Sharpa Raises $668M as Its Robot Runs Dairy Queen's 55-Step Workflow
The AMW Read
It is a meaningful segment-level commercialization test with a $668 million round, but it largely confirms the same funding and Dairy Queen launch AMW covered days earlier.
Sharpa Raises $668M as Its Robot Runs Dairy Queen's 55-Step Workflow
Sharpa said it has raised more than RMB 4.5 billion ($668 million) at a RMB 22 billion ($3.3 billion) post-money valuation, backed by Alibaba, Meituan, Tencent, JD.com, Transsion, and venture firms. At Dairy Queen's Wujiang Road outlet in Shanghai, operated by CFB Group, it says its robot can prepare and hand over a Blizzard Treat through 55 linked steps, using the store's existing equipment, ingredients, layout, and procedures.
The commercial claim matters because Sharpa is testing a core bottleneck in physical AI: whether a robot can complete a long, interdependent workflow inside a site designed for people rather than a specially engineered cell. Sharpa says the system links task planning, action execution, state assessment, and error correction, and is intended to operate during the store's 10 a.m. to 10 p.m. business day without continuous human control. A repeatable zero-retrofit deployment would make restaurant automation easier to replicate; a one-off demonstration would not settle the reliability or economics question.
For builders, the bar is now operational rather than purely technical: demonstrate intervention rates, uptime, task-completion quality, and deployment cost across normal order fluctuations. For investors, the $668 million raise makes Sharpa a well-funded test of whether dexterous manipulation can move from OEM and research settings into revenue-bearing service sites, but the key evidence will be repeat performance across locations and workflows rather than a single store's launch.
#Robotics #EmbodiedAI #DexterousManipulation #RestaurantAutomation #AIInvestment
