Initial commit: TFGrid Economics documentation site
Some checks failed
Deploy to GitHub Pages / Deploy to GitHub Pages (push) Has been cancelled
Some checks failed
Deploy to GitHub Pages / Deploy to GitHub Pages (push) Has been cancelled
- Based on working minting_plan repository - Configured for threefold.info/economics deployment - Added ops documentation for server deployment - Updated baseUrl and URL configuration
This commit is contained in:
418
docs/node-economics/tf-bid-packages.md
Normal file
418
docs/node-economics/tf-bid-packages.md
Normal file
@@ -0,0 +1,418 @@
|
||||
---
|
||||
sidebar_position: 4
|
||||
---
|
||||
|
||||
# ThreeFold Bid Packages
|
||||
|
||||
ThreeFold commits to purchasing specific capacity packages at defined minimum specifications and pricing. These "bids" represent guaranteed demand that farmers can fulfill by providing qualifying capacity.
|
||||
|
||||
:::info What Are Bids?
|
||||
Bids are ThreeFold's standing offers to purchase specific types of slices. If your node meets the minimum requirements, you can fulfill these bids and receive the stated pricing.
|
||||
:::
|
||||
|
||||
---
|
||||
|
||||
## Standard Bid Packages
|
||||
|
||||
### Standard Compute Slice
|
||||
|
||||
**Target Workloads**: General web applications, APIs, microservices, development environments
|
||||
|
||||
**Minimum Requirements**:
|
||||
|
||||
| Specification | Minimum Value |
|
||||
|--------------|---------------|
|
||||
| **Performance** | 2,000 Passmark |
|
||||
| **Memory** | 4 GB |
|
||||
| **GPU Memory** | - |
|
||||
|
||||
**Pricing**:
|
||||
- **1.2 CC/month** per slice
|
||||
- Can scale based on reservation duration (see [Reservation Pricing](#reservation-duration-pricing) below)
|
||||
|
||||
**Total ThreeFold Commitment**:
|
||||
- **20,000 slices** at this specification
|
||||
- **Total monthly value**: 24,000 CC (~$3,072 at $0.128/CC)
|
||||
|
||||
**Who Can Provide**:
|
||||
Most node types qualify:
|
||||
- 3Node Home ✅
|
||||
- All Mini series (1, 2, 3) ✅
|
||||
- Large TFGrid nodes ✅
|
||||
- AI nodes (but better used for AI bids) ✅
|
||||
|
||||
---
|
||||
|
||||
### Memory-Heavy Slice
|
||||
|
||||
**Target Workloads**: Databases (PostgreSQL, MySQL, Redis), caching layers, in-memory processing
|
||||
|
||||
**Minimum Requirements**:
|
||||
|
||||
| Specification | Minimum Value |
|
||||
|--------------|---------------|
|
||||
| **Memory** | 8 GB |
|
||||
|
||||
**Pricing**:
|
||||
- **2.4 CC/month** per slice (2x standard compute price)
|
||||
- Can scale based on reservation duration
|
||||
|
||||
**Total ThreeFold Commitment**:
|
||||
- **2,000 slices** at this specification
|
||||
- **Total monthly value**: 4,800 CC (~$614)
|
||||
|
||||
**Who Can Provide**:
|
||||
Nodes with higher memory configurations:
|
||||
- Mini 2 ✅
|
||||
- Mini 3 ✅
|
||||
- Large TFGrid Node ✅ (512 GB RAM - excellent fit)
|
||||
- All AI nodes ✅
|
||||
|
||||
:::tip High-Memory Premium
|
||||
Memory-heavy slices command 2x pricing of standard compute. If your node has abundant RAM, prioritize these bids.
|
||||
:::
|
||||
|
||||
---
|
||||
|
||||
## AI & GPU Bid Packages
|
||||
|
||||
### AI Slice (Standard)
|
||||
|
||||
**Target Workloads**: AI inference, small model training, ML development
|
||||
|
||||
**Minimum Requirements**:
|
||||
|
||||
| Specification | Minimum Value |
|
||||
|--------------|---------------|
|
||||
| **Memory** | 124 GB |
|
||||
|
||||
**Pricing**:
|
||||
- **50 CC/month** per slice
|
||||
- Based on full-node AI configuration (e.g., AI 395+ new nodes)
|
||||
|
||||
**Total ThreeFold Commitment**:
|
||||
- **2,000 slices** at this specification
|
||||
- **Total monthly value**: 100,000 CC (~$12,800)
|
||||
|
||||
**Who Can Provide**:
|
||||
High-spec AI nodes:
|
||||
- Mini AI 1 ✅ (entry point)
|
||||
- Nvidia 6000 variants ✅
|
||||
- Tenstorrent 8x ✅
|
||||
|
||||
**Note**: This represents full-node access for AI workloads with shared GPU resources.
|
||||
|
||||
---
|
||||
|
||||
### AI Slice Big (Nvidia 1x)
|
||||
|
||||
**Target Workloads**: Professional AI training, large model inference, production ML pipelines
|
||||
|
||||
**Minimum Requirements**:
|
||||
|
||||
| Specification | Minimum Value |
|
||||
|--------------|---------------|
|
||||
| **Memory** | 124 GB |
|
||||
| **GPU Memory** | 96 GB (dedicated) |
|
||||
|
||||
**Pricing**:
|
||||
- **500 CC/month** per slice
|
||||
- Dedicated GPU access (not shared)
|
||||
|
||||
**Total ThreeFold Commitment**:
|
||||
- **50 slices** at this specification
|
||||
- **Total monthly value**: 25,000 CC (~$3,200)
|
||||
|
||||
**Who Can Provide**:
|
||||
- Nvidia 6000 (Single GPU) ✅
|
||||
- Nvidia 6000 (Dual GPU) ✅ (can provide 2 slices)
|
||||
|
||||
**Key Feature**: **96 GB dedicated GPU memory** - significantly larger than typical cloud offerings (16-40 GB)
|
||||
|
||||
---
|
||||
|
||||
### AI Slice Big (Nvidia 2x)
|
||||
|
||||
**Target Workloads**: Large-scale AI training, multi-GPU model parallelism, enterprise ML operations
|
||||
|
||||
**Minimum Requirements**:
|
||||
|
||||
| Specification | Minimum Value |
|
||||
|--------------|---------------|
|
||||
| **Memory** | 250 GB |
|
||||
| **GPU Memory** | 192 GB (dedicated) |
|
||||
|
||||
**Pricing**:
|
||||
- **1,000 CC/month** per slice
|
||||
- Dual dedicated GPU access
|
||||
|
||||
**Total ThreeFold Commitment**:
|
||||
- **50 slices** at this specification
|
||||
- **Total monthly value**: 50,000 CC (~$6,400)
|
||||
|
||||
**Who Can Provide**:
|
||||
- Nvidia 6000 (Dual GPU) ✅
|
||||
|
||||
**Key Feature**: **192 GB AI Memory** for massive model training and inference
|
||||
|
||||
---
|
||||
|
||||
### AI Slice Big (Tenstorrent 8x)
|
||||
|
||||
**Target Workloads**: Next-generation AI acceleration, cutting-edge ML research, high-throughput inference
|
||||
|
||||
**Minimum Requirements**:
|
||||
|
||||
| Specification | Minimum Value |
|
||||
|--------------|---------------|
|
||||
| **Memory** | 250 GB |
|
||||
| **GPU Memory** | 192 GB (AI memory) |
|
||||
|
||||
**Pricing**:
|
||||
- **1,000 CC/month** per slice
|
||||
- 8x Tenstorrent accelerators
|
||||
|
||||
**Total ThreeFold Commitment**:
|
||||
- **50 slices** at this specification
|
||||
- **Total monthly value**: 50,000 CC (~$6,400)
|
||||
|
||||
**Who Can Provide**:
|
||||
- Tenstorrent 8x nodes ✅
|
||||
|
||||
**Key Advantage**: **2,000 TPS** (vs 700 for Nvidia) - significantly higher throughput for specific AI workloads
|
||||
|
||||
:::note Tenstorrent vs. Nvidia
|
||||
While Nvidia is industry-standard with broader ecosystem support, Tenstorrent offers superior performance (2,000 TPS vs 700 TPS) for certain AI workloads at the same price point.
|
||||
:::
|
||||
|
||||
---
|
||||
|
||||
## Total Market Summary
|
||||
|
||||
**Combined ThreeFold Commitments**:
|
||||
|
||||
| Bid Package | Slices | Monthly Value | Annual Value |
|
||||
|-------------|--------|---------------|--------------|
|
||||
| Standard Compute | 20,000 | 24,000 CC | 288,000 CC |
|
||||
| Memory-Heavy | 2,000 | 4,800 CC | 57,600 CC |
|
||||
| AI Slice (Standard) | 2,000 | 100,000 CC | 1,200,000 CC |
|
||||
| AI Big (Nvidia 1x) | 50 | 25,000 CC | 300,000 CC |
|
||||
| AI Big (Nvidia 2x) | 50 | 50,000 CC | 600,000 CC |
|
||||
| AI Big (Tenstorrent 8x) | 50 | 50,000 CC | 600,000 CC |
|
||||
| **TOTAL** | **24,150** | **253,800 CC** | **3,045,600 CC** |
|
||||
|
||||
**Annual Value in USD** (at $0.128/CC): ~$389,837
|
||||
|
||||
:::tip Guaranteed Demand
|
||||
These bids represent **committed ThreeFold demand**. This is baseline capacity the platform will purchase, providing farmers with guaranteed revenue opportunities.
|
||||
:::
|
||||
|
||||
---
|
||||
|
||||
## Reservation Duration Pricing
|
||||
|
||||
The pricing listed above is **baseline monthly pricing**. Actual prices vary based on minimum reservation duration:
|
||||
|
||||
### Pricing Tiers by Duration
|
||||
|
||||
**Example pricing structure**:
|
||||
|
||||
| Min Reservation | Price Modifier | Example (Standard Compute) |
|
||||
|-----------------|----------------|---------------------------|
|
||||
| **1 hour** | 1.0x - 10.0x | 1.2 - 12.0 CC/month |
|
||||
| **1 day** | 0.9x - 8.0x | 1.08 - 9.6 CC/month |
|
||||
| **1 week** | 0.8x - 6.0x | 0.96 - 7.2 CC/month |
|
||||
| **1 month** | 0.7x - 4.0x | 0.84 - 4.8 CC/month |
|
||||
| **3 months** | 0.6x - 3.0x | 0.72 - 3.6 CC/month |
|
||||
| **1 year** | 0.5x - 2.0x | 0.60 - 2.4 CC/month |
|
||||
|
||||
**Key Principle**: Longer commitments = lower per-unit pricing
|
||||
|
||||
**Benefits**:
|
||||
- **For farmers**: More predictable income, can offer discounts
|
||||
- **For users**: Cost savings for long-term workloads
|
||||
- **For platform**: Improved capacity planning
|
||||
|
||||
:::info Flexible Market
|
||||
These are baseline bids. Farmers can set custom pricing when the marketplace allows farmer price-setting (coming soon). See [Implementation Roadmap](/implementation-roadmap#farmer-price-setting-capability).
|
||||
:::
|
||||
|
||||
---
|
||||
|
||||
## How to Qualify for Bids
|
||||
|
||||
### 1. Meet Minimum Specifications
|
||||
|
||||
Your node must meet or exceed the minimum requirements for the bid package:
|
||||
- Check [Node Specifications](/node-economics/node-specifications) for your hardware
|
||||
- Ensure adequate CPU, RAM, storage, and GPU (if applicable)
|
||||
|
||||
### 2. Maintain Uptime
|
||||
|
||||
High-quality, reliable capacity:
|
||||
- Target: 99%+ uptime
|
||||
- Stable internet connection
|
||||
- Proactive monitoring and maintenance
|
||||
|
||||
### 3. Proper Node Configuration
|
||||
|
||||
- Correct slice partitioning in node setup
|
||||
- Accurate resource reporting
|
||||
- Compatible with TFGrid v4 (when released) or v3 (currently)
|
||||
|
||||
### 4. Pricing Within Range
|
||||
|
||||
- Set pricing at or below bid price ceiling
|
||||
- Competitive with other farmers
|
||||
- Consider reservation duration discounts
|
||||
|
||||
---
|
||||
|
||||
## Beyond ThreeFold Bids
|
||||
|
||||
These bids represent **ThreeFold's committed demand**, not the total market:
|
||||
|
||||
### Additional Demand Sources
|
||||
|
||||
**End-User Applications**:
|
||||
- Applications built on TFGrid by partner teams
|
||||
- Direct user deployments
|
||||
- Enterprise customers
|
||||
|
||||
**Marketplace Transactions**:
|
||||
- Custom capacity requests
|
||||
- Bid-based capacity allocation (vast.ai model coming)
|
||||
- Private arrangements between users and farmers
|
||||
|
||||
**Partner Ecosystems**:
|
||||
- Integration with other platforms
|
||||
- White-label cloud offerings
|
||||
- Reseller channels
|
||||
|
||||
**Offtaker Agreements**:
|
||||
- Large-scale enterprise commitments
|
||||
- To be announced when commercially appropriate
|
||||
- Potentially significantly larger than baseline bids
|
||||
|
||||
:::note Market Growth
|
||||
As the ecosystem matures, demand will expand beyond these baseline bids. Early farmers establishing reputation and capacity will be well-positioned for growth.
|
||||
:::
|
||||
|
||||
---
|
||||
|
||||
## Strategic Considerations for Farmers
|
||||
|
||||
### Capacity Planning
|
||||
|
||||
**Question**: Which bid packages should I target?
|
||||
|
||||
**Factors to consider**:
|
||||
1. **Hardware you can afford**: Start with budget constraints
|
||||
2. **Market competition**: Standard compute is crowded; AI is underserved
|
||||
3. **Growth potential**: AI demand is expanding rapidly
|
||||
4. **Risk tolerance**: Higher investment (AI) = higher potential returns but more risk
|
||||
|
||||
### Investment Strategies
|
||||
|
||||
**Conservative**:
|
||||
- Focus on Mini 1-3 nodes
|
||||
- Target standard compute and memory-heavy bids
|
||||
- Lower investment, proven demand
|
||||
- Slower but steadier ROI
|
||||
|
||||
**Balanced**:
|
||||
- Mix of standard nodes and entry AI (Mini AI 1)
|
||||
- Diversify across bid types
|
||||
- Moderate investment and risk
|
||||
- Flexible capacity allocation
|
||||
|
||||
**Aggressive**:
|
||||
- Focus on high-end AI nodes (Nvidia 6000, Tenstorrent)
|
||||
- Target AI bid packages
|
||||
- High investment, high potential returns
|
||||
- Bet on AI growth trajectory
|
||||
|
||||
### Geographic Considerations
|
||||
|
||||
**Latency-Sensitive Workloads**:
|
||||
- Closer to major population centers = potential pricing premium
|
||||
- Edge computing use cases value low latency
|
||||
- Consider location when pricing
|
||||
|
||||
**Energy Costs**:
|
||||
- High electricity costs reduce profitability
|
||||
- Consider renewable energy for cost savings
|
||||
- Factor into pricing strategy
|
||||
|
||||
---
|
||||
|
||||
## Worked Example: Fulfilling Bids
|
||||
|
||||
### Scenario: You Have a Large TFGrid Node
|
||||
|
||||
**Your Node Specs**:
|
||||
- 36 CPU cores, 20,000 Passmark
|
||||
- 512 GB RAM
|
||||
- 4,000 GB SSD
|
||||
- Max 100 slices, typically provides 20
|
||||
|
||||
**Bid Fulfillment Strategy**:
|
||||
|
||||
**Option A: All Memory-Heavy**
|
||||
- Leverage your 512 GB RAM advantage
|
||||
- Provide 20 memory-heavy slices at 8 GB each = 160 GB used
|
||||
- Income: 20 × 2.4 CC = 48 CC/month minimum
|
||||
- Premium positioning due to high RAM
|
||||
|
||||
**Option B: Mixed Strategy**
|
||||
- 10 memory-heavy slices: 10 × 2.4 = 24 CC
|
||||
- 10 standard compute slices: 10 × 1.2 = 12 CC
|
||||
- Total: 36 CC/month minimum
|
||||
- More flexibility, broader market
|
||||
|
||||
**Option C: Target Higher Pricing**
|
||||
- Compete in marketplace when available
|
||||
- Offer custom pricing for longer reservations
|
||||
- Potential: 40-100 CC/month with market pricing
|
||||
|
||||
**Recommendation**: Option A or C - leverage your high RAM to target premium memory-heavy demand.
|
||||
|
||||
---
|
||||
|
||||
## Next Steps
|
||||
|
||||
### For New Farmers
|
||||
|
||||
1. **Review [Node Specifications](/node-economics/node-specifications)** - Choose hardware
|
||||
2. **Identify bid packages you can fulfill** - Match specs to bids
|
||||
3. **Calculate ROI with [Pricing Model](/node-economics/pricing-model)** - Ensure profitability
|
||||
4. **Acquire hardware and configure node** - Get started
|
||||
|
||||
### For Existing Farmers
|
||||
|
||||
1. **Evaluate which bids you currently fulfill** - Optimize allocation
|
||||
2. **Consider upgrades** - Add RAM for memory-heavy bids, or GPU for AI bids
|
||||
3. **Plan for certification** - Q4 2025 certified nodes coming
|
||||
4. **Set competitive pricing** - When marketplace allows
|
||||
|
||||
---
|
||||
|
||||
:::tip Community Resources
|
||||
- **[ThreeFold Forum](https://forum.threefold.io)** - Discuss strategies with other farmers
|
||||
- **[Telegram](https://t.me/threefoldfarmers)** - Real-time community support
|
||||
- **[Community Calls](https://bit.ly/tfcommunitycall)** - Weekly updates and Q&A
|
||||
|
||||
Share your experiences and learn from others building the decentralized cloud!
|
||||
:::
|
||||
|
||||
---
|
||||
|
||||
## Important Notes
|
||||
|
||||
**(*) Revenue Distribution**: All pricing listed is gross revenue. Farmers receive 80%, with 10% burned and 10% to ThreeFold operations.
|
||||
|
||||
**CC Acquisition**: Cloud Credits can be purchased at fiat equivalent of 1/1000 gram of gold ≈ $0.128 USD (Oct 8, 2024 pricing) or by converting 2 TFT per 1 CC.
|
||||
|
||||
**Pricing Flexibility**: Bids represent minimum commitments at stated prices. Marketplace implementation will allow farmers to set custom pricing for competitive differentiation.
|
||||
|
||||
**Duration Matters**: The pricing reflects monthly rates, but can vary significantly based on minimum reservation duration (hour, day, week, month, year).
|
Reference in New Issue
Block a user