2. What Problem Is Menlo Solving?

Robotics doesn't have a hardware problem. Most well funded teams can build a pretty good prototype. Robotics has a business model problem: robot intelligence needs internet scale data, and no single company can generate that alone.

The Three Broken Models

1. Vertical integration. Solve the single hardest use case with one perfect robot and hope to sell a few thousand units. This bets a huge amount of capital on one difficult SKU before the market, or the intelligence, is ready.
2. Buy the data. Assume physical intelligence will generalize once you acquire enough proprietary data from a handful of deals. This is a premature, capital intensive bet that one company's model will eventually fix everything, top down.
3. Robots as a service. Promise a customer near perfect reliability, then charge a subscription once you've hit it. This moves at the speed of one company chasing one reliability bar for one customer at a time.

None of these solve the actual bottleneck. They're all trying to manufacture, top down, what only an open, bottoms up ecosystem can generate at scale.

Menlo's Read

We believe the robotics data problem is better solved bottoms up, with more capital efficiency, by unlocking open innovation across 100,000 developers chasing the next billion dollar idea, the same way an open developer ecosystem unlocked mobile software instead of one company writing every app.

In practice, that means collapsing the bottlenecks that keep robotics locked to a narrow specialist community:

1. No Common Language for Skills

The AI agent boom has produced millions of developers who can reason, plan, and build with high level abstractions. Robotics is still stuck one level down, in motor control and implementation detail, with no shared way to describe what a robot should do rather than how its motors move. That gap keeps the broader software ecosystem on the sidelines.

2. No Open Hardware Standard

Humanoid robotics still lacks an open, modular hardware standard. Each vendor ships a vertically integrated stack with proprietary hardware and software, so there's no writing once and running on any humanoid, and no commoditization where factories compete on price and volume. The result is vendor lock in and an ecosystem stuck in high cost, low volume infancy.

3. Slow Iteration

Robotics lacks mature simulation and data collection infrastructure, so iteration is expensive and slow. Real world failures cost time, money, and hardware, and months of work can pass before anyone discovers something doesn't work.

4. Closed Supply Chains

Today's humanoids are largely vertically integrated with closed, permissioned supply chains. No standardization means no competition among actuator and parts suppliers, no local repair networks, and prices that stay high because there's no ecosystem pushing them down.