Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Fix the driver behind crashes, sound loss and screen glitches3Repair Windows errors before they cause bigger problemsiTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Micron says it is shipping production-capable HBM3E 12-high memory to key industry partners for qualification, and its product page says samples are available. That means the 36GB stack is moving through the AI hardware supply chain—not launching as a retail memory module for PCs or standalone graphics cards. Micron has named AMD Instinct MI350 Series as an accelerator family that integrates the memory.
What “shipping” means for Micron’s 36GB HBM3E
Micron’s HBM3E blog says the company is “now shipping production-capable HBM3E 12-high to key industry partners for qualification across the AI ecosystem.” Its product page describes the 36GB 12-high cube as production-capable and available, and says samples are available now. These are signs of partner qualification and sampling, not evidence of a consumer retail launch or of broad volume availability.
HBM (high-bandwidth memory) is designed to be integrated alongside a processor in a packaged accelerator. The 12-high label refers to the vertically stacked memory configuration; the 36GB figure is capacity per placement. It is a component for accelerator manufacturers and platform builders, not a plug-in DIMM or a standalone GPU upgrade a PC owner can install.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Micron HBM3E 12-high specifications
| Measure | Micron 12-high 36GB | Context |
|---|---|---|
| Capacity | 36GB per placement | Micron current product and blog pages |
| Bandwidth | More than 1.2 TB/s | Micron current product and blog pages; this is the stack figure, not the total bandwidth of an accelerator platform |
| Pin speed | Greater than 9.2 Gb/s | Micron current product and blog pages |
| Capacity comparison | 50% more than 24GB 8-high | Micron’s stated comparison in its HBM3E blog |
| Power comparison | Up to 20% lower power | Micron’s 2025 comparison against a cited competing 8-high 24GB product; this is a vendor-reported maximum, not a universal system-level result |
Bandwidth figures at different levels should not be conflated. Micron gives more than 1.2 TB/s for a 36GB HBM3E placement, while its MI350 announcement gives up to 8 TB/s for the accelerator platform. The latter is a platform-level figure and should not be read as the bandwidth of one memory stack.
#1 Best Overall
- NVIDIA Volta GV100 Architecture — 4,608 CUDA Cores, 640 1st-Gen Tensor Cores delivering 14 TFLOPS FP32 and 112 TFLOPS deep learning performance for AI training, inference, HPC, and scientific computing workloads
- 32GB HBM2 ECC Memory — 900 GB/s Bandwidth — High-bandwidth memory on a 4096-bit bus with ECC error correction provides the memory capacity and throughput required for the largest AI models, simulations, and datasets
- PCIe 3.0 x16 Interface — 250W TDP — Standard PCIe Gen3 connectivity with passive cooling designed for enterprise rack server deployment in HPE ProLiant, Dell PowerEdge, and Supermicro platforms with adequate chassis airflow
- NVLink — Scale to 96GB Unified Memory — Connect two V100 GPUs via NVLink at 300 GB/s bi-directional bandwidth to scale GPU memory from 32GB to 96GB for larger AI training and HPC workloads
- Multi-Precision Computing — Supports FP64 (7 TFLOPS), FP32 (14 TFLOPS), FP16 (112 TFLOPS) and INT8 precision modes for flexible deployment across training, inference, and scientific simulation workloads
Which AI accelerators use the 36GB stack?
AMD Instinct MI350 Series
The clearest named integration is AMD Instinct MI350 Series. In a June 12, 2025 announcement, Micron said its HBM3E 36GB 12-high is integrated into upcoming MI350 Series solutions. Micron reported 288GB of HBM3E and up to 8 TB/s of bandwidth on MI350 GPU platforms, and said one MI350 GPU can support up to 520 billion parameters. These are Micron’s platform claims; they do not by themselves establish the availability date or configuration of every MI350 product.
Micron’s 288GB platform figure is the total it reports for the MI350 GPU platform, whereas 36GB is capacity per memory placement. The two figures describe different levels of the system.
Rank #2
- High-Performance AI Processing: The MX3 is designed to handle the most demanding AI computer vision workloads, delivering exceptional performance and efficiency.
- Flexible Integration: The MX3 can be easily integrated into your existing systems via its M.2 M-key form factor and support for Linux operating systems.
- Energy Efficient: The MX3 is designed to provide high performance while minimizing power consumption.
- Comprehensive Software Development Kit (SDK): The MX3 is supported by a comprehensive SDK that simplifies development and deployment.
- Hardware compatability: The MX3 is compatible with the PCI-SIG M.2 M-key 2280 Specification. It can be used with the Raspberry Pi 5 with a M-key 2280 HAT.
Other accelerators
Micron describes partner qualification across the AI ecosystem, but the information available here does not name additional accelerator products using this exact 36GB 12-high stack. Do not assume that a GPU supports it just because the GPU uses HBM3E: memory generation, stack capacity, package design and qualification are platform-specific.
Why the extra capacity matters for AI workloads
Micron positions HBM3E for generative-AI training and inference, deep learning and high-performance computing. More memory capacity can allow a larger model or more of its data to fit close to the processor, reducing reliance on CPU offload or transfers between GPUs. Micron says 36GB can let a model such as Llama 2 with 70 billion parameters run on a single processor. That is a vendor-described capability, not a guarantee that every 70-billion-parameter model, precision, batch size or workload will fit or perform equally well.
Rank #3
- ✅Powered by 26 Tera-Operations Per Second (TOPS) Hailo-8 AI Processor. 2.5W typical power consumption
- ✅Scalable, enabling simultaneous processing of multi-streams & multi-models
- ✅Enabling real-time, low latency and high-efficiency AI inferencing on the edge devices
- ✅Supports TensorFlow, TensorFlow Lite, ONNX, Keras, Pytorch frameworks
- ✅Supports Linux and Windows. Supports the temperature range of -40°C to 85°C
The practical benefit depends on the complete accelerator configuration and workload. Capacity is only one factor: software support, model precision, memory bandwidth, compute resources and how the platform distributes data also affect performance. Micron’s 50% capacity comparison is specifically against its cited 8-high 24GB reference, rather than every HBM product on the market.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Can you buy Micron HBM3E 36GB for a PC or graphics card?
No consumer purchase path is established for this component. The product is intended for integration into accelerator packages, and Micron’s stated availability is partner qualification and sample availability. It is not a user-installable PC memory module or an ordinary retail graphics-card upgrade. For buyers evaluating AI hardware, the relevant product is a complete compatible accelerator platform, such as an appropriately configured MI350 system, rather than the memory stack by itself.
Quick Recap
Best Value
- Professional AI & Creator Workstation: AMD Radeon AI PRO R9700 GPU with 32GB GDDR6 is engineered for AI development, professional content creation, and compute-intensive workloads.
- Massive 32GB Memory Capacity: 32GB of GDDR6 memory on a 256-bit bus provides ample bandwidth for large AI models, 8K video editing, and complex 3D rendering.
- Advanced RDNA 4 with AI Accelerators: 64 Compute Units with 3rd Gen Ray Tracing and dedicated 2nd Gen AI Accelerators for groundbreaking AI performance and visual computing.
- Professional Blower Cooling: Efficient single blower design exhausts heat directly out of the chassis, ideal for multi-GPU workstation and server configurations.
- Enterprise-Grade Thermal Solution: Vapor chamber heatsink with industrial Honeywell PTM7950 thermal interface material ensures reliable cooling under sustained professional loads.
Rank #4
- 48GB AI graphics accelerator
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
