AI compute. Cheaper by design.

Solana-native AI infrastructure

Designed to run regardless of market cycles.

Capital Yield Compute Revenue

Compounds with demand.

Capital Providers

Yield plus infrastructure exposure

Deploy capital through Primis. Base staking yield plus revenue from real AI workloads.

Architecture designed for institutional-scale allocations.

  • Base yield preserved
  • Usage revenue distributed
  • Liquid positions
primis.protocol/portfolio v1.2.4
Connected
Epoch 421 Last update: 3h ago Revenue settles per epoch
Total Earned Since deposit · 47 days
+0.0 SOL
All values settle per epoch
Yield from staking +0.0 SOL
Revenue from compute +0.0 SOL
Effective APY
—
Staking + usage
Staked
0 SOL
Principal
Network Util.
63%
Current epoch
Yield Allocation
This epoch Last 30d All time
Staker: 70%
Subsidy: 20%
Reserve: 10%
Earnings · Last 30 days
Recent Revenue Derived from network usage
fine-tune #f8k2
+0.24 SOL
embed-batch #a3m9
+0.18 SOL
train #t1x7
+0.82 SOL
AI Teams

Compute with predictable pricing

Access subsidized compute through a standard API. No contracts, no minimums.

Infrastructure designed for sustained workloads.

  • Below-market pricing
  • Standard REST API
  • Monthly spend capped
  • No burst pricing
terminal
project-atlas global · A100/H100 pooled
Max hourly rate enforced
$ primis jobs submit --type fine-tune --model llama-3-8b

Allocating GPUs...
Scheduling workload...
Subsidy applied

{
"job_id": "job_8f7d3k2l",
"status": "running",
"estimated_cost": "$12.40"
}

Job running.
Job Status No surprise overruns
GPU allocation 4x A100
Est. completion ~2h 15m
Subsidy rate Applied
Budget status Within cap

Early access

Primis is onboarding a limited number of teams.