Join the movement and participate on Uniswap.
Participate
Nils just returned from China with two clear signals: the hardware for real-world data collection is arriving faster than expected, and Auki's existing customer base puts it in an unusually strong position to capitalize.
The team visited a major data factory in Changzhou with 150 robots running teleoperated data collection. These facilities produce high-quality manipulation data, but the scenarios remain artificial. The bigger development is new hardware — a headset plus handheld grippers with wrist tracking — that lets humans collect robot-friendly egocentric data while actually doing the job in real stores.
This changes the economics dramatically. Teleoperation data sells for $90–$200 per hour. The new approach is estimated to be 20 to 50 times cheaper. With the hardware expected in September, Auki can begin collecting real customer-environment data this year.
Nils noted the strategic fork this creates:
"This is a big discussion internally for us like what path do we want to go — would we want to sell this data or do we want to become a model company ourselves."
The data would come from live retail environments rather than controlled Chinese or Malaysian replicas, giving Auki (or its partners) a differentiated training set for tasks like product fronting.
On the Realman RS-02 side, the team shifted to Cartesian movements, removed unnecessary full-arm motion, and increased speed. The immediate next milestone is simultaneous two-arm operation — "dual wielding."
This week the robot achieved verifiable full-fixture autonomy: it can now scan an entire shelf without teleoperation and return structured data to the store owner. Current speed on a 10-meter fixture is roughly one hour; dual-arm planning is expected to improve that significantly.
Robin’s camera-based scanning flow for Cactus removed the need for an external barcode scanner. The system now reads barcodes directly from the camera feed and places them correctly in 3D space. One retailer who had previously stepped back because of friction is now back in play after seeing the latest demo — a potential $8 million per year contract.
The same computer-vision improvements benefit both the phone-based copilot and the robots.
JB and Charlie added self-service portal generation to WebDMT. Customers and third-party developers can now upload their own designs, generate Auki-enabled QR codes, and print locally.
Planogram tooling also advanced. Store managers can now visualize current placement data against product data, design the desired shelf layout, and automatically generate tasks for staff to execute the change. The interface is a clear step above legacy planogram software.
Nils highlighted an important clarification for future positioning: from the customer’s perspective, Auki already delivers a more vertically integrated solution than many "vertically integrated" robotics companies. Auki provides the full operating layer — perception, copilot, planograms, tasking, and integration with retailer systems — even while partnering on the robot hardware.
A new, more realistic demo store is under construction by one of Auki’s investors. It will include multiple shelving types, freezers, coolers, and glass-door units. Setup is targeted for mid-August.
Auki is making the physical world accessible to AI by building the real world web: a way for robots and digital devices like smart glasses and phones to browse, navigate, and search physical locations.
70% of the world economy is still tied to physical locations and labor, so making the physical world accessible to AI represents a 3X increase in the TAM of AI in general. Auki's goal is to become the decentralized nervous system of AI in the physical world, providing collaborative spatial reasoning for the next 100bn devices on Earth and beyond.
X | Discord | LinkedIn | YouTube | Whitepaper | auki.com
