Nov 2, 2026, 11:05 AM ETInterview

What iPhone Packaging Reveals About the Robotics Opportunity

Current Operations: Tutor Intelligence employs ~160 people based primarily in Boston, with additional Go-to-Market (GTM) and technician teams across the U.S.
Deployment Metrics: Tutor robots have helped manufacture and ship over 250 million products to American consumers.
Customer Base: Deployed across a spectrum from small "mom and pop" businesses to Fortune 50/100 supply chain leaders; majority of U.S. customers are new robot buyers (greenfield market).
Market Context: In 2025, the U.S. deployed only 38,000 industrial robots, leaving ~90% of U.S. factories and warehouses without any robotics, representing a massive untapped opportunity.

Full-Stack Philosophy: CEO Josh Gruenstein argues robotics requires a full-stack approach (hardware, software, models, supply chain, customers), not just isolated layer innovation.
Data Flywheel: The company utilizes human "tutors" to tele-operate robots and collect label data, creating a networked dataset for training frontier AI models.
Economic Alignment: The business model ensures data collection is an economic byproduct of revenue-generating tasks, avoiding the "no guarantee of valuable data" risk of passive data scraping.
Unit Economics Focus: The strategy prioritizes trade-offs across the stack to lower the cost per task, competing directly with human labor rather than traditional legacy automation.
Hardware Strategy: Tutor designs and manufactures its own robot bodies in Watertown, Massachusetts, while sourcing off-the-shelf components (actuators, cameras) from OEMs.

Cassie: A single-arm, brute-force robot designed for heavy-duty work (e.g., pallet/case transport), utilizing foundation models (200-400M parameters) to predict affordances and grasps.
Sunny: A wheeled, semi-humanoid robot designed for 3-10 year reliability in industrial settings, featuring higher payload, speed, and collaborative safety standards for bimanual manipulation.
Form Factor Strategy: Currently covers 80-90% of manufacturing/logistics tasks; future expansions will target legs for outdoor/construction use and different form factors for home environments.
Current Limitations: Robots cannot currently operate on unpaved surfaces (gravel) or in unstructured outdoor environments due to traction and hardware limitations.

Sales Cycle Speed: Average time from first call to contract signing is 40-60 days; average time from contract to on-site production is ~40 days.
Deployment Velocity: Robots can be standing up in production within hours or a day of arrival due to standardized, non-customized deployment.
Risk Mitigation: The company removes upfront Capital Expenditure (CapEx) risk for customers, facilitating faster adoption among risk-averse manufacturers.
Pricing Model: Pricing includes both upfront hardware costs and ongoing operational costs (inference, teleoperation, maintenance), shifting from traditional high CapEx ROI models to ongoing OpEx.
User Experience: Customers view working robots as "machines" rather than "beings" once functional, though Human-Robot Interaction (HRI) strategies (e.g., naming robots) foster ownership and reduce friction among line workers.

Model Iteration: The company releases software updates daily and hardware iterations every two weeks, contrasting with the industry norm of single-version deployments.
Simulation Integration: Leveraging "frontier models" (e.g., ASTR) and 3D spatial training data to accelerate learning, allowing robots to apply general knowledge to specific physical tasks.
Reinforcement Learning: Utilizing a "Data Factory" fleet of 100 robots in Boston to conduct real-world reinforcement learning at scale, a feat previously impossible for academic or small industry teams.
Supply Chain Verticalization: Tutor plans to continue verticalizing by bringing more component manufacturing in-house as volume grows, aiming to build a mature U.S. robotic supply chain.

U.S. vs. China Dynamics: While global supply chains remain interconnected, the U.S. is building industrial capacity via greenfield general-purpose robotics, whereas China faces the challenge of retrofitting legacy, specialized automation infrastructure.
Future Labor Impact: The availability of flexible robotic labor is expected to lower the cost of physical goods, potentially shifting manufacturing back to the U.S. and enabling a "cloud-like" proliferation of physical businesses.
Market Correction: The industry is moving away from "custom machine" engineering toward "general purpose" platforms, mirroring the evolution of computing from specialized hardware to general CPUs/GPUs.
Talent Needs: Tutor is hiring across every layer of the stack, including researchers, engineers, hardware specialists, and technicians, emphasizing the need for interdisciplinary teams capable of managing the entire system.