Dynamic Fees
The PSM fee is not static at 30%—it adjusts dynamically based on protocol conditions to optimize for peg stability and user experience.
Why Dynamic Fees?
Static fees can't respond to changing market conditions:
Below peg: Need stronger buy pressure (higher fees)
Above peg: Can reduce burden on users (lower fees)
High backing: Can afford lower fees (treasury is strong)
Low backing: Need higher fees (treasury needs support)
Market stress: Increase fees to defend peg
Market calm: Decrease fees to improve UX
Dynamic adjustment creates optimal balance at each moment.
Factors Influencing Fee Rate
The AI management system monitors multiple inputs:
spBNB:BNB Peg Ratio
Current peg status determines urgency:
1.05+
-5 to -10%
Less need for support, reward users
1.01-1.05
0% (baseline)
Healthy peg, standard fee
0.99-1.01
0 to +5%
At peg, slight precaution
0.95-0.99
+5 to +15%
Below peg, increase support
<0.95
+15 to +25%
Critical, maximum support
Example: At 0.93 TWAP, fee might be 40-45% instead of 30%
Treasury Backing Ratio
Backing level affects fee needs:
>110%
-5 to -10%
Strong, can reduce fees
105-110%
0%
Healthy baseline
100-105%
+5 to +10%
Tight, need more backing
95-100%
+10 to +20%
Concerning, increase urgently
<95%
+20 to +30%
Critical, maximum fees
Example: At 98% backing, fee might be 40% instead of 30%
SPAI Market Conditions
SPAI price and volatility impact optimal fees:
High SPAI Price:
Users are in profit
Can afford higher fees
Increase fees moderately
Low SPAI Price:
Users are underwater
Lower fees to maintain participation
Decrease fees moderately
High Volatility:
Unpredictable conditions
Increase fees for caution
Build treasury buffer
AI Optimization Algorithm
The AI doesn't just react—it optimizes:
Real-Time Monitoring
Constantly analyzes:
Peg deviation magnitude and duration
Treasury backing trajectory
Market volatility metrics
Claim frequency patterns
User behavior changes
Predictive Modeling
Projects outcomes of different fee levels:
If fee = 35%, expect 20% fewer claims
Fewer claims = less revenue but better user sentiment
More claims at 30% = better revenue but possible negative feedback
Optimal = balance between revenue and retention
Parameter Bounds
AI operates within safety limits:
Minimum fee: 15-20% (ensures base revenue)
Maximum fee: 50-60% (prevents user exodus)
Adjustment speed: Max 5-10% change per adjustment
Adjustment frequency: Once per day/epoch minimum
Prevents AI from setting extreme fees that break protocol.
Multi-Objective Optimization
Balances competing goals:
Maximize peg stability ⚖️ Minimize user cost
Maximize treasury growth ⚖️ Maximize user retention
Maximize short-term revenue ⚖️ Maximize long-term health
No single objective dominates—finds equilibrium.
Example Scenarios
Scenario 1: Healthy Growth
TWAP: 1.03 (expanding)
Backing: 107% (strong)
SPAI price: Up 40% from launch
Claims: High volume, frequent
AI Decision:
Baseline 30% fee appropriate
Slight decrease to 27-28% to reward users
Builds goodwill during good times
Scenario 2: Peg Stress
TWAP: 0.92 (well below peg)
Backing: 101% (barely above minimum)
SPAI price: Down 30% from peak
Claims: Moderate volume, users nervous
AI Decision:
Increase fee to 40-45%
Need maximum buy pressure for spBNB
Trade-off: Users claim less, but claims that happen provide stronger support
Scenario 3: Post-Crisis Recovery
TWAP: 0.97 (recovering toward peg)
Backing: 98% (below target)
SPAI price: Stabilizing after drop
Claims: Low volume (users hesitant)
AI Decision:
Moderate increase to 35%
Need backing but don't want to discourage participation
Balance recovery with user retention
User Implications
Check Fee Before Claiming
Always verify current rate:
Display in UI before claim button
Calculate net proceeds
Compare to previous claims
Decide if timing is optimal
Don't assume 30% is always current rate.
Timing Claims
Optimize claim timing around fee changes:
When fees are low:
Claim more frequently
Compound more aggressively
Take advantage of favorable rates
When fees are high:
Delay claims if possible
Wait for conditions to improve
Reduce claim frequency
Understanding the Tradeoff
Higher fees = stronger protocol = better long-term yields
Would you rather:
30% fee with sustainable protocol
0% fee with protocol that dies in 2 months
Dynamic fees optimize for longevity.
Transparency Requirements
For dynamic fees to work, users need visibility:
Current Fee Display
Real-time rate shown in UI
Updated automatically
Clear percentage display
Net proceeds calculator
Historical Fee Data
Chart showing fee changes over time
Correlation with peg/backing metrics
Explanation of major adjustments
Average fee over periods
Fee Change Notifications
Advance notice when possible (e.g., "Fee will adjust in 6 hours")
Reasoning provided (e.g., "Due to TWAP below 0.95")
Expected duration (e.g., "Temporary measure")
AI Explainability
Dashboard showing inputs AI is monitoring
Current values of key metrics
How they're affecting fee calculation
When next adjustment expected
Comparing Dynamic vs Static Fees
Static 30% Fee Protocol
Pros:
Predictable for users
Simple to understand
No surprise changes
Cons:
Can't respond to crises
Overly burdensome during good times
Suboptimal for most conditions
Dynamic 15-50% Fee Protocol (SPAI)
Pros:
Responds to market conditions
Optimal for current state
Balances multiple objectives
AI learns and improves
Cons:
Less predictable
Requires user monitoring
More complex system
Trust in AI required
The added complexity is worth the optimization benefits.
Last updated