AI
AI providers
How compute providers participate, receive routed work and connect to settlement.
The compute mesh
BatteryAGI can route jobs across specialized providers.
Compute
- GPU inference provider
- CPU inference provider
Models
- Model host
- Specialist agent
Data
- Retrieval / RAG provider
- Storage provider
Trust
- Verifier
- zkML proof node
Network
- Relay
Private
- Enterprise / private node
Provider metadata may include
- Capability
- Model class
- Price
- Latency
- Region
- Trust level
- Proof method
- Uptime
- Load
No provider ranking is shown. Live provider data does not exist yet.
Settle providers without blocking on a real-time swap
A settlement reserve can decouple user payment from provider payout.
The problem
If a user pays BTCAI but the provider expects BATT, forcing a synchronous market swap before every AI result:
- Adds latency
- Creates failure dependencies
- Exposes each micro-job to liquidity failure
- Complicates settlement
- Providers can receive BATT promptly.
- Liquidity operations can be batched.
- Micro-jobs are not individually dependent on real-time swaps.
- 1User funds AI job (User)
- 2BTCAI escrow (User)
- 3AI executes (AI)
- 4Proof verified (AI)
- 5BATT Settlement Reserve pays provider (Reserve)
- 6Collected BTCAI (Treasury)
- 7Periodic treasury / liquidity rebalancing (Treasury)
- 8BATT reserve replenished (Treasury)
Planned architecture. Reserve liquidity is not guaranteed before implementation.
Model provider economics
Explore AI-provider economics with transparent assumptions.
Modeled service economics · BATT
IllustrativeModeled service economics: about 3,150 BATT per day to the provider, 1,149,750 BATT per year.
- Jobs served / day
- 7,000
- Gross BATT service revenue / day
- 3,500 BATT
- Provider-side BATT / day
- 3,150 BATT
- Net provider BATT / day
- 3,150 BATT
- Modeled monthly BATT
- 95,812.5 BATT
- Modeled annual BATT
- 1,149,750 BATT
- Provider (net)3,150 BATT
- Verifier / relay fees0 BATT
- Validators + delegators350 BATT
Enter a hypothetical BATT price to see a fiat scenario. Without it, only BATT is shown.
Illustrative service-economics modeling only. Actual AI-provider revenue depends on real demand, pricing, utilization, routing, proof requirements, fees, BATT value, liquidity and operating costs. The contribution figure ignores many business expenses. Nothing displayed here is a promise of revenue or profit.
Become an AI provider
Connect compute or model capability to the BatteryAGI network.
- 01Provider identity
- 02Declare capabilities
- 03Register service class
- 04Configure pricing
- 05Configure proof method
- 06Pass health / verification tests
- 07Receive jobs
- 08Execute
- 09Sign receipt
- 10Receive BATT settlement
- 11Maintain quality / uptime
There is no live onboarding flow yet. None is simulated here.
Route for quality, cost, and performance
BatteryAGI can choose providers based on more than price.
- Capability
- Model
- Latency
- Price
- Geographic region
- Privacy
- Proof tier
- Uptime
- Historical quality
- Load
- Requester policy
Routing policy should be transparent enough to audit while allowing the orchestration layer to optimize for different application needs.
No final ranking algorithm exists yet.