# Menlo

## 1-line

Menlo is building the brains for humanoids.

### Why Humanoids?

We believe that humanoids will be the dominant, scale-manufactured form factor for robots:

- Humanoids can use the same tools, interfaces, and environments as humans
- Human form factor datasets are growing exponentially, through CCTV and bodyworn cameras
- Humans attribute greater value (and willingness to pay) to humanoid robots, due to [Anthropomorphic Bias](https://en.wikipedia.org/wiki/Anthropomorphism)

We define humanoids as a spectrum of human-form factor designs (head, 2 arms, torso, "legs") designed to operate in human-centric environments, that perform tasks that are typically done by humans.

Menlo is agnostic to whether humanoids are wheeled or legged, and whether they have human-like features or a more utilitarian design.

## Why does Menlo exist?

Robotic adoption has been limited due to the following factors:

- Today's robotic models are task-specific and trap capex in narrow vertical TAM with limited cross-domain compounding
- Robotics is capital-intensive and hardware-bound, which limits iteration speed and increases cost of failure
- Robotics relies on bespoke physical data collection, which is expensive, slow, and often inadequate in terms of volume and diversity
- Robots today are sold as closed applicances, which throttles experimentation when modern robotics requires hardware-software co-design (e.g. adding sensors, trying different end-effectors)

## What Menlo does:

### Menlo Robot

> https://robot.menlo.ai

- Menlo Robot is an Embodied AI Agent abstraction, that adapts and builds upon the trillions invested in the LLM, VLM, CV stacks, and modern robotics models (e.g. Pi0, Gr00t) to enable humanoid robotics
- We leverage a synthetic data strategy to utilize existing commodity datasets (e.g. CCTV videos, bodyworn cameras) with a proprietary pipeline for transformation and domain randomization
- A Menlo Robot can either be deployed virtually in Menlo's World Simulator, or embodied in a Humanoid robot (e.g. Menlo's Asimov) with a combination of cloud and edge models

### Menlo Neocortex

> https://neocortex.menlo.ai

- Menlo's Neocortex is proprietary robotic operating system that is AI-Agent and humanoid centric, vs. the traditional ROS2 stack
- Menlo Pi is a proprietary x86-based robot processing unit (RPU) that provides GPS and 5G connectivity, and is optimized to run Neocortex (and optionally Brain models at the edge)

### Menlo World Simulator

> https://uranus.menlo.ai

- Menlo's World Simulator lowers the cost of failure, enable rapid experimentation and infinite data generation, and expand access to people who don't have access to physical robots
- Menlo's World Simulator works with businesses and property owners to build photorealistic digital twins of real-life environments, enabling virtual proof-of-concepts of robot deployments by researchers and vendors

### Menlo Asimov

> https://asimov.menlo.ai

- Menlo's Asimov is an open-source humanoid that enables researchers hardware-software co-design
- Menlo's Asimov is a reference design for Menlo Brain and Neocortex, and can either be 3D printed or purchased from certified manufacturers (similar to OpenArm[^1]).
- Menlo's Asimov is designed to provide best-of-class Real2Sim and Sim2Real transfer with Menlo's World Simulator

### Menlo Cloud

> https://cloud.menlo.ai

- Menlo Cloud is a Robotics cloud that provides IaaS services for robotics researchers and deployments
- Menlo Cloud supports common data engines to collect robot telemetry and data for both real-world and simulated environments (e.g. Rerun.io, Prometheus/Grafana)
- Menlo Cloud provides cloud-based LLM, VLM and CV inference for Menlo Robot deployments
- Menlo Cloud powers Menlo's World Simulator, providing scalable compute for photorealistic rendering and physics engines

[^1]: https://docs.openarm.dev/purchase/