One vision. Every circuit connected.
This whole plan is one machine. Hover a node to see what it powers. Click to fly to that subsystem.
The assets are real.
The structure isn't finished.
We customize enterprise-grade robots to fit how each business actually operates — source, customize, deploy — as a fully managed service.
Nursing-round telepresence
The $11.8M LOI — 300 units into healthcare facilities, automating vitals rounds, delivery, and non-patient tasks.
Charging-station manager robots
Robots that manage electric-vehicle downtime — physically plugging in vehicles that need a human hand today. A pilot is underway right now.
Hotel operations
Back-of-house logistics, delivery, and guest-facing service in hotels — a staffing-shortage industry ready for the PureUpside wedge.
We are the deployment layer
Hardware manufacturers build platforms. Physical-AI companies build intelligence. Neither can put a working robot inside a real business. We partner with both sides and build everything in between.
Hardware OEMs
Platform supply — arms, chassis, lifts, hands
OPEN DROIDS
Task fit · reliability engineering · on-site integration · managed operations
Physical-AI Models
Intelligence supply — manipulation & embodied reasoning
In commercial service deployment, nobody else plays this role — InOrbit orchestrates warehouses, Formant sells fleet tooling, Formic and Intrinsic live in industrial manufacturing. The commercial deployment layer is empty. We take it — with VLA-native orchestration, an always-upside contract, and India-leveraged economics.
The API layer for all robots
One interface between every business problem and the right robot — regardless of vendor, regardless of use case. The experts in between, so neither side gets ripped off.
The right robot, not our robot
Buyers don't know robots; robot makers don't know the buyer's operation. We consult across every use case and match the platform that actually fits — ours or anyone's.
Nobody gets ripped off
Buyers don't overpay for the wrong platform; OEMs get qualified, deployment-ready customers. We monetize deployment and RaaS — never information asymmetry.
The LOI queue is the proof
Every concierge conversation is captured via Google sign-in, curated into a plan, and converted toward pre-orders and LOIs — a visible queue of real demand for whoever gets robots out of labs and into production.
> curated plans → follow-up → pre-order/LOI
> every entry = evidence of demand
Robots that build robots
Long-term, Open Droids is not just deploying robots. We're building toward self-replication: develop one robot that can assemble the next one — because robots building robots is what makes the $5k robot possible.
A factory factory
A production system where robots build more robots — capacity that compounds instead of scaling linearly.
Deployments fund the mission
Every deployment generates the revenue, the data, and the reliability engineering self-replication will demand.
Trust, earned deliberately
None of it happens unless clients — and people — trust these robots far beyond what the industry has earned. We build that trust deployment by deployment, by design.
The $5k robot is the prize. The $50k robot earns it.
Nobody starts at the cheap end of a hardware cost curve — you descend it. We start where reliability is worth $50k, win trust deployment by deployment, and ride the flywheel and robots-that-build-robots down to the home robot.
Healthcare, EV charging, hospitality — where reliability is worth the BOM, and where trust is earned.
The descent begins: every deployment makes the next robot cheaper to build and faster to trust.
The robot in every home — the goal the whole machine exists to reach.
The customer never sees the BOM
Under PureUpside + RaaS, a $50k robot on 60-month terms is under ~$1k a month — against an $80–120k/yr labor line. The hardware price is invisible. The 98% completion guarantee is not.
Cheap-first filled the graveyard
Serve ran a negative gross margin at 2,000 robots. The money dies in failures, truck rolls, and churn — not in BOM. Over the life of the contract, the expensive robot is the cheap one.
Fanatically cheap where invisible
Partner components, one rogent base image, India engineering, the duty split — R&D leverage saves by a mile. Never the part the customer touches.
The home robot isn't rejected — it's scheduled. Every deployment, every datapoint, every robot-built robot is one step down the curve.
Get the robots out. And collect.
Bigger is better — more orders force cleaner execution. First on the ground wins what no lab can produce: real-world task data and real-world failures.
We collect more than the models consume
Cameras, audio, everything — but only while the robot is working: executing tasks, running agent workflows, interacting with people. Never passive, never always-on. Partners only ever touch a slice.
Partners get their slice free
Physical-AI partners get free access to deployment data — that free flow is what cements the partnerships.
The surplus is ours to sell — consented & rights-cleared
Idle-time perception, human interaction, agent traces — captured event-triggered with on-device de-identification, healthcare fenced under BAAs. What we sell is the consented, rights-cleared dataset; that's what makes it sellable at all.
Annotation is the edge
A large, extremely accurate annotation team in India turns raw logs into labeled, high-value datasets at very low cost.
Let everyone build on these robots
The biggest key: thousands of roboticists and software engineers creating on our platforms — without owning a robot. We are not building robots stuck in time.
Hosted simulation platform
A website where anyone can experiment, write algorithms, and vibe-code their way up. // UNDER DEVELOPMENT — ALMOST READY
Clone it. Run it. It works.
Movement, grasping, core task primitives — R2D3_ros2 is live and public today; the rogent-v3 orchestrator opens with the platform launch. Together: the community's starting point.
Robot Orchestration System
Our core IP: the orchestration layer that wraps physical-AI models and gives the robot agentic interaction with its surroundings — built on Gemini Robotics-ER, publicly available via API since July 30, 2026.
The API door is already open
Gemini Robotics-ER went public API on July 30, 2026 — we can build on it today, no permission needed. The partnership play: one robot demonstrably working end-to-end on it, then walk it into DeepMind.
Real money for real progress
Community builders earn bounties for every "solve" or meaningful contribution to the models and code — the stack actively improves because improving it pays. Standout contributors get onboarded full-time from the community — the bounty board doubles as our hiring pipeline.
The loop closes in deployment: real → sim → real plus reinforcement learning in the field. Collect, fine-tune, redeploy — every robot keeps improving.
First contact → deployed robot
One pipeline. Six stages. Two continents. Hover each station.
Outreach & Contract
SALESLand the PureUpside agreement — the always-upside contract. The $0-cost automation audit is the wedge in the door.
Audit & Discovery
ON-SITECapture problems, map what to automate, review layouts, workflows, throughput, payroll baselines.
Sourcing Decision
PER USE CASEAdapt an existing design, create a whole new robot, or procure a platform from China.
R&D, Build & Test
INDIAHardware design + software harness by top engineers and standout students. Partner models supply intelligence; we build the rest for whole-task reliability.
Integration & Assembly
USAOnly actuators & mechanical parts ship over. Every sensor and Jetson is bought in the US — NDAA & FCC compliance is the point, the value prop — assembled at our SF site, locked with CBP binding rulings before scale. A US-based robot: simple, safe to use.
Deploy & Operate
CUSTOMER SITE30-day proof-of-savings pilot (≥33% cost cut, ≥98% completion), convert waves, operate under RaaS.
What's on. What we switch on next.
Not roles — objectives. An honest board of what the current team covers against what the pipeline demands.
On-site discovery, workflow mapping, solution architecture.
Full-stack robotics — ROS, autonomy, manipulation — through whole-task reliability.
OEM + physical-AI partnerships — carried alongside build work; short-term sustainable only.
Founder-level investor access and the PureUpside contract engine.
The CEO seat is empty; the Founder sits above the structure. Finding the operator is a near-term objective.
No one owns outreach and closing PureUpside agreements at volume.
The Founder owns fundraising; missing is an experienced operator beside him.
The India facility and US assembly site don't exist yet. Both get stood up.
Post-conversion, someone must own uptime and the 98% guarantee daily.
An experienced CEO via the Founder's network + deliberate search. Everything accelerates once this exists.
3 sales & business-development hires at the foundation — driving contracts and pre-orders from day one, scaled aggressively from there.
Solved by the first gap: an operator beside the Founder in investor conversations.
The India design/build/test facility + the US assembly site — funded and stood up, not hired.
Grows out of deployment experience; formalized when the first wave converts to RaaS.
Derived. Not asserted.
A bottoms-up estimate built from component research — India facility, US site, tariff stack, executive comp, pilot economics, sim infrastructure.
| Component | Low | Expected | High |
|---|---|---|---|
| PHASE 1 — R&D FOUNDATION (MONTHS 0–9/12) · ROBOTS AREN'T SOFTWARE — DEV TAKES REAL TIME | |||
| CEO cash (9–12 mo) | $120k | $150k | $180k |
| Sales & business development — 3 (9–12 mo)> at the foundation: driving contracts & pre-orders from day one | $180k | $270k | $380k |
| Core SF engineering (2 senior + intern cohort, 9–12 mo)> two senior builders; intern cohort included | $220k | $250k | $300k |
| India starter team + workspace (scaled down)> in Phase 1 from day one — robot R&D can't wait for Phase 2 | $70k | $115k | $180k |
| Robot R&D prototypes (2–3, BOM + tariffs) | $70k | $110k | $180k |
| Shipping & misc (freight, crating, spares, tools)> everything robot development needs — deployments don't happen out of the blue | $25k | $45k | $80k |
| 3–4 provider-funded pilots | $60k | $100k | $150k |
| Sim platform (dev share + infra)> carried by cloud-provider credits & partnerships | $8k | $18k | $35k |
| Legal / entity setup> $0 — entity already established | — | — | — |
| Phase 1 subtotal | $0.75M | $1.06M | $1.49M |
| PHASE 2 — THE MACHINE (MONTHS ~12–24) | |||
| India facility scale-up (one-time: fit-out, lab) | $40k | $70k | $160k |
| India full 12-mo run (rent + 6 eng + interns + ops) | $135k | $205k | $330k |
| Annotation & data collection> partner-run, not in-house hires — separate partnership cost | $25k | $45k | $75k |
| SF assembly site: setup + 12-mo lease + 2 techs | $230k | $290k | $400k |
| ~10 deployment robots (BOM + tariffs + freight) | $300k | $420k | $600k |
| US team growth (+3–5 eng, +sales, months 12–24) | $800k | $1.05M | $1.4M |
| Core team continuation (months 12–24)> the Phase 1 team keeps running while the machine scales | $520k | $670k | $860k |
| Field service & deployment engineering> lean forward-deployed engineers — 1 FDE + 1 field tech covers the first 10–20 robots | $200k | $300k | $430k |
| Site integration — 4–6 revenue sites @ $15–30k> mapping, elevators, doors, workflow config (Diligent/Aethon norm) | $60k | $120k | $180k |
| Fleet sustainment: spares inventory + maintenance> 10–15%/yr of fleet hardware value | $60k | $110k | $170k |
| Insurance & healthcare compliance> GL/PL, per-robot policies, vendor credentialing — get quotes | $50k | $90k | $160k |
| 24/7 remote monitoring / teleassist (India, 3–5 FTE) | $40k | $80k | $140k |
| Sim scale + internal Isaac + cloud> under GPU credits from cloud/NVIDIA partnerships | $10k | $25k | $50k |
| Ops & G&A (residual) | $70k | $100k | $150k |
| Phase 2 subtotal | $2.54M | $3.58M | $5.11M |
| MARGIN + FLEET CAPITAL | |||
| Total spend (24 mo) | $3.3M | $4.6M | $6.6M |
| Contingency margin (+25% expected) | — | +$1.16M | +$0.9M |
| Equity raise target | $6M – $7.5M | ||
| Fleet hardware at LOI conversion (150–300 units ≈ $5.6–18M)> NOT equity — pre-arranged $4–8M equipment-debt facility, drawn only on signed conversion orders; Series A (~$25M median) triggers at months 18–24 for the 50→300 ramp | $4–8M debt facility + Series A trigger | ||
SF payroll
~40–50% of total spend, repriced at a Bay Area premium (~10–15% over national). Seed-stage CEO cash averages $147k; the 3 foundation sales/BD hires run $100–150k OTE each. Engineering stays lean: 2 seniors + interns.
Hardware under two shocks
NVIDIA raised Jetson prices 33–101% in July 2026; China parts carry a ~25–40% tariff stack. Our split: only actuators & mechanicals cross the border — NDAA-compliant sensors and GPUs bought in the US, with CBP rulings locking the origin math.
The India entity
Cheap per unit — mid engineers ₹12–25 LPA, annotation <$1/hr, 4,000 sq ft lab ≈ $39k/yr — but a real second entity: ~$220–350k/yr + setup.
Equity funds the team. Debt funds the robots.
The comparables are brutal: Diligent Robotics burned ~$100M to reach ~100 hospital robots and sold for $25.7M; Serve ran a negative gross margin at 2,000 robots; Rapid Robotics abandoned RaaS entirely. The winners split the capital: Chef Robotics paired $20.6M equity with a $22.5M SVB equipment facility; Formic stacked $100M+ of dedicated debt. We do the same — the $6–7.5M equity never buys fleet hardware; a pre-arranged $4–8M equipment-debt facility (2025 terms: 8–15%, dropping to ~6–9% asset-backed as contracts season) draws only on signed PureUpside conversion orders, and a Series A at months 18–24 — triggered by 10 robots live and proven unit economics — funds the 50→300-unit ramp.
// verify before investor use: post-July-2026 tariff rate · senior India salary band · sim GPU density · $25–40k robot BOM
Three phases
Each phase begins when the previous one's milestones are real — power flows forward only when earned.
Foundation
- New operating structure adopted — clear ownership of every objective
- CEO search launched
- 3 sales & BD hires onboard — contracts and pre-orders driven from day one, scaled aggressively from there
- India starter cohort onboarded — robot R&D underway from day one; hardware takes real time, deployments don't happen out of the blue
- Simulation platform launches publicly
- AI concierge front door live — Google-login intake feeding the pre-order/LOI queue
- DeepMind milestone: one robot end-to-end on the embodied-reasoning model
- First PureUpside audits signed; engagements advanced to deployment scopes
The Machine
- India design/build/test facility live with first engineering cohort
- US assembly & integration site operational
- First contract-driven robots designed, built, shipped, assembled, deployed
- Annotation team online — data flywheel turning
- Community building on the combined starting-point repos
Compounding
- Wave conversions stacking long-term RaaS revenue
- Data revenue flowing as a second stream
- Fleet operations mature — 98% completion held at scale
- Real→sim→real RL loop running across the fleet
- First dedicated R&D toward self-replication — robots assembling robots
- Cost-down program opens — the descent toward the home robot begins