despiegk
a00a487a53
Some checks failed
Deploy to GitHub Pages / Deploy to GitHub Pages (push) Has been cancelled
119 lines
8.6 KiB
Markdown
119 lines
8.6 KiB
Markdown
|
||
# Example Node Configurations (Q4 2025)
|
||
|
||
This document outlines example compute nodes, their specifications, estimated costs, performance profiles, and potential monthly income based on the **ThreeFold / GeoMind** model for decentralized AI, compute, and storage nodes.
|
||
|
||
Nodes are divided into **Certified** and **Non-Certified** types, each with defined hardware specs and income models.
|
||
|
||
---
|
||
|
||
## ⚙️ Node Specifications
|
||
|
||
| **Type** | **Estimate Cost (USD)** | **CPU Cores** | **Passmark** | **MEM (GB)** | **SSD (GB)** | **AI (GB)** | **TPS** |
|
||
| --------------------------------- | ----------------------- | ------------- | ------------ | ------------ | ------------ | ----------- |
|
||
| **Non-Certified** | | | | | | | |
|
||
| 3Node Home | $500 | 16 | 22,000 | 32 | 1,000 | – | – |
|
||
| Large TFGrid Node | $2,000 | 36 | 20,000 | 512 | 4,000 | – | – |
|
||
| Large TFGrid Node 2 | $1,200 | 36 | 20,000 | 128 | 2,000 | – | – |
|
||
| **Certified (new batch Q4 2025)** | | | | | | | |
|
||
| Mini 1 | $350 | 16 | 16,000 | 32 | 1,000 | – | – |
|
||
| Mini 2 | $800 | 24 | 25,000 | 64 | 2,000 | – | – |
|
||
| Mini 3 | $1,000 | 24 | 25,000 | 128 | 4,000 | – | – |
|
||
| Mini AI 1 | $2,000 | 32 | 50,000 | 128 | 4,000 | 128 | 20 |
|
||
| Nvidia 6000 (1x) | $15,000 | 24 | 35,000 | 128 | 4,000 | 96 | 350 |
|
||
| Nvidia 6000 (2x) | $30,000 | 24 | 35,000 | 128 | 8,000 | 192 | 700 |
|
||
| Tenstorrent 8x | $30,000 | 24 | 35,000 | 128 | 8,000 | 192 | 1,000 |
|
||
|
||
---
|
||
|
||
## Node Economics
|
||
|
||
### Hardware Cost and Income Potential
|
||
|
||
| **Model** | **Max # Slices** | **HW Cost / month (5 Y)** | **Min Income / Month** | **Max Income / Month** | **Nr Slices** |
|
||
| ---------------- | ---------------- | ------------------------- | ---------------------- | ---------------------- | ------------- |
|
||
| Mini 1 | 25 | 8 | 10.92 | 126 | 7 |
|
||
| Mini 2 | 25 | 13 | 23 | 270 | 15 |
|
||
| Mini 3 | 25 | 17 | 48.36 | 558 | 25 |
|
||
| Mini AI 1 | 4 | 33 | 52 | 600 | 4 |
|
||
| Nvidia 6000 (1x) | 1 | 250 | 390 | 4,500 | 1 |
|
||
| Nvidia 6000 (2x) | 2 | 500 | 780 | 9,000 | 2 |
|
||
| Tenstorrent 8x | 2 | 500 | 780 | 9,000 | 2 |
|
||
|
||
> 💡 **Income split:**
|
||
> 80 % to hoster, 10 % burned, 10 % to ThreeFold (protocol).
|
||
|
||
---
|
||
|
||
## Cloud Credit (CC) Pricing
|
||
|
||
| **Currency Basis** | **Value** |
|
||
| ----------------------- | -------------------------------- |
|
||
| 1 CC = 1/1000 g of gold | ≈ 0.128 USD *(as of Oct 8 2025)* |
|
||
| Equivalent | 2 TFT (TF token units) |
|
||
|
||
---
|
||
|
||
## Pricing for 1 Slice
|
||
|
||
| **Cores** | **Perf** | **Mem (GB)** | **SSD (GB)** | **TPS** | **Min Price / Month (CC)** | **Max Price / Month (CC)** | **Min Price / Hour (CC)** | **Max Price / Hour (CC)** |
|
||
| --------- | -------- | ------------ | ------------ | ------- | -------------------------- | -------------------------- | ------------------------- | ------------------------- |
|
||
| 2.3 | 3,143 | 4 | 143 | 0 | 1.2 | 12 | 0.0033 | 0.0250 |
|
||
| 1.8 | 1,000 | 25.4 | 200 | 0 | 7.6 | 76.2 | 0.0212 | 0.1588 |
|
||
| 1.8 | 1,000 | 6.2 | 100 | 0 | 1.9 | 18.6 | 0.0052 | 0.0388 |
|
||
| 2.3 | 2,286 | 4 | 143 | 0 | 1.2 | 12 | 0.0033 | 0.0250 |
|
||
| 1.6 | 1,667 | 4 | 133 | 0 | 1.2 | 12 | 0.0033 | 0.0250 |
|
||
| 1.0 | 1,000 | 5.0 | 160 | 0 | 1.5 | 14.9 | 0.0041 | 0.0310 |
|
||
| 8.0 | 12,500 | 31 | 1,000 | 5 | 10 | 100 | 0.0278 | 0.2083 |
|
||
| 24 | 35,000 | 124 | 4,000 | 350 | 300 | 3,000 | 0.8 | 6.3 |
|
||
| 12 | 17,500 | 62 | 4,000 | 350 | 300 | 3,000 | 0.8 | 6.3 |
|
||
| 12 | 17,500 | 62 | 4,000 | 1,000 | 300 | 3,000 | 0.8 | 6.3 |
|
||
|
||
> **Note:**
|
||
> These calculations assume **no CPU oversubscription** — in real deployments, users can utilize higher effective CPU capacity.
|
||
|
||
---
|
||
|
||
## Bid Tiers (ThreeFold Willing to Set)
|
||
|
||
| **Type** | **# Nodes** | **Min Perf** | **MEM (GB)** | **GPU (GB)** | **Price / Month (CC)** |
|
||
| --------------------------- | ----------- | ------------ | ------------ | ------------ | ---------------------- |
|
||
| Std Compute Slice | 20,000 | 2,000 | 4 | 1.2 | |
|
||
| Mem Heavy Slice | 2,000 | 2,000 | 8 | 2.4 | |
|
||
| AI Slice | 2,000 | 25,000 | 124 | 50 | |
|
||
| AI Slice Big Nvidia 1x | 50 | 35,000 | 124 | 96 | 500 |
|
||
| AI Slice Big Nvidia 2x | 50 | 35,000 | 250 | 192 | 1,000 |
|
||
| AI Slice Big Tenstorrent 8x | 50 | 35,000 | 250 | 192 | 1,000 |
|
||
|
||
**Total:**
|
||
|
||
* **24,150 slices**
|
||
* **253,800 GB total memory**
|
||
* **≈ 3,965,625 CC monthly potential (≈ $507,600)**
|
||
|
||
> *Based on new AI 395+ nodes. Dedicated memory & GPU per slice (full node model).*
|
||
|
||
---
|
||
|
||
## Pricing Logic
|
||
|
||
* Bids can be set per **1 hour**, **1 month**, or longer durations.
|
||
* Prices adjust dynamically based on reservation length (shorter = higher price).
|
||
* Reflects **minimum reservation duration** logic from Mycelium / ThreeFold’s marketplace.
|
||
|
||
---
|
||
|
||
## Summary Explanation
|
||
|
||
* Each **node** is a physical or certified compute machine that can be divided into **slices** — independent virtual allocations (e.g., 2 cores, 4 GB RAM, 143 GB SSD = 1 slice).
|
||
* **Certified Nodes** (Mini, AI, GPU variants) are hardware validated for stable and efficient participation in the decentralized grid.
|
||
* Each node has:
|
||
|
||
* A **capital cost** (hardware)
|
||
* A **hardware amortization** over 5 years (monthly cost)
|
||
* A potential **monthly income** based on slice sales (compute / AI / storage)
|
||
* **Hosters** receive 80 % of income, 10 % is protocol burn, 10 % to ThreeFold.
|
||
* The income range shows **minimum guaranteed** and **maximum potential** usage scenarios.
|
||
* **Cloud Credits (CC)** are the internal accounting unit pegged to gold, maintaining stability and fairness.
|
||
|