directly.

Whether you’re developing advanced neural networks, experimenting with generative AI, building AI research workstations, or exploring scalable machine learning infrastructure, the Blackhole P150a provides a powerful and flexible hardware platform.

Key Features & Highlights

• 664 TFLOPS BLOCKFP8 AI Performance – High-throughput tensor computing for supported AI workloads.
• 120 Tensix Processing Cores – Specialized architecture designed for efficient AI acceleration.
• 16 SiFive X280 RISC-V Cores – Programmable processor architecture for advanced computing.
• 32GB GDDR6 Memory – High-capacity onboard memory for AI models and data-intensive applications.
• 512GB/s Memory Bandwidth – Rapid memory access for demanding AI processing tasks.
• 180MB On-Chip SRAM – Fast local memory for efficient data movement.
• PCI Express 5.0 x16 Interface – High-speed connectivity with compatible workstations.
• 4 × QSFP-DD 800Gbps Ports – Dedicated high-bandwidth links for compatible Blackhole accelerators.
• Up to 1.35GHz AI Clock – High-performance Blackhole processor architecture.
• 300W Total Board Power – Performance-focused AI accelerator design.
• Active Blower-Style Cooling – Designed for installation in compatible desktop workstations.
• Open-Source Software Stack – Developer access to Tenstorrent’s software and programming tools.

Advanced Blackhole AI Processor Architecture

The Tenstorrent Blackhole P150a uses a specialized processor architecture designed to efficiently execute tensor operations and support modern artificial intelligence applications.

Its 120 Tensix cores provide dedicated computational resources for supported neural network operations, while its integrated RISC-V cores enable programmable processing capabilities.

The architecture combines high-performance computation, fast onboard memory, and dedicated interconnect technology to support advanced AI development.

664 TFLOPS of AI Computing Power

With up to 664 trillion floating-point operations per second in BLOCKFP8 format, the Blackhole P150a offers substantial theoretical compute throughput for compatible low-precision AI workloads.

This makes it an attractive option for researchers exploring neural networks, transformer-based models, and specialized machine learning applications.

The 664 TFLOPS figure represents peak theoretical BLOCKFP8 performance, not guaranteed real-world application throughput.

32GB GDDR6 Memory with 512GB/s Bandwidth

Memory performance is essential for modern artificial intelligence applications.

The Blackhole P150a features:

• 32GB GDDR6 onboard memory
• 512GB/s memory bandwidth
• 256-bit memory bus
• 16 GT/s memory speed
• 180MB on-chip SRAM

This high-bandwidth memory architecture supports rapid movement of model parameters, tensors, and intermediate data during supported AI operations.

The 32GB memory capacity also provides additional flexibility for larger models compared with the 28GB Blackhole P100a.

Multi-Card AI Computing with 800Gbps QSFP-DD Connectivity

One of the most important advantages of the Tenstorrent Blackhole P150a is its dedicated high-speed interconnect capability.

The accelerator features four passive QSFP-DD ports, each providing 800Gbps connectivity.

These interfaces allow compatible Tenstorrent Blackhole P150a and P150b cards to communicate directly, supporting scalable multi-accelerator configurations.

Benefits of Multi-Card Connectivity

Scalable AI Infrastructure

Connect compatible Blackhole accelerators to build more powerful multi-card AI development environments.

Expanded Aggregate Memory

Multiple interconnected accelerators provide additional distributed memory capacity for supported workloads.

High-Speed Accelerator Communication

Dedicated 800Gbps interfaces support high-bandwidth communication between compatible Blackhole cards.

Flexible Research Configurations

Developers can experiment with distributed AI workloads and multi-device computing architectures.

QSFP-DD connectivity is intended for supported Tenstorrent Blackhole accelerator interconnections. It is not a standard Ethernet networking interface. Appropriate cables, compatible software, and supported topology are required.

Applications & Use Cases

1. Artificial Intelligence Development

Develop and experiment with modern AI applications using Tenstorrent’s open-source software ecosystem.

2. Large Language Models & Generative AI

Explore supported transformer architectures, language model inference, and generative AI applications.

3. Deep Learning & Neural Networks

Build and optimize supported neural network applications involving tensor operations and parallel computation.

4. Computer Vision & Image Processing

Research supported image classification, object detection, visual recognition, and AI-assisted image processing workloads.

5. AI Research Laboratories

An excellent platform for universities, research organizations, engineering institutions, and AI innovation laboratories.

6. Multi-Accelerator AI Workstations

Create scalable AI development configurations by connecting compatible Blackhole accelerators through dedicated QSFP-DD interconnects.

7. Custom AI Kernel Development

Explore low-level AI programming and optimize specialized computational kernels using Tenstorrent’s software tools.

Complete Technical Specifications

System Requirements & Compatibility

For proper installation and operation, users should prepare a compatible workstation with:

• CPU: Compatible x86-64 host processor
• Motherboard: PCIe 5.0 x16 slot recommended
• System Memory: At least 64GB recommended for development
• Storage: Sufficient SSD space for software, models, and datasets
• Operating System: Supported Linux environment, including compatible Ubuntu releases
• Power Supply: ATX 3.1 certified or better
• External Power: Compatible 12V-2x6 power connector
• Cooling: Adequate workstation ventilation
• Expansion Space: Dual-slot physical clearance

For multi-card installations, verify motherboard slot spacing, system power capacity, interconnect cables, and software support.

Why Choose the Tenstorrent Blackhole P150a?

1. Powerful AI Processing

With 664 TFLOPS peak BLOCKFP8 compute performance, the P150a is designed for demanding AI development workloads.

2. Larger 32GB Memory

The additional onboard memory compared with the P100a offers greater flexibility for memory-intensive applications.

3. Scalable Multi-Card Architecture

Four dedicated QSFP-DD ports enable supported multi-accelerator configurations.

4. Open-Source Development

Tenstorrent provides an open-source programming environment for researchers and developers.

5. Advanced RISC-V Architecture

The combination of Tensix cores and RISC-V processors offers a distinctive approach to artificial intelligence computing.

6. Workstation-Friendly Design

The active-cooled PCIe form factor allows installation in compatible desktop systems without requiring a dedicated rack-mounted server.

Frequently Asked Questions

Q1. What is the Tenstorrent Blackhole P150a?

The P150a is a high-performance PCIe AI accelerator designed for artificial intelligence development, machine learning, neural network research, and multi-card computing.

Q2. How much memory does the Blackhole P150a have?

It includes 32GB GDDR6 memory with 512GB/s memory bandwidth.

Q3. How powerful is the Blackhole P150a?

It provides up to 664 TFLOPS of theoretical BLOCKFP8 compute performance.

Q4. What is the difference between Blackhole P100a and P150a?

The P150a offers 32GB GDDR6 instead of 28GB, higher memory bandwidth, and four dedicated 800Gbps QSFP-DD interconnect ports for multi-card configurations.

Q5. Can multiple P150a cards work together?

Yes. Compatible Blackhole P150a and P150b accelerators can be interconnected using supported QSFP-DD cables and software configurations.

Q6. Does the Blackhole P150a support CUDA?

No. It uses Tenstorrent’s own software ecosystem and is not a drop-in replacement for NVIDIA CUDA hardware.

Q7. Is the P150a suitable for large language models?

It supports compatible AI and language model workloads. Actual model compatibility depends on available memory, software implementation, precision, and multi-card configuration.

Q8. Can the P150a be installed in a desktop PC?

Yes, provided the workstation meets the physical, electrical, cooling, and software requirements.

Q9. Does the P150a include active cooling?

Yes. The P150a includes active blower-style cooling. The P150b is the passive-cooled alternative intended for appropriately cooled server environments.

Buy Tenstorrent Blackhole P150a AI Accelerator in India

Upgrade your AI development infrastructure with the Tenstorrent Blackhole P150a – a powerful, scalable AI accelerator designed for next-generation computing.

Featuring 32GB GDDR6 memory, 664 TFLOPS peak BLOCKFP8 performance, PCIe 5.0 connectivity, and dedicated 800Gbps multi-card interconnects, the P150a is an excellent choice for specialized AI research and development.

Order Now | Request Best Price | Enquire for Availability | Contact for Bulk Orders

Product availability, delivery timelines, warranty, and final pricing are subject to seller confirmation.

SEO keywords

Primary keyword: Tenstorrent Blackhole P150a India

Secondary keywords: Buy Tenstorrent P150a India, Blackhole P150a 32GB Price, Tenstorrent AI Accelerator India, 664 TFLOPS AI Card, RISC-V AI Processor, PCIe 5.0 AI Accelerator, 32GB AI Computing Card, Tenstorrent Blackhole GPU Alternative, AI Research Hardware India.