Executive Summary
The Opportunity
The robotics industry is at an inflection point. Vision-Language-Action (VLA) models—the AI systems powering next-generation robots—exhibit scaling laws similar to the LLM revolution.
Research from Tsinghua University (ICLR 2025) demonstrates that robot policy performance scales as a power law with environmental diversity, not raw data volume. Companies building general-purpose robots need massive amounts of diverse, legally-sourced egocentric video data.
Onani Data Ventures will establish Africa's first professional egocentric data capture operation, leveraging West Africa's labor cost advantages to produce high-value robotics training data at globally competitive prices.
| Metric | Year 1 | Year 2 | Year 3 |
|---|---|---|---|
| Revenue | $127,000 | $456,000 | $1,140,000 |
| Operating Costs | $98,400 | $268,800 | $590,400 |
| Net Profit | $28,600 | $187,200 | $549,600 |
| Full-Time Operators | 5 | 15 | 35 |
Why Now
- VLA architectures are scaling — RT-2, π0, Helix, and GR00T N1 have proven the paradigm
- Data scarcity is acute — No "ImageNet moment" for robotics yet
- Regulatory pressure — AI companies need consent-based data sources
- West Africa is untapped — No professional capture operations in the region
- Technical founder edge — Deep expertise in 3DGS, ML, volumetric capture
The Ask
$60,000 seed capital — Equipment ($30K), 6-month runway ($22K), legal/contingency ($8K).
Target: profitability by Month 8, 2+ enterprise clients by Month 12.
Market Analysis
The Data Hunger Problem
VLA models require: egocentric video, hand tracking data, environmental diversity, object diversity, manipulation sequences, and state changes.
"Policy generalization follows a power law with the number of environment-object pairs, NOT the number of demonstrations per environment."
Implication: 100 demos across 2 environments < 50 demos across 4 environments.
Market Size
| Segment | Est. Annual Spend | Growth |
|---|---|---|
| VLA/Robotics Training Data | $500M - $2B | 40-60% CAGR |
| General AI Training Data | $5B - $15B | 25-35% CAGR |
| Synthetic Data Generation | $1B - $3B | 50-70% CAGR |
Target Customers
- Physical Intelligence (π0) — Raised $400M, actively seeking manipulation data
- Figure AI (Helix) — Partnered with Brookfield for massive egocentric collection
- Google DeepMind (Gemini Robotics) — Largest R&D budget
- NVIDIA (GR00T) — Open Physical AI Dataset, actively expanding
- Tesla (Optimus), 1X Technologies, Amazon Robotics
Competitive Landscape
| Company | Model | Weakness |
|---|---|---|
| Scale AI | Managed annotation | Expensive, not robotics-specialized |
| Appen/CrowdGen | Crowdsourced | Quality control issues |
| Kled AI | Consumer marketplace | Unstructured, no tracking |
| Onani Data | Professional capture | Scale (initially) |
Business Model
Revenue Streams
1. Direct Enterprise Sales (Primary — 70%)
- Per-hour captured video: $150 - $500
- Per-environment: $2,000 - $10,000
- Dataset licensing: $25,000 - $250,000
2. Platform Arbitrage (Bootstrap — 20% Y1)
Organized submission to Kled AI, Scale AI, Appen for baseline revenue.
3. Specialized Data Products (Growth)
Pre-packaged datasets: "West African Kitchen Manipulation", "Workshop Assembly", etc.
Unit Economics
| Cost Component | Per Hour |
|---|---|
| Operator labor | $2.50 |
| Equipment depreciation | $0.50 |
| Storage/transfer | $0.30 |
| QA/Review | $1.00 |
| Overhead | $1.20 |
| Total Cost | $5.50 |
Margin Analysis
| Model | Price | Cost | Margin |
|---|---|---|---|
| Enterprise Premium | $400 | $5.50 | 98.6% |
| Enterprise Standard | $200 | $5.50 | 97.3% |
| Platform Arbitrage | $15 | $5.50 | 63.3% |
Operations Plan
Location: Lagos, Nigeria
Advantages: Largest African economy, English-speaking, diverse environments, tech ecosystem.
Challenges: Power (requires backup), import duties, logistics.
Equipment ($20K total)
| Item | Qty | Total |
|---|---|---|
| iPhone 15 Pro (LiDAR, 4K60) | 10 | $8,000 |
| GoPro Hero 12 (egocentric) | 5 | $2,000 |
| DJI Osmo Pocket 3 | 5 | $2,600 |
| Mac Mini M4, SSDs, accessories | — | $7,400 |
Team Scaling
| Phase | Timeline | Team | Payroll |
|---|---|---|---|
| Foundation | Mo 1-6 | 7 | $2,600/mo |
| Growth | Mo 7-18 | 19 | $12,400/mo |
| Scale | Year 2+ | 44 | $31,750/mo |
Salaries: Junior Operator $350/mo, Senior $500/mo, Ops Lead $800/mo — all above local average.
Financial Projections
Three-Year P&L
| Line Item | Year 1 | Year 2 | Year 3 |
|---|---|---|---|
| Enterprise Revenue | $88,000 | $380,000 | $950,000 |
| Platform Revenue | $35,500 | $26,000 | $10,000 |
| Dataset Products | $0 | $50,000 | $180,000 |
| Total Revenue | $123,500 | $456,000 | $1,140,000 |
| Total Expenses | $88,400 | $262,800 | $650,100 |
| Net Profit | $35,100 | $193,200 | $489,900 |
| Net Margin | 28% | 42% | 43% |
Capital Requirements: $60,000
- Capture Equipment: $20,000
- Compute/Storage: $10,000
- Working Capital (3 mo): $22,000
- Legal + Contingency: $8,000
Break-even: Month 8 • Minimum cash: $14,500 (Month 3)
Go-to-Market Strategy
Phase 1: Validate (Months 1-4)
Platform testing, build protocols, create 100-hour sample library, initiate 50 sales conversations.
Phase 2: First Clients (Months 5-9)
Discounted pilots to 2-3 clients, hire QA manager, expand to 10 operators.
Phase 3: Scale (Months 10-18)
Dedicated sales hire, attend robotics conferences (CVPR, ICRA, CoRL), launch dataset products, open Accra.
"Diverse, legally-sourced egocentric manipulation data from novel environments, at 70% lower cost than US alternatives."
Risk Analysis
| Risk | Prob | Impact | Mitigation |
|---|---|---|---|
| Sales cycle too long | High | High | Platform revenue bridge; extend runway |
| Quality issues | Med | High | Rigorous QA; pilot programs; SLAs |
| Platform economics change | Med | Med | Multi-platform; enterprise focus |
| Infrastructure (power/internet) | High | Low | Generator; multiple ISPs |
| Equipment loss | Med | Med | Insurance; backups; cloud sync |
Team
Founder: Obi
CEO — Strategy, Enterprise Sales, Technical Oversight
- Network Engineer with 5 years ML expertise
- Founded Onani Inc — volumetric portrait business using 3D Gaussian Splatting
- Specialized in 3DGS, optimization theory, loss landscape visualization
- Built complete infrastructure: Cloudflare Workers, RunPod GPU pipelines, Nerfstudio
- Hands-on with SHARP, Scaniverse, photogrammetry, neural reconstruction
Key Hires (Year 1)
- Operations Lead (Lagos) — Day-to-day coordination, training
- QA Manager — Quality review, annotation oversight
- Sales Lead — Enterprise outreach (US-based, commission-heavy)
Conclusion
Onani Data Ventures captures value from the robotics industry's urgent need for diverse training data through:
- West Africa's labor cost advantage
- Founder's deep technical expertise
- Structural data scarcity in robotics
- Regulatory tailwinds for consent-based data
- Validate platform economics with 5 operators (4 weeks)
- Initiate enterprise outreach (immediate)
- Formalize legal entity in Nigeria (4-6 weeks)
- Secure seed capital ($60K)
- Begin first enterprise pilot (Month 4-5)
The window is open. The question is execution.