# Tangible Robots

> Source: https://aiwiki.ai/wiki/tangible_robots
> Updated: 2026-06-07
> Categories: Humanoid Robots, Robotics Companies
> From AI Wiki (https://aiwiki.ai), a free encyclopedia of artificial intelligence. Quote with attribution.

**Tangible Robots** (also known as **Tangible Robotics**) is an American robotics startup based in [San Francisco](/wiki/san_francisco), [California](/wiki/california), [United States](/wiki/united_states), developing dexterous [humanoid robot](/wiki/humanoid_robot)s for household use. Founded by Bipasha Sen, Shankara N. Vaidyanathan, and Benjamin Soria, the company emerged from stealth mode in 2025 with its flagship product, the [Eggie](/wiki/eggie_robot) wheeled humanoid robot. Tangible Robots emphasizes dexterity, compliance, and whole-body control, enabling its robots to operate in the cluttered, unpredictable environments typical of real homes.[1][2]

The company sits inside a wider 2025 wave of consumer-facing humanoid startups that includes [Sunday Robotics](/wiki/sunday_robotics), [Cartwheel Robotics](/wiki/cartwheel_robotics), and [1X Technologies](/wiki/1x_technologies). Tangible's bet is sharper than most peers: while rivals pair simple grippers with heavy software, Tangible insists that humanlike, five-fingered hardware is the only practical way to operate inside a world built around human hands. The stated mission is to "put robots in every home," and Bipasha Sen has framed the engineering philosophy with the line, "What is AGI without a sense of touch?"[2][3]

## History

Tangible Robots was co-founded in 2024 by Bipasha Sen, who left an [MIT](/wiki/mit) PhD program to pursue building robots for homes, along with fellow researchers Shankara N. Vaidyanathan and Benjamin Soria.[1][3] Sen had been a first-year PhD student at MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL), where she was supported by the Ida Green Fellowship and advised by Professor Pulkit Agrawal. Her prior research, conducted at IIIT Hyderabad, Brown University, and the University of Toronto, focused on 3D shape completion (SCARP, ICRA 2023), neural radiance fields (HyP-NeRF, NeurIPS 2023), and scene understanding (ConceptGraphs, ICRA 2024).[4] Before graduate school she worked at Microsoft as a data scientist on the recommendation system inside [Outlook](/wiki/outlook).[5]

The early prototyping work was done while the team was incubated at Founders, Inc., a San Francisco hardware-focused accelerator led by General Partner Hubert Thieblot. Thieblot personally backed the company and Founders, Inc. provided initial seed support and physical workshop space during the stealth phase.[2][6]

The company completed a pre-seed investment round in late 2024, followed by a $3 million seed round in 2025 led by India-based Blume Ventures and Micelio Technology Fund. Total disclosed funding as of 2025 was approximately $4 million across the two early-stage rounds, although individual contributions from US-based angels were not disclosed.[1][6]

Tangible Robots emerged from stealth in November 2025 with a 26-second public reveal video of its [Eggie](/wiki/eggie_robot) humanoid robot wiping up a spilled coffee on a kitchen counter. The release positioned the company in the rapidly growing market for home humanoid robots and arrived only days after the public debut of competing startup [Sunday Robotics](/wiki/sunday_robotics) and its Memo robot.[1][2]

## Founders and leadership

The founding team brings together backgrounds in machine learning research, surgical robotics, and large-scale autonomous deployment. The company has stated that its leadership includes graduates of MIT, [UC Berkeley](/wiki/uc_berkeley), and [USC](/wiki/usc), with prior experience deploying thousands of autonomous robots and developing control systems for advanced surgical robotics.[3]

| Founder | Role | Background |
|---------|------|------------|
| Bipasha Sen | CEO and Co-Founder | Former MIT CSAIL PhD candidate advised by Pulkit Agrawal; previously a Data Scientist at Microsoft; researcher at IIIT Hyderabad, Brown University, and the University of Toronto |
| Shankara N. Vaidyanathan | Co-Founder | Leads core robotics and control systems work |
| Benjamin Soria | Co-Founder | Heads mechanical and software engineering |

Sen serves as the public face of the company and the main spokesperson on technical philosophy. She has spoken publicly about her decision to leave MIT, framing it less as a dropout and more as a swap of research timelines for product timelines. In a Homegrown India profile she described her vision: "I see robots as a means to augment our physical world like AR/VR once promised."[5]

## Products

### Eggie

[Eggie](/wiki/eggie_robot) is Tangible Robots' flagship mobile humanoid robot, unveiled in November 2025. The robot was designed for the home market rather than the warehouse, factory, or hospitality verticals favored by many of its larger competitors.[1][2]

| Specification | Details |
|--------------|--------|
| Design | Wheeled humanoid upper body |
| Height | Approximately 160 cm |
| Weight | Approximately 45 kg |
| Hands | Anthropomorphic five-fingered, two |
| Fingers per hand | 5 |
| Degrees of freedom | Roughly 20 overall, 3 in each hand |
| Head | Rectangular with two front-facing cameras |
| Cameras | Dual 1080p RGB plus depth |
| Materials | Smooth white composite panels over an aluminum frame, exposed black joints |
| Mobility | Wheeled base (no bipedal legs) |
| Max speed | About 3 km/h |
| Arm payload | Around 1 to 2 kg per arm |
| Runtime | Approximately 7 hours per charge |
| Connectivity | Wi-Fi, Bluetooth |
| Latency | Under 200 ms end-to-end |
| Target market | Consumer, home |
| Price (announced) | $32,000 |
| Status | Prototype, not yet commercially available |

Eggie features a pair of complex, anthropomorphic, five-fingered hands, which distinguish it from many competing home robots that use simpler grippers or fewer fingers. The five-fingered design enables Eggie to manipulate a wider range of household objects, including those that require pinch grips, wrap grips, and other complex hand configurations.[1][2]

#### Design philosophy

Eggie's design is clean and minimalist, with smooth white panels covering much of its structure and exposed black joints at the wrists and elbows. The robot has a rectangular head with two front-facing cameras that serve as its primary visual sensors and give it a face that reads as friendly rather than menacing. The aesthetic aim is to be approachable and non-threatening in a home environment.[2]

Like other wheeled humanoids targeting the home market, Tangible has opted for a wheeled base rather than bipedal legs, prioritizing stability and energy efficiency over the complexity of legs. This practical choice reflects the flat, indoor nature of most household environments and also gives the company much longer continuous operation windows for data collection compared to legged platforms that must be tethered or charged frequently.[7]

#### Capabilities

Eggie is built for household tasks in cluttered, unpredictable home environments. The robot's control system emphasizes three behaviors:

- **Dexterity**: Five-fingered hands with compliant control for handling diverse objects, from delicate glassware to bulky pots.
- **Compliance**: The ability to adapt to unexpected contacts and disturbances safely, rather than rigidly resisting them.
- **Whole-body control**: Coordinated movement of base, torso, arms, and hands so the robot acts as a unified system rather than a stack of independent limbs.

The stealth launch demo showed Eggie picking up a white mug with one hand while wiping a coffee spill on the counter with a red cloth held in the other hand, an example of what Tangible calls a "contact-rich" task. Reported autonomous behaviors include wiping spills, holding mugs, coordinated two-handed manipulation, and general household object handling. Tangible has been explicit that Eggie is intended for messy, unstructured environments such as cluttered sinks, fridges, and countertops, rather than clean factory floors or staged demo rooms.[1][2][8]

#### Frank

In addition to Eggie, Tangible has discussed an internal whole-body control system named **Frank**, used to coordinate household chores across the robot's base, torso, arms, and hands. Frank was co-led by multiple members of the founding team and represents around a year of development work on tightly integrated hardware and remote teleoperation interfaces.[5]

## Technology

Tangible Robots' core technology focuses on three areas: dexterous manipulation, compliant control, and whole-body coordination. The team is unusually opinionated about hardware choices, arguing that machine learning alone cannot rescue a robot from a poorly designed gripper. As a corollary, they invest heavily in real-world data collection rather than relying on simulators or synthetic data.[1][3]

### Dexterous manipulation

The company's primary differentiator is its approach to dexterous manipulation. Eggie's five-fingered anthropomorphic hands are designed to handle the enormous variety of objects found in homes, from delicate glassware to heavy pots. Most competing wheeled humanoids in the home market favor simpler parallel-jaw grippers, on the bet that strong AI can do more with less hardware. Tangible's view is the opposite: human-built environments expect human-shaped manipulators, and the cost of a simpler hand is paid every time the robot tries to grip a kettle handle, a tea bag, or a fridge drawer.[2]

### Compliant control

Tangible's control system emphasizes compliance, meaning the robot yields to unexpected forces rather than rigidly fighting them. This is critical for safety in home environments where collisions with furniture, pets, and people are inevitable. Compliance also helps the robot perform contact-rich manipulation: pressing a sponge into a counter, easing a heavy pan into a sink, or settling a dish onto a stack without shattering it.

### Whole-body coordination

The robot coordinates its wheeled base, torso, and dual arms as a unified system, enabling complex tasks that require simultaneous movement and manipulation. Reaching deep into a fridge, for example, calls for the base to nose forward, the torso to lean, and an arm to extend at the same time. Tangible's Frank control system is the in-house framework that handles this coordination.[5]

### Training data and real-world collection

Unlike many robotics firms that lean on simulation and synthetic data, Tangible focuses on real-world data collection. Engineers wear a sensor-packed partial exoskeleton that captures human-level touch and force information during tasks like cleaning and picking up objects. Visiting journalist Robert Scoble observed engineers wearing this exoskeleton equipment during a site visit, noting that it captures specifically how much pressure to apply to different materials, a signal that is hard to recover from cameras alone.[1][2] The company also stresses that its policies are trained on data collected from hundreds of unique environments and thousands of interactions rather than purely from simulation.[3]

### Software stack

Public information about the software stack is limited, but the company has described the platform as integrating imitation learning, reinforcement learning, dexterous manipulation, and vision-language-action models. Eggie reportedly runs Linux with a [ROS](/wiki/ros)-based environment and uses cloud LLM APIs for high-level reasoning when needed. End-to-end control latency is reported under 200 ms.[3][7]

## Funding

| Round | Amount | Lead Investor | Date |
|-------|--------|--------------|------|
| Pre-seed | Undisclosed (approximately $1 million) | Founders, Inc. and angels including Hubert Thieblot | Late 2024 |
| Seed | $3 million | Blume Ventures, Micelio Technology Fund | 2025 |
| **Total disclosed** | **About $4 million** | | |

The investor list is notable for two reasons. First, the lead seed investors, Blume Ventures and Micelio Technology Fund, are India-based, an unusual pairing for a San Francisco humanoid robotics company and a reflection of Sen's standing as an Indian-origin AI founder. Second, reported recruiting offers in the $1 to $2 million range are aggressive relative to disclosed funding, suggesting either substantial undisclosed capital or an expected follow-on round.[1][2][6]

## Hiring and team building

Tangible has been aggressive about hiring. Sen has used the company's stealth-exit moment as a recruiting event, pitching openings in imitation learning, reinforcement learning, dexterous manipulation, and vision-language-action models. Reported total compensation for top research candidates ranges between $1 million and $2 million annually, high even by current Bay Area robotics standards.[2][3]

## Market position and competitors

Tangible Robots is competing in the emerging market for consumer home humanoid robots, a category that has become unusually crowded in 2025. Key competitors and adjacent players include:

| Company | Robot | Form Factor | Notes |
|---------|-------|-------------|-------|
| [Sunday Robotics](/wiki/sunday_robotics) | Memo | Wheeled humanoid | Direct competitor; launched in stealth just days before Tangible. Uses simpler grippers and a full-stack AI bet |
| [Cartwheel Robotics](/wiki/cartwheel_robotics) | Cartwheel | Bipedal humanoid | Consumer-friendly companion humanoid |
| [1X Technologies](/wiki/1x_technologies) | Neo | Bipedal humanoid | Larger, better-funded; consumer home robot |
| [Figure AI](/wiki/figure_ai) | Figure 02 | Bipedal humanoid | Primarily industrial, with home ambitions |
| [Tesla](/wiki/tesla) | [Optimus](/wiki/tesla_optimus) | Bipedal humanoid | General-purpose, factory-first roadmap |
| [Unitree](/wiki/unitree) | G1-D | Wheeled semi-humanoid | Lower-cost wheeled competitor |

In its public materials, Tangible draws the strongest contrast with Sunday Robotics. Both companies are building wheeled, friendly-looking mobile manipulators for the home. The most significant hardware difference between Eggie and Sunday's Memo is the hand: Sunday has bet on powering a simpler gripper with stronger AI, while Tangible has bet that anthropomorphic five-fingered hardware is essential for a world built around human hands.[2]

## Reception

The November 2025 stealth-exit video drew strong reactions from the robotics commentariat. Roboticist Chris Paxton called the moment "a good day for wheeled semi-humanoids," placing Eggie and Sunday's Memo in the same emerging category and arguing that wheeled platforms are a pragmatic foundation for AI-focused research while legged technology catches up.[7] AOL and several smart-home outlets ran comparison features placing Eggie alongside [1X](/wiki/1x_technologies)'s Neo, NVIDIA's Isaac platform, and Sunday's Memo as the most visible 2025 entrants in the domestic robot category.[8] Trade publications focused on the dexterous hands as the defining choice, with mikekalil.com characterizing Tangible as the company "betting on dexterous hands" rather than software-only differentiation.[1]

Skeptics noted that the company is still pre-product and that the $32,000 listed price is more of a positioning signal than a commercial commitment, given that Eggie remains a prototype.[7]

## Vision and philosophy

Sen has spoken about an ambition that runs further than household chores. The pitch is that physical AI, robots that can perceive, plan, and manipulate inside human spaces, is the next platform shift after smartphones, and that contact-rich manipulation is the technical bottleneck. The repeated rhetorical question from the founding team is, "What is AGI without a sense of touch?"[1][5]

The company has also tied its mission to a broader argument about [embodied AI](/wiki/embodied_ai): that a learning system disconnected from physical interaction cannot fully understand the world, and that home robotics is the most demanding testbed because it is messy and personal.

## See also

- [Eggie (robot)](/wiki/eggie_robot)
- [Humanoid robot](/wiki/humanoid_robot)
- [Service robot](/wiki/service_robot)
- [Dexterous manipulation](/wiki/dexterous_manipulation)
- [Sunday Robotics](/wiki/sunday_robotics)
- [Cartwheel Robotics](/wiki/cartwheel_robotics)
- [1X Technologies](/wiki/1x_technologies)
- [Figure AI](/wiki/figure_ai)
- [Embodied AI](/wiki/embodied_ai)
- [Imitation learning](/wiki/imitation_learning)
- [Reinforcement learning](/wiki/reinforcement_learning)
- [Vision-language-action model](/wiki/vision-language-action_model)

## References

1. Humanoids Daily. "Meet Eggie: Tangible Robotics Joins the Wheeled-Humanoid Race, Betting on Dexterous Hands." 2025. https://www.humanoidsdaily.com/feed/meet-eggie-tangible-robotics-joins-the-wheeled-humanoid-race-betting-on-dexterous-hands
2. Mike Kalil. "Tangible Robots Unveils Eggie Humanoid for Homes." mikekalil.com, 2025. https://mikekalil.com/blog/tangible-robots-eggie/
3. Tangible Robots. "Tangible: General-purpose manipulation robots for real-world environments." Company website. https://tangiblerobots.ai/
4. Bipasha Sen. "Bipasha Sen, PhD student at MIT CSAIL." Personal academic website. https://bipashasen.github.io/
5. Homegrown India. "What Comes After Alexa? Bipasha Sen's Tangible Robots Wants To Build It." 2025. https://homegrown.co.in/homegrown-voices/tangible-robots-bipasha-sen-is-reimagining-a-future-of-commonplace-robotics
6. Founders, Inc. "Tangible: Humanoid robots for your home." Portfolio page. https://f.inc/portfolio/tangible/
7. Chris Paxton. "A Good Day for Wheeled Semi-Humanoids." It Can Think Substack, 2025. https://itcanthink.substack.com/p/a-good-day-for-wheeled-semi-humanoids
8. AOL News. "Eggie, Neo, Isaac and Memo are domestic robots. But would you let them load your dishwasher?" 2025. https://www.aol.com/articles/eggie-neo-isaac-memo-domestic-000730928.html
9. Humanoid Press. "Eggie, Wheeled Humanoid Home Robot by Tangible Robotics." Humanoid Press Database. https://humanoid.press/database/humanoid-press-database-eggie/
10. Humanoid Guide. "Eggie Humanoid Robot by Tangible Robots." Product page. https://humanoid.guide/product/eggie/