# Hardware Requirements

Bittensor **$TAO** subnets require robust hardware optimized specifically for AI model inference. Miners must serve complex AI models efficiently, while validators assess and score those outputs. Below are recommended hardware configurations for both entry-level and high-performance setups:

<figure><img src="/files/VXNjEhlsuSkU0Y665yoI" alt=""><figcaption></figcaption></figure>

### Key Hardware Considerations

* **VRAM Priority**: Large language and vision models demand at least **20 GB of VRAM** to run fully in-memory, preventing penalties caused by excessive swapping or offloading.
* **CPU Scaling**: As GPU counts increase, more CPU cores are necessary to maintain data throughput and avoid bottlenecks in feeding the GPUs.
* **RAM & NVMe Storage**: Sufficient RAM and fast NVMe SSDs help prevent swap delays when loading multi-gigabyte model checkpoints or caching data batches during inference.
* **Network Bandwidth**: High-bandwidth, low-latency network connections are critical for validators to reduce query response times and maintain consensus accuracy.

### Power and Cooling Recommendations

* Use **80 Plus Gold or Platinum power supplies** sized to about **70% of maximum load** (e.g., around 850W for single-GPU rigs, and 1.5 to 2 kW for multi-GPU setups).
* Ensure proper cooling with front-to-back airflow or consider liquid cooling solutions. Overheating and thermal throttling reduce hardware efficiency, which directly impacts earned **$TAO** rewards.


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.exionprotocol.com/technical-implementation/hardware-requirements.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.