In a warehouse, the final step before a package ships is often the most stubborn. After goods are picked, sorted, and conveyed, someone still has to fold a box, insert the right items, add dunnage, and slap on a label. It's a repetitive, ergonomically taxing job that has largely resisted automation. Amaru Robotics, a Mountain View startup founded in 2024, is building a robotic cell designed to drop into that exact spot, aiming to automate what it calls "the last manual step in fulfillment" [Amaru Robotics, 2024].
The Drop-In Automation Wedge
Amaru's bet hinges on minimizing disruption. The company is not selling a vision of a fully automated greenfield warehouse. Instead, it's building a dual-arm, AI-powered robotic cell that fits where a manual pack station already stands, working with the existing flow of goods and requiring no facility redesign or major IT project [Amaru Robotics, 2024]. The value proposition is practical: a Robotics-as-a-Service (RaaS) model that includes installation, service, and continuous improvements without a large upfront capital investment. For a logistics manager, the calculus is about swapping out a labor line item for a service contract, not engineering a new building. The company says its robots handle a variety of e-commerce SKUs, packing them into boxes or bags using two arms guided by AI vision [Amaru Robotics, 2024].
The Team and Early Traction
The founders are not named in public materials, but the company's careers page states the team includes second-time founders with prior experience at agricultural robotics firm Traptic, vertical farming company Bowery Farming, and autonomous vehicle developer Zoox [Amaru Robotics, 2024]. Public profiles list Paul Phan as Co-Founder and CTO, with a background at those same companies [Paul Phan Contact Information, 2026]. This pedigree suggests a team familiar with the long, hard road of deploying complex robotics in unpredictable real-world environments, from harvesting strawberries to navigating city streets.
Traction, while early, is pointed in the right direction. The company reports it has a paying customer and that its robots are "heading to production" [Amaru Robotics, 2024]. It is actively hiring for roles critical to scaling, including a senior AI Robotics Software Engineer focused on making manipulation policies reliable in live environments and an on-site operations technician to keep machines running at customer sites [Amaru Robotics, 2024]. These hires signal a transition from prototype to supported product.
The Market Tailwind and the Hard Part
The segment Amaru is targeting is small but growing fast. Packing robots represented about 21.3% of the e-commerce fulfillment robot application market in 2025, and that segment is growing at a 13.5% compound annual rate [Dataintelo, 2026]. More specifically, e-commerce companies are the fastest-growing adopters of robotic packaging arms, with installations climbing at an 18.5% CAGR [Dataintelo, 2026]. The tailwind is clear: labor scarcity and rising wages in logistics are pushing operators to automate wherever they can.
The counter-bet, however, is just as clear. Success in this niche requires overcoming significant hurdles that have tripped up many robotics startups.
- Real-world reliability. A lab demo that packs ten boxes an hour is not a product. The system must work at high speed, day after day, with millions of different SKU shapes and packaging materials, without constant human intervention.
- The service motion. Hardware breaks. The RaaS model is attractive because it offloads maintenance risk, but it also means Amaru must build a field service and support organization capable of keeping uptime promises. That is a costly, operational heavy lift.
- Economic proof. The service fee must convincingly undercut the fully loaded cost of a human packer, including turnover and management overhead, while still leaving Amaru with a healthy margin. That equation gets harder outside of peak season or in lower-wage regions.
The company's early focus on a single, repeatable task and a drop-in installation is the right wedge. The real test will be the renewal motion. After the first production cells are installed, do they work well enough that the customer orders ten more? Does the total cost of ownership justify expanding the rollout to other facilities? For now, having a paying customer moving toward production is the strongest signal that Amaru is being let in the door to find out.
The Realistic Competitive Set
Amaru's ideal customer is a third-party logistics provider (3PL) or a dedicated e-commerce fulfillment center operator already at scale, for whom labor volatility is a direct threat to margin and service-level agreements. These are pragmatic buyers, not early tech adopters. They will compare Amaru not to other venture-backed robotics startups, but to the status quo of human labor and to large, established automation integrators who might offer a more customized,and expensive,system. The competitive set is therefore bifurcated: the incumbent human workforce on one side, and bulky, fixed automation on the other. Amaru's bet is that its agile, software-updatable cell can carve out a middle path. The next twelve months will be about proving that first deployment can scale from a single station to a full pack line, turning a promising pilot into a repeatable sales playbook.
Sources
- [Amaru Robotics, 2024] Careers page | https://amaru.ai/careers
- [Amaru Robotics, 2024] Company homepage | https://amaru.ai/
- [Paul Phan Contact Information, 2026] RocketReach profile | https://rocketreach.co/paul-phan-email_39442926
- [Dataintelo, 2026] E-Commerce Fulfillment Robot Market Report | https://dataintelo.com/
- [Dataintelo, 2026] Robotic Packaging Arm Market Report | https://dataintelo.com/
- [ProjectStartups, 2026] Robotics Seed Recently Funded Startups | https://projectstartups.com/pages/robotics-seed/