top of page
Won by pricing collaborative robots on proven productivity gains rather than upfront capital cost, removing the single biggest reason most warehouses still have zero automation.
1
MODEL
BUSINESS MODEL
On-Demand Services, Product + Service Hybrid
model bm
HOW THEY BUILT IT
- Co-founded in 2019 by Anthony Jules and robotics pioneer Rodney Brooks, building Carter, a software-defined collaborative mobile robot that can switch between fulfillment picking, point-to-point transport, and mobile sorting wall functions without new hardware.
- Sells Carter exclusively through a Robotics-as-a-Service (RaaS) model where customers only begin payments once jointly defined performance metrics are met, removing upfront capital risk entirely.
- Structured a 'Crawl, Walk, Run' phased deployment model (piloted with ShipLab) that lets customers validate ROI at each stage before expanding the robot fleet or task scope.
- Partnered with Foxconn for manufacturing scale and with DHL Supply Chain as a flagship deployment (60%+ productivity gains reported at a Las Vegas facility), using a top-tier 3PL's validation as its central case study.
HOW TO ARCHITECT IT
1) Remove capital risk entirely by tying payment to proven performance metrics, because your buyer's real objection isn't 'does it work' but 'what if it doesn't.' 2) Build a single software-defined hardware platform that performs multiple warehouse functions, because most facilities can't justify single-purpose automation. 3) Structure deployment in phases (pilot, then scale) so a skeptical operations leader can say yes to a small bet before a big one. 4) Win one prestigious, large-scale reference customer (DHL) and let their name and case-study numbers do the selling to every smaller prospect afterward.
DISTRIBUTION MODEL
Enterprise Sales, Partnership Distribution
dm
HOW THEY OPERATIONALIZED
- Direct enterprise sales into 3PLs, retailers, and manufacturers, backed by pre-deployment simulations using the customer's own historical data to forecast productivity gains before a single robot ships.
- Manufacturing partnership with Foxconn specifically to scale production fast enough to meet demand from large logistics customers without building its own factory.
- Expanding geographically through existing enterprise relationships — the DHL partnership extended from North America into Latin America (Mexico retail operations) rather than opening new accounts cold in each region.
HOW TO REPLICATE WHAT WORKED
What worked: publishing a specific, credible productivity number (60%+ picking productivity gains at a DHL facility that was 'already efficient to begin with') gives prospective customers a benchmark far more persuasive than generic automation marketing claims.
The trap: RaaS pricing that defers payment until performance targets are met requires Robust AI to carry significant working-capital and deployment risk itself — a founder copying this model needs the balance sheet to absorb that risk across many simultaneous deployments before it becomes a scalable playbook rather than a cash-flow problem.
| PATTERNS OF THIS MODEL
PATTERNS IN ROBOTICS-AS-A-SERVICE WITH OUTCOME-LINKED PAYMENT:
1. TIE PAYMENT TO PROVEN PERFORMANCE AND YOU REMOVE THE REAL OBJECTION. The buyer's fear is not "does it work" but "what if it doesn't." Payments beginning only when jointly defined metrics are met eliminates capital risk — the same structural logic as outcome pricing in AI software.
2. ONE SOFTWARE-DEFINED PLATFORM ACROSS MULTIPLE TASKS beats single-purpose automation, because most facilities cannot justify a dedicated robot per function. Carter switching between picking, transport and sorting is what makes the ROI close.
3. PHASE DEPLOYMENT SO A SCEPTICAL OPERATIONS LEADER CAN SAY YES SMALL. "Crawl, Walk, Run" turns a capital decision into a sequence of validations.
4. ONE PRESTIGIOUS REFERENCE DOES THE SELLING. A reported ~60% productivity improvement at DHL's Las Vegas facility led to a five-year DHL Supply Chain alliance extending to Mexico City, plus Saddle Creek and O'Neill Logistics deployments, with Foxconn manufacturing and an Aptiv autonomy partnership.
CAUTION: RaaS defers revenue and consumes working capital — the model only survives with manufacturing partners and patient balance-sheet funding.
What companies with this model reveal
| OPPORTUNITY INTELLIGENCE
GOLDMINE 1 — REMOVE THE CAPITAL RISK, NOT JUST THE CAPITAL.
Standard: the buyer's real objection is not "does it work" but "what if it doesn't." Robotics-as-a-Service where payment begins only once jointly defined performance metrics are met transfers the risk to the vendor — and closes deals nothing else closes.
GOLDMINE 2 — ONE SOFTWARE-DEFINED PLATFORM, MULTIPLE FUNCTIONS.
Standard: Carter switches between fulfilment picking, transport and mobile sorting without new hardware. Most facilities cannot justify single-purpose automation.
GOLDMINE 3 — PHASE THE DEPLOYMENT SO "YES" IS SMALL.
Standard: a crawl-walk-run structure lets a sceptical operations leader validate ROI before scaling. Then let one prestigious reference (DHL Supply Chain, 60%+ productivity gains at a Las Vegas facility) sell to everyone smaller.
THE PIT — RaaS MEANS YOU FINANCE YOUR CUSTOMER'S CAPEX FROM YOUR OWN BALANCE SHEET.
Every robot deployed is cash out before revenue in, with payment contingent on performance you must first prove. Growth consumes capital at hardware intensity while being valued on software multiples — the structural tension in every RaaS business.
THE SECOND PIT — FOXCONN MANUFACTURING SCALE IS A COMMITMENT, NOT AN OPTION.
Contract manufacturing at scale requires volume forecasts you cannot yet make.
MOVE WITH CAUTION — WAREHOUSE ROBOTICS IS CAPITAL-SATURATED.
Amazon in-sources, and Locus, 6 River and Geek+ are funded. Founder pedigree (Rodney Brooks) buys attention, not unit economics.
Untapped Business Model / Gaps / Goldmines / Pits
Patterns & Insights
2
MARKET
mkt mt es
MARKET TYPE
Emerging Market
WHY THEY WON
Warehouse automation is a genuinely under-penetrated market — by Robust AI's own cited data, roughly 80% of warehouses lack any automation at all, not even a conveyor belt, even as the addressable market is forecast to more than quadruple over the next decade. Robust AI achieved early traction by targeting exactly the operators for whom the capital cost of automation (not the technology) was the blocker, using RaaS to remove that specific barrier. Transferable principle: in a market where the technology already works but adoption still lags, the business-model innovation (how you charge) can matter more than the product innovation.
ENTRY STRATEGY
Greenfield Entry
EXECUTION
Robust AI built its own robotics platform (Carter) and RaaS commercial model from its 2019 founding rather than entering via acquisition or licensing an existing robot maker's hardware, evidenced by its proprietary eight-camera, lidar-free semantic-mapping approach designed in-house.
FOOTHOLD STRATEGY
fs
Lighthouse Customer Strategy
DHL Supply Chain, the world's largest 3PL, served as Robust AI's lighthouse customer — a deployment whose reported 60%+ productivity gain became the credibility anchor for every subsequent sales conversation — and the company expanded from that flagship account into other 3PLs, retail, and manufacturing operators who trust DHL's operational rigor as a proxy for their own evaluation.
GROWTH CAMPAIGN
CAMPAIGNS THAT WORKED
Case-study-led enterprise marketing built around named, quantified customer outcomes (DHL's 60% productivity gain, ShipLab's phased 'Crawl, Walk, Run' rollout) combined with industry-recognition press (Fast Company's Most Innovative Companies list, RBR50 Robotics Innovation Award) to build third-party credibility in a technical buying category.
KEY LEARNING
If your buyer's objection is capital risk rather than product doubt, restructure your pricing model (RaaS, pay-on-performance) rather than just improving your product pitch. If you need enterprise credibility fast in a technical category, land one flagship, name-brand customer and quantify their results precisely — that one case study does more selling than broad marketing spend.
gc
Market Context
| MARKET INTELLIGENCE
THE STANDARD: In a market where the TECHNOLOGY ALREADY WORKS BUT ADOPTION LAGS, the business-model innovation matters more than the product innovation.
RULE 1 — IDENTIFY WHETHER THE BLOCKER IS CAPABILITY OR CAPITAL.
Robust.AI cites roughly 80% of warehouses having no automation at all. If the barrier is upfront cost rather than technical feasibility, Robotics-as-a-Service removes the actual objection — and RaaS is a financing decision dressed as a pricing page.
RULE 2 — PERFORMANCE-LINKED PAYMENT IS THE STRONGEST FORM OF THIS, AND THE MOST DEMANDING.
Payments beginning only once jointly defined metrics are met eliminates buyer risk and transfers it entirely to you. It requires pre-deployment simulation and near-certainty about outcomes.
RULE 3 — AUGMENTATION IS A DIFFERENT SALE FROM REPLACEMENT.
"Robots empower people" sells to operations leaders facing labour shortage rather than to finance teams pursuing headcount cuts. Reported ~60% day-one picking productivity gains at DHL Las Vegas is the proof shape this positioning requires.
RULE 4 — RAAS MAKES YOU A BALANCE-SHEET BUSINESS. You fund the hardware and recover over years; growth is gated by capital, not demand.
RULE 5 — LARGE 3PLs AND MANUFACTURING PARTNERS ARE THE DISTRIBUTION STRATEGY.
Five-year DHL Supply Chain alliance, Saddle Creek deployment, Foxconn for production scale, Aptiv for perception and safety certification. In hardware categories, partnerships substitute for a sales force and for capex.
EVIDENCE: Founded 2019 (Anthony Jules, Rodney Brooks); $22.5M (2020), $20M (2023), plus an undisclosed round with APL Ventures and 15th Rock (2025). Revenue and fleet size undisclosed.
MARKET TYPE: Emerging Market (warehouse automation), unlocked by business model rather than technology.
| MARKET ENTRY PLAYBOOK
THE STANDARD: IN ROBOTICS THE ENTRY DECISION IS NOT THE ROBOT — IT IS WHO CARRIES THE CAPITAL RISK. Robots-as-a-service moves it to your balance sheet in exchange for adoption.
RULE 1 — RaaS REMOVES THE BUYER'S CAPEX OBJECTION AND CREATES YOUR FUNDING PROBLEM.
Charging per hour makes the sale easy and means you finance every deployment; plan debt facilities alongside equity.
RULE 2 — COLLABORATION WITH PEOPLE IS A DIFFERENT ARCHITECTURE FROM AUTOMATION.
Working alongside existing workers in existing buildings avoids the facility redesign that blocks full automation — which is what makes short pilots possible.
RULE 3 — THE PILOT-TO-FLEET GAP IS WHERE ROBOTICS COMPANIES DIE.
Pilots prove nothing. Instrument cost per task against the human baseline from deployment one.
EVIDENCE: founded 2019; built its own Carter collaborative warehouse robot and RaaS model with a proprietary multi-camera, lidar-free mapping approach; raised $20M (2022) and a reported $42.5M (2024). Fleet size and unit economics undisclosed.
How to enter
| FOOTHOLD STRATEGY PLAYBOOK
THE STANDARD: IN INDUSTRIAL AUTOMATION, ONE FLAGSHIP DEPLOYMENT WITH A PUBLISHED PRODUCTIVITY NUMBER IS THE ENTIRE SALES ASSET. Operations buyers do not buy capability; they buy a result someone comparable already measured.
RULE 1 — WIN THE OPERATOR WHOSE RIGOUR IS A PROXY FOR EVERYONE ELSE'S. A deployment at the world's largest third-party logistics provider substitutes for the prospect's own evaluation, because they trust that firm's testing more than their own.
RULE 2 — LOW BARRIER TO ENTRY BEATS HIGH CAPABILITY IN BROWNFIELD SITES. A robot that delivers value on day one without warehouse redesign, integration or new infrastructure reaches ROI faster than a fully autonomous system requiring a rebuild — and speed to ROI is what buyers actually optimise.
RULE 3 — PHASED CAPABILITY IS A COMMERCIAL DESIGN, NOT A TECHNICAL COMPROMISE. Starting with motor-assisted movement and activating autonomy later lets sites of different maturity buy the same product.
RULE 4 — POSITIONING AUTOMATION AS AUGMENTATION REMOVES THE INTERNAL BLOCKER. Where workforce opposition can stall a deployment, "empower people, not replace them" is a procurement strategy.
EVIDENCE: DHL Supply Chain was the lighthouse: Carter delivered over 60% productivity gains within weeks at DHL's Las Vegas facility, and that number anchored subsequent sales. In December 2025 DHL signed a five-year strategic alliance to deploy Carter in Mexico, starting with 15 units and showing ~30% gains there, with WMS integration planned by 2026. Robust.AI added Saddle Creek Logistics, a Foxconn manufacturing partnership and an Aptiv perception partnership, and raised a round with APL Ventures and Japan's 15th Rock (amount not disclosed). Founded 2019 by Anthony Jules and Rodney Brooks. Revenue and total funding are not comprehensively disclosed.
How to get the first strong position
MARKET PATTERNS & PLAYBOOK
3
MONEY
money rev pri
REVENUE MODEL
Subscription, Performance Fees
PRICING MODEL
Value-Based Pricing
WHY THEY WON
Recurring RaaS subscription fees per deployed robot, explicitly structured so customer payments only begin once jointly agreed productivity/performance metrics are met — effectively a performance-fee gate layered on top of a subscription, aligning Robust AI's revenue directly with the customer's realized ROI.
Price is not tied to hardware cost or a flat per-robot rate but to demonstrated performance improvement at each customer's specific facility, validated through pre-deployment simulation using the customer's own historical operational data before any commercial terms are finalized.
TARGET AUDIENCE
CUSTOMER BUYING BEHAVIOUR
tg cb
Third-party logistics providers (3PLs), retail and manufacturing operations needing flexible warehouse automation without large capital outlays
Procurement-driven, pilot-first evaluation (phased 'Crawl, Walk, Run' deployment) with operations leadership requiring proof of productivity gains before scaling fleet size or task scope
| PRICING INTELLIGENCE
What makes this model effective & make customers pay
THE STANDARD: When your product replaces labour that cannot be hired, price against the unfillable role. Robotics is sold per-hour against a wage, not per-unit against a machine.
RULE 1 — ROBOTS-AS-A-SERVICE REMOVES THE CAPITAL APPROVAL THAT KILLS MOST DEALS.
A monthly fee compared to a warehouse wage is an operational decision. A six-figure purchase is a board decision with a much longer cycle and a much lower close rate.
RULE 2 — COLLABORATIVE DEPLOYMENT PRICES LOWER RISK, WHICH IS WHAT THE BUYER IS ACTUALLY BUYING.
A robot that works alongside existing staff and existing layouts avoids the facility redesign that makes full automation unaffordable. Incremental adoption is the pricing advantage.
RULE 3 — YOUR PRICE CEILING IS A FULLY-LOADED WAGE INCLUDING TURNOVER AND RECRUITMENT.
Warehouse labour churn is severe and constant. Include the cost of continuously rehiring, not just the hourly rate.
RULE 4 — LABOUR SCARCITY, NOT LABOUR COST, IS THE REAL DRIVER.
Operators buy because they cannot fill shifts at any price. That is a stronger and more durable motivation than savings, and it survives a downturn better.
DISCLOSURE: Robust AI does not publish pricing, deployment counts or current revenue in reliable public sources; funding figures vary between trackers.
THE WILLINGNESS-TO-PAY INSIGHT: The operations director is buying certainty that tomorrow's shift is covered. Where your customer's constraint is people they physically cannot recruit, price against the hire that never happened — an unbounded number, because the alternative is orders going unfulfilled.
PRICE & REVENUE
| Revenue Risk - The biggest threat to revenue stability
THE STANDARD: Gating payment on jointly agreed productivity metrics is the most customer-friendly structure in robotics and the most punishing on working capital — you fund the hardware and wait.
RULE 1 — PERFORMANCE-GATED REVENUE MEANS YOU CAPITALISE THE PILOT. Robots are built, shipped and deployed before any payment begins. Cash conversion is measured in quarters, and a failed metric means the cost stays with you.
RULE 2 — THE METRIC IS NEGOTIATED WITH THE CUSTOMER, SO YOUR REVENUE DEPENDS ON THEIR MEASUREMENT. Disputes over whether the productivity target was met are collection risk with no clean escalation path.
RULE 3 — MARQUEE LOGOS CONCENTRATE REVENUE RATHER THAN DIVERSIFYING IT. A five-year strategic alliance with DHL Supply Chain (Mexico City), a reported 60% productivity improvement at DHL Las Vegas, and a Saddle Creek Logistics partnership are strong proof points — and they mean a small number of relationships carry the plan.
RULE 4 — WAREHOUSE AUTOMATION DEMAND FOLLOWS 3PL CAPEX CYCLES. Deployment budgets are approved annually and frozen quickly when freight volumes soften.
RULE 5 — CAPITAL INTENSITY MEETS A CATEGORY AWASH IN MONEY. Humanoid and warehouse robotics competitors have raised at multi-billion valuations; a RaaS model with deferred revenue must keep raising against them.
NOT DISCLOSED: no revenue, ARR, fleet size or unit economics published. Robust.AI was ranked #2 in Fast Company's Robotics and Engineering category (2026).
Where the model can break
4
MOTION
GROWTH EXPANSION MODEL
COMPETITIVE STRATEGY
motion ge cs
Geographic Expansion
HOW THEY EXPAND
Against traditional single-function automated guided vehicles (AGVs) and fixed conveyor systems, Robust AI differentiates on its software-defined, multi-function robot (picking, transport, sorting in one unit) plus its zero-capex RaaS pricing, rather than competing purely on hardware specs or price per unit.
Differentiation
HOW THEY COMPETE
Against traditional single-function automated guided vehicles (AGVs) and fixed conveyor systems, Robust AI differentiates on its software-defined, multi-function robot (picking, transport, sorting in one unit) plus its zero-capex RaaS pricing, rather than competing purely on hardware specs or price per unit.
GROWTH ENGINE
GTM
ge n gtm
Partnership Growth
A successful, quantified deployment at one facility of a large customer (DHL) creates internal pressure and precedent to expand Carter to additional facilities within the same company, while the published productivity numbers attract new enterprise prospects evaluating automation for the first time — the loop is limited by manufacturing capacity and the operational bandwidth to run simulations and phased pilots for each new account.
Enterprise sales built on quantified case studies and staged pilot programs, reinforced by manufacturing-scale partnerships (Foxconn) that let Robust AI credibly promise fast delivery to large logistics customers evaluating automation for the first time.
SUSTAINING MOATS
Switching Costs, High Customer Lock-In, Brand Power, Technology Advantage (complex enterprise scenarios)
moat
Every new deployment adds real-world operational data (facility layouts, task patterns, productivity benchmarks) that sharpens Robust AI's pre-deployment simulation accuracy, making its ROI promises more reliable than a newer entrant's, while its Foxconn manufacturing partnership and DHL enterprise relationship compound into distribution advantages a new robotics startup would need years to build.
| MOAT INTELLIGENCE
THE STANDARD: In robotics the moat is deployment economics, not autonomy. Whoever makes the robot pay back inside a customer's budget cycle wins, regardless of whose AI is better.
RULE 1 — SELL LABOUR ARBITRAGE ON A PAYBACK PERIOD, NOT CAPABILITY. Warehouse operators buy against a labour-cost line. A better robot paying back in four years loses to a simpler one paying back in twelve months.
RULE 2 — COLLABORATIVE DESIGN AVOIDS THE MOST EXPENSIVE BARRIER: FACILITY CHANGE. Robots that work alongside people in existing aisles remove the capital project that kills most automation deals. That is a commercial advantage, not an engineering one.
RULE 3 — FOUNDER REPUTATION RAISES CAPITAL AND DOES NOT SHIP UNITS. A celebrated pedigree opens funding and press. Deployment counts are the only evidence that matters, and they are what this category consistently fails to produce.
RULE 4 — ROBOTICS-AS-A-SERVICE SHIFTS CUSTOMER RISK ONTO YOUR BALANCE SHEET. It closes deals and converts you into a capital-intensive leasing business.
EVIDENCE:
- Robotics company founded by a team including Rodney Brooks (iRobot, Rethink Robotics), building collaborative warehouse robots designed to work alongside human workers without facility reconfiguration.
- I DID NOT VERIFY CURRENT FUNDING, VALUATION, DEPLOYMENT COUNT, REVENUE OR OPERATING STATUS. Confirm before citing — this category moves fast and several well-funded peers have failed.
- Structural context: contested by Amazon Robotics internally, Locus Robotics, 6 River Systems (Ocado), Fetch (Zebra) and a wide AMR field. Rethink Robotics, a prior venture by the same founder, shut down in 2018 despite strong technology and reputation.
THE SIGNAL: the reference case is the founder's own previous company — excellent robots, celebrated leadership, insufficient unit economics. In robotics, the moat is the payback period.
Why this company remains defensible
ARR & TAKEAWAY
ARR Journey - what to do at each stage
PRE-$1M ARR — NEGATE THE ASSUMPTION EVERYONE SHARES
The founding method is explicit: find what every competitor assumes implicitly and negate it. Here that meant no manipulators, no humanoid form, and a human always in the loop.
Design for the worker, not the demo. A cart that can be grabbed and pushed is adopted; a machine that cannot be overridden is resisted.
Assemble credibility deliberately (co-founders include Rodney Brooks; the founding group drew from SRI, Alphabet and academia).
$1–5M ARR — PAID TRIALS, NOT FREE PILOTS
A paid trial in a live distribution centre tells you what a free pilot never will.
Sell throughput per worker; warehouses buy labour productivity, not autonomy.
WATCH: robots per site after 90 days. In-facility expansion is the only real proof.
$5–10M ARR — SELL FLEETS TO 3PLs, NOT UNITS TO WAREHOUSES
Third-party logistics operators run many facilities and replicate what works. (Reported deployments include a 3PL running 24 robots across two distribution centres.)
Price as robotics-as-a-service so the customer pays operating cost, not capital cost.
$10–50M ARR — HARDWARE ECONOMICS DECIDE EVERYTHING
Unit cost, service, spares and uptime determine whether RaaS margins turn positive. Instrument cost-to-serve per robot per month.
Capitalise for manufacturing scale-up — this is where hardware companies die.
$50–100M ARR — THE CATEGORY'S FUNDING CYCLE IS A RISK
Brooks has publicly argued the humanoid robotics boom will deflate. A correction reprices even well-run companies.
Hold cash, keep deployments revenue-positive, avoid dependence on a single large customer.
NOTE PLAINLY: no revenue disclosed; band placement is inference.
$100M+ ARR — NOT IN EVIDENCE
Rule: in physical AI, hype funds the category and deployments decide who survives it. Optimise for robots working in real facilities, not for the narrative.
COPY PLAYBOOK : What Worked → What Failed → What to Replicate → What to Avoid
THE STANDARD: Guaranteeing the productivity outcome removes the buyer's risk and transfers it entirely onto your balance sheet — which only works if you can carry that risk across many simultaneous deployments.
SEQUENCE:
1. Publish a specific, credible number from a demanding site: 60%+ picking productivity gains at a DHL facility already considered efficient.
2. Position robots as augmenting workers rather than replacing them, which materially eases adoption with operations leaders and unions.
3. Sell Robotics-as-a-Service with payment starting only once jointly defined performance metrics are met.
4. Make one hardware platform software-defined so the same robot serves picking, transport and sorting — flexibility without new capital per use case.
5. Secure manufacturing and component partnerships (Foxconn for production scale, Aptiv for perception and safety certification) rather than building everything.
WHAT WORKED:
- A guaranteed, quantified outcome written into customer agreements, which is far more persuasive than generic automation marketing.
- Converting a pilot into a five-year strategic alliance with DHL Supply Chain, plus deployments with Saddle Creek and others.
- Roughly $42.5M raised across rounds including APL Ventures and 15th Rock — modest capital for a robotics company at this deployment stage.
CAUTIONS:
1. DEFERRED-PAYMENT RaaS MEANS YOU FINANCE THE DEPLOYMENT. Working capital and hardware risk sit with you until targets are hit; without a balance sheet to absorb that across many sites simultaneously, it is a cash-flow problem rather than a scalable model.
2. THE LATEST ROUND SIZE WAS NOT DISCLOSED, and no valuation is public — for a hardware company scaling deployments, that opacity is itself worth noting.
3. WAREHOUSE ROBOTICS IS CAPITAL-HEAVY AND CROWDED, with far better-funded competitors and 3PL customers who can wait out any vendor's timeline.
bottom of page