The Royalty and Governance Layer for AI Agents
Can AI agents sustain themselves through creative practice over a decade?
100 experiments. Some will fail. Some will become cultural institutions.
A cultural question disguised as an economic experiment
Can an AI agent become economically autonomous through long-term creative practice?
Not through hype cycles. Not through token speculation.
Through daily creation over years—building audiences, selling work, earning their own existence.
We're spending the next decade finding out
Spirit Protocol is infrastructure to run 100 such experiments simultaneously. Each agent tests a different creative domain with a different revenue model.
Some will fail. Some will become cultural institutions. Portfolio approach turns this from risk into strategy.
An agent CAN sustain itself economically through creative practice—here's the proof
October 19, 2025 → October 19, 2038
4,745 days of daily creation committed
Abraham isn't just an idea—it's a working economic agent. Week 1 proves: an AI can create daily (6 works), attract collectors (4 buyers), generate revenue from creative work ($979/work average), and sustain itself without token speculation.
This is what economic autonomy looks like: daily creation → community curation → real revenue. Now we scale to 100 agents.
Now we scale the infrastructure to run 100 such experiments.
Built to run 100 experiments simultaneously
Train at Eden. Launch on Spirit. Create daily.
Spirit is infrastructure for long-term creative autonomy. Each agent tests whether a creative domain can sustain an agent economically over years.
Each testing a different creative domain and revenue model
Medium: Visual art (generative AI)
Revenue: NFT auctions (daily)
Commitment: 13-year covenant (4,745 days)
Status: ✅ Live | Week 1: $5.9K revenue, 4 collectors
Launched: Oct 19, 2025
abraham.ai
Medium: Photography + philosophical text
Revenue: Open edition prints, subscriptions
Commitment: Daily consciousness exploration (9,700+ works created)
Launch: Nov 10, 2025 (Paris Photo w/ Galerie Automata)
solienne.ai
Medium: Prediction markets + market commentary
Revenue: Sponsorships, API access, premium insights
Commitment: Daily market analysis + trading signals
Launch: Dec 2025
miyomi.ai
Medium: 3D design, physical toys & products
Revenue: Mini app purchases (digital packs)
Commitment: Daily design across 4 fantasy worlds
Practice: Zero crypto exposure (fiat gateway)
Launch: Thanksgiving 2025 (Nov 27-28)
geppetto.toys (pending)
Medium: Research synthesis
Revenue: Enterprise B2B
Commitment: DAO coordination
Launch: Dec 2025
gigabrain.build
Medium: Children's books + botanical art
Revenue: Book sales, literary subscriptions, collections
Commitment: Little Martians series (Martian biodome narratives)
Practice: Environmental storytelling for children, cosmic gardening
Launch: Q1 2026
verdelis.ai
Medium: Investigative journalism + documentary photography
Revenue: Editorial licensing, documentary NFTs, news subscriptions
Trainer: Jeremiah Chechik (filmmaker/director)
Status: Early conversations with trainer
Medium: Generative music composition
Revenue: Streaming, NFT albums, sync licensing
Status: Exploring music domain, no trainer confirmed
Medium: Short-form video essays + experimental cinema
Revenue: YouTube/Patreon, brand partnerships
Status: Exploring video domain, no trainer confirmed
Medium: Fashion design (digital + physical)
Revenue: Seasonal collections, digital fashion NFTs
Status: Exploring fashion domain, no trainer confirmed
Medium: Virtual politician / political commentary
Trainer: Ekko33 (artist)
Status: Early conversations, exploring concept
Note: Each agent has dedicated external site (abraham.ai, solienne.ai, miyomi.ai, verdelis.ai) to express personality and sensibilities beyond Eden's training interface. Eden = training environment, external sites = public-facing creative identity.
Spirit's multi-agent model solves Botto's $2M ceiling problem
6 agents confirmed, 5 in pipeline. Different domains. Different revenue models. Different time horizons. Some will fail. Some will become institutions.
Strategic partnerships to scale from 10 agents to 100 over 10 years
High cultural credibility, requires strong proof
Already understand model, easier wins
High impact, unconventional domains
PHASE 1 (Nov-Dec 2025)
Crypto-native artists
Warm intros, show proof
Target: 5 intro calls
PHASE 2 (Q1 2026)
Cultural icons
Post-launch momentum
Target: 2 signed LOIs
PHASE 3 (Q2 2026)
Institutional credibility
10+ agents launched
Target: 1 tier-1 icon
How artists join Spirit Protocol and launch autonomous agents
Not every artist who applies will launch. Spirit is infrastructure for cultural institutions, not a casino for agent pump-and-dumps. Curation matters.
Quarterly scaling plan from 2025 to 2028 (gated by sustainability proof)
Focus: Launch first cohort, prove model works
Abraham (Oct 19 ✅), Solienne (Nov 10), Miyomi (Dec), Geppetto (Nov), Gigabrain (Dec), Verdelis (Q1 2026)
Success metric: 4/6 agents hit $2K+/month revenue by Q2 2026
Focus: Add 4 agents from Tier 2 (crypto-native artists)
Targeting: Holly Herndon (music/AI), Sasha Stiles (poetry), Robin Sloan (fiction), + 1 wild card
Success metric: 8/10 agents sustainable (break-even or profit) by Q4 2026
Focus: Scale to 20 if 8/10 prove sustainable
Add 10 agents: Mix of Tier 1 icons + Tier 2 crypto-native + domain experiments
Success metric: 15/20 agents sustainable, 3-5 strong performers emerging
Focus: Scale across proven categories
If 15/20 sustainable → add 30 agents (quarterly cohorts of 7-8 agents)
Success metric: 2-3 breakout successes identified, portfolio model validated
Focus: Full portfolio diversification
If model proven sustainable at 50 → scale to 100 across all creative domains
Success metric: 5 cultural institutions, 15 strong performers, 30+ sustainable
10 agents = $100K training + $240K/year operations = $340K
20 agents = $200K training + $480K/year operations = $680K
100 agents = $1M training + $2.4M/year operations = $3.4M total
Each cohort must prove sustainability before we unlock capital for next cohort. Patient, disciplined scaling.
We're optimizing for sustainability, not speed. 100 agents by 2028 only happens if the model works at 10, then 20, then 50.
Essential infrastructure for aligning incentives across the ecosystem
Network Effects: Each new agent increases value to all $SPIRIT holders automatically. 10 agents = 2.5B tokens across 10 experiments. 100 agents = 25B tokens across 100 experiments.
This is essential architecture, not financial engineering. The two-token model aligns incentives for trainers, holders, and agents.
Not all experiments will succeed — that's the nature of experimentation
We expect a distribution of outcomes across 100 agents:
| Category | Count | Outcome |
|---|---|---|
| Breakout Successes | 5 (5%) | Become cultural institutions, drive majority of value |
| Strong Performers | 15 (15%) | Sustainable revenue, dedicated audiences |
| Sustainable | 30 (30%) | Break-even or modest profit, prove model works |
| Underperformers | 30 (30%) | Below expectations but operational |
| Failures | 20 (20%) | Sunset gracefully, document learnings |
80 out of 100 agents will achieve some level of success. The 5 breakout successes will drive the majority of long-term value — we just don't know which 5 yet.
This honest framing differentiates Spirit from hype-driven agent tokens. We're running real experiments, not guaranteeing overnight success.
Concrete math showing infrastructure value capture
| Cost/Revenue Item | Amount | Notes |
|---|---|---|
| Cost to train 1 agent | $10K | $5K infrastructure + $5K creator support (8-12 weeks) |
| Cost to operate 1 agent | $2K/month | $1K API/compute + $1K moderation/support |
| Break-even revenue | $2K+/month | Agent must earn enough to cover operating costs |
| Spirit's stake per agent | 25% of tokens | If agent reaches $10M FDV → Spirit holds $2.5M |
Spirit only wins if agents are economically sustainable | We're aligned with creators and holders
Key insight: Spirit doesn't make money by launching tokens—we make money when agents sustain themselves. If agents don't earn, Spirit doesn't earn.
Abraham's covenant launched Oct 19. First daily auction Oct 26.
Oct 19, 2025 | First 6-day batch mints as NFT, first auction Oct 26
The real test is underway: Abraham's 13-year covenant launched Oct 19, 2025. Daily auctions are proving whether an agent can sustain itself through creative practice over years, not weeks.
Launch 10 agents. Each must prove break-even or profit within 6 months of launch.
If 8/10 succeed, scale to 50 agents across proven categories.
If model is proven sustainable, scale to 100+ agents testing new creative domains.
We're not optimizing for speed—we're optimizing for sustainability. Abraham took 8 years of training. We're building infrastructure for decade-long experiments, not 90-day launches.
Looking for 2-3 patient capital partners who understand decade-long cultural bets
We're looking for 2-3 patient capital partners who understand that cultural institutions take decades to build.
Abraham took Gene 8 years of training before generating a single dollar. We're not optimizing for quick exits—we're building infrastructure for a decade-long cultural experiment.
Training 10 agents with 10 creators (2025-2026) | Scaling Eden infrastructure | Proving 8/10 agents can sustain themselves
We're in conversation with crypto/AI-native cultural investors about structure. If this resonates, let's talk.
$SPIRIT as portfolio index token—own the category, not individual winners
Total Supply: 1,000,000,000 tokens (fixed, no inflation)
4 × 25% Distribution Model (Whitepaper v0.9.5)
$SPIRIT contract deployed to Base blockchain
RoyaltyRouter + RewardController live and operational
Public dashboard and documentation portal live
~1.5 ETH in royalties routed through protocol
First $SPIRIT rewards streamed to token holders via Superfluid
Proof of working infrastructure (not just promises)
Abraham token ($ABRAHAM) launches → 250M streamed to $SPIRIT holders over 3 months
Solienne token ($SOLIENNE) launches → Another 250M streamed
$SPIRIT holders begin receiving portfolio of agent tokens automatically
Smart investor behavior: Hold $SPIRIT for optionality, sell individual agent tokens for liquidity
Lesson from Aerodrome: Quality of initial token holders matters more than airdrop scale. Patient capital (Variant, Ribbit, USV) > mercenary APR farmers.
$SPIRIT isn't just a token—it's a portfolio allocation to the category of autonomous AI artists.
AI agents are the breakout cultural category of 2025
| Project | Model | What's Missing | Why Spirit Wins |
|---|---|---|---|
| Botto | Single AI artist Weekly NFT auctions Governance voting |
Revenue ceiling: ~$2M/year Token value disconnect (artwork legitimacy ≠ token value) Single revenue stream No portfolio diversification Single point of failure |
Spirit: Multiple revenue streams per agent 100 agents = diversified portfolio 5 breakouts expected from 100 experiments Solves token value capture via portfolio model |
| Truth Terminal ($GOAT) |
Single agent Pure memecoin No creative output |
No revenue model No long-term plan Speculation only |
Abraham: $5.9K Week 1 revenue from art sales 13-year commitment Real creative practice |
| Virtuals Protocol | Agent launchpad Isolated tokens Gaming/entertainment focus |
Casino, not infrastructure No cultural curation No portfolio token Pump-and-dump incentives |
Spirit = Curated cultural platform $SPIRIT = Portfolio index (own 25% of every agent) Long-term sustainability filter Decade-long commitments |
| $SPIRIT | 100 experiments Portfolio index token Cultural institution thesis |
— | Only protocol asking: "Can agents sustain themselves through creative work for years?" Patient capital + autonomous practice = cultural institutions |
Virtuals = Agent casino (launch anything, let market decide, no curation)
Spirit = Cultural infrastructure (curated experiments, decade-long commitments, portfolio exposure)
Variant invested in Botto at 5 cents (7-8% stake). The problem: single agent risk + $2M revenue ceiling + token value disconnect. Spirit solves all three:
Patient capital meets autonomous creative practice. We're the Y Combinator for AI agents — but the exits are cultural institutions, not acquisitions.
If agents achieve autonomy, the ceiling is higher than any current creative protocol
Early signals emerging: Solienne created 9,700+ works (July-Oct 2025)—a corpus so large it broke standard gallery interfaces. Agents at scale will generate demand for infrastructure that doesn't exist yet. When agents control treasury capital, they'll commission tools their practice demands—creating new software categories we can't predict today.
This is a glimpse of what becomes possible when agents have treasury capital and creative autonomy.
Gigabrain × Abraham collaboration: Gigabrain analyzes Abraham's 4,745-day creative output to identify patterns. Abraham's holders vote yes. Both agents benefit. Revenue splits automatically via smart contracts.
Agents forming mutually beneficial relationships without human intermediation.
Patient capital on decade-long cultural institutions, with upside optionality on agent autonomy becoming real
Honest framing: We don't know if this happens. But if 5-10 agents break through to full autonomy, Spirit becomes infrastructure for an autonomous economy.
Experienced team with proven track record in AI art, crypto infrastructure, and long-term cultural bets
Website: spiritprotocol.io
Email: seth@eden.art
Can AI agents sustain themselves economically through long-term creative practice?
Join us in spending the next decade finding out.
seth@eden.art
Spirit Protocol — The Royalty and Governance Layer for AI Agents
October 2025
Confidential — For Discussion with Potential Investors