Skip to content
← All projectsDeal memo ·Reviewed 43d ago
Physical AIWatchingResearch CoverageSeedVerified

FrodoBots

DePIN cho mini-robots — telerobotic + AI — physical AI play

Research Coverage

Lookout covers this project based on publicly available information. Lookout does not represent, endorse, or have a commercial relationship with this project. Tier assignments reflect independent editorial judgment.

Executive summary

Watching On the radar — strong on some axes, needs more signal.

The Lookout view: the most substance-backed of the embodied-AI cohort on the data-supply side — real hardware in the wild on six continents and an actually-cited 2,000-hour open dataset, not a deck. Protocol Labs co-developing BitRobot is a strong validator. Two caveats temper the enthusiasm: sidewalk-rover data is a narrow slice of what robotics foundation models need, and value accrual now hinges on an unlaunched token plus a DePIN narrative that has to prove subnet demand. (We corrected our record here — the prior $2.5M / 2023 / 1kx entry was materially wrong.)

Key metrics

Stage
Seed
Raised
$8.0M
Founded
2022
Team
10
Geography
Singapore
Chain
Solana
Token
(no token yet)

Market opportunity

Why this, why now.

Real-world robot training data is the binding constraint for embodied AI, and FrodoBots attacks it from the cheapest possible end — gamified sidewalk rovers driven by a global crowd, already yielding the open-sourced 2,000-hour FrodoBots-2K dataset. The BitRobot subnet architecture (with Protocol Labs) generalizes this into a DePIN for many robotics data tasks. Timing is good: data scarcity is the bottleneck everyone from the big labs down is now funding.

Team assessment

Founder track record.

Michael Cho

@micoolcho

Singapore-based serial entrepreneur who started FrodoBots as a weekend project to gamify robotics data collection.

Sam Cho

Co-founder; co-started FrodoBots' early gamified-robot experiments.

Competitive position

Where it sits.

FrodoBots' moat is distribution and traction — hundreds of deployed rovers across six continents and a genuinely cited dataset, more real-world embodied data than most crypto-robotics peers combined. Versus PrismaX (teleop of robot arms for manipulation data) it owns navigation/mobility data instead; versus OpenMind it's data-supply, not OS. The risk is that low-cost sidewalk-rover data is narrow — navigation, not the dexterous manipulation data the foundation-model labs most want.

vs PrismaX vs OpenMind vs Tesla / Waymo fleetsvs AgiBot

7-axis evaluation

The full read.

Signal mix · 7 axes

3 Strong4 Neutral0 Weak
01

Team & Execution

Strong

Shipped real hardware (mini-robots) with real users (gaming). Most tangible Embodied-AI product in the category — not a deck.

02

Tech & Differentiation

Strong

Telerobotic gaming network generates real teleoperation data. Hardware + crypto incentives is a defensible combination.

03

Tokenomics & Economics

Neutral

No token yet. Earn-via-play model promising but economics unproven until token launch.

04

Traction & Adoption

Strong

Real users playing with real robots. ET Fugi / global robot network is genuine traction, rare in Embodied-AI.

05

Funding & Backers

Neutral

1kx + Castle Island backing is quality but round size is modest. Enough runway to prove the scaling thesis.

06

Narrative & Market Fit

Neutral

Embodied-AI x crypto is early. FrodoBots is the most real product but the question is whether it scales beyond novelty.

07

Risk Vectors

Neutral

Novelty risk — fun today, durable business unclear. Hardware scaling is capital-intensive and slow.

Lookout risk view

What could break it.

  • Data niche may be too narrow — sidewalk-navigation footage is far from the dexterous-manipulation data frontier labs pay most for.
  • The BitRobot pivot leans heavily on Protocol Labs involvement and an unlaunched token; execution and token design are both unproven.
  • Hardware + global fleet ops is capital- and logistics-heavy for an ~$8M-funded team, capping scale versus better-capitalized rivals.

VC fit

VCs that fit this deal.

Data confidence: Verified

Facts sourced · take is Lookout judgment

No advisory relationship at time of writing. If that changes, this memo updates first.

Is this your project?

Spot inaccurate data or want to update your listing? Reach out and we'll prioritize your review.