The Three Business Models
Before writing a line of code or buying a motor, decide which business model you are building. This determines your capital requirements, team composition, and timeline to revenue.
| Model | Examples | Capital to First Revenue | Time to Revenue | Gross Margin |
|---|---|---|---|---|
| Hardware (sell robots) | Unitree, Agility, SVRC (OpenArm) | $2M-$50M | 18-36 months | 30-50% |
| Software / AI (sell brains) | Covariant, Physical Intelligence, Realtime Robotics | $1M-$10M | 12-24 months | 70-90% |
| Data / Services (sell outcomes) | SVRC (data services), Formic (RaaS) | $200K-$2M | 3-12 months | 50-70% |
Technical Co-Founder Requirements
Robotics companies fail when the founding team lacks either hardware or software depth. At minimum, the technical founding team needs:
- Mechanical/electrical engineer: Can design a custom PCB, select actuators, build a prototype with 3D printing and CNC. Minimum: MS in ME/EE or equivalent industry experience.
- Robotics software engineer: Proficient in ROS2, motion planning (MoveIt2), control systems, and Python/C++. Can deploy and debug on real hardware. Experience with imitation learning or RL is increasingly essential.
- Domain expert: Deep understanding of the target customer's workflow. In logistics, someone who has worked on a warehouse floor. In manufacturing, someone who has programmed an industrial robot on a production line.
A solo founder building a robotics hardware company is a red flag for investors. The breadth of required expertise (mechanical, electrical, software, ML, domain) is too wide for one person.
MVP Definition for Hardware Startups
In robotics, "minimum viable product" does not mean a polished product. It means the simplest system that demonstrates your core value proposition to a paying customer. Stages:
- Demo (Month 0-3): Teleoperated or scripted demonstration on a lab bench using off-the-shelf components. Proves the task is technically feasible. Cost: $5K-$20K in hardware.
- Pilot (Month 3-9): Semi-autonomous system running at a customer site for 1-4 weeks. Mix of off-the-shelf and custom components. Proves the system works in a real environment. Cost: $20K-$100K.
- Alpha product (Month 9-18): Custom hardware and software, sold to 3-5 early customers. Reliability target: 90%+ uptime during customer use. Cost to develop: $200K-$1M.
- Production product (Month 18-36): Design for manufacturing (DFM), injection-molded enclosures, certified electronics, warranty and support infrastructure. Cost: $1M-$10M to reach production.
Manufacturing Path
Prototype (1-10 units): 3D printing (FDM/SLA), hand-assembled PCBs, off-the-shelf actuators. Lead time: 1-4 weeks. Cost per unit: 3-10x final production cost.
Pilot production (10-100 units): CNC machined parts, PCBA from a contract manufacturer (PCBWay, JLCPCB for small volumes, Flex or Jabil for larger). Start DFM review. Lead time: 6-12 weeks. Cost per unit: 1.5-3x final cost.
Volume production (100+ units): Injection molding for enclosures (tooling: $10K-$50K per mold), automated PCBA, quality control process. Lead time: 12-20 weeks for first production run. Target BOM cost: 25-40% of selling price.
Funding Milestones
- Pre-seed ($250K-$1M): Build the demo and validate with 2-3 potential customers. Friends and family, angel investors, SBIR grants. Use this stage to prove the problem is real and your team can build.
- Seed ($1M-$5M): Build the pilot system and deploy at 1-3 customer sites. Hardware-focused VCs (Lux Capital, Eclipse Ventures, DCVC) and climate/deep-tech funds. You need a Letter of Intent from at least one enterprise customer.
- Series A ($5M-$20M): Scale from pilot to alpha product with 10+ paying customers. Prove unit economics: customer acquisition cost (CAC), lifetime value (LTV), and gross margin. Traditional robotics VCs and strategic investors (Toyota Ventures, Samsung NEXT).
Key Failure Modes
- Building hardware nobody asked for: The most common failure. Always start with customer discovery, not with a cool robot design.
- Underestimating integration complexity: The robot works in your lab. At the customer site, the lighting is different, the table is at a different height, the objects have different textures. Budget 3x the time you think integration will take.
- Running out of cash before production: Hardware startups have longer cash conversion cycles. You spend 18 months and $2M before the first unit ships. If you raise too little, you die in the "valley of death" between prototype and production.
- Ignoring regulatory requirements: FCC testing (required for any electronic device sold in the US) takes 4-8 weeks and costs $5K-$15K. CE marking for EU. UL certification for safety-critical components. Start early.
- No recurring revenue: Selling hardware is a one-time transaction. Build recurring revenue through software subscriptions, data services, maintenance contracts, or consumable supplies.
Regulatory Path
- FCC (US): Required for all electronic devices. Part 15 for unintentional radiators (most robots). Budget $5K-$15K and 4-8 weeks. Use a test lab like Intertek or UL.
- CE (EU): Self-declaration for most robots. Machinery Directive 2006/42/EC applies. EMC Directive 2014/30/EU for electromagnetic compatibility. Budget $10K-$25K for testing.
- UL (safety): UL 1740 for robots in industrial environments. Not legally required in the US but increasingly expected by enterprise customers and insurance companies.
- ISO 10218/ISO/TS 15066: Required if your robot operates near humans. See our Collaborative Robot Safety Standards guide.