Atlas
A humanoid robot developed by Boston Dynamics. Early versions of Atlas featured only a small number of degrees of freedom in the hands, and the platform is used as a comparative reference when assessing the evolution of dexterous hand technology in humanoid robotics.
Core Facts
Tesla's Optimus, Figure AI's Figure 02, and Boston Dynamics' Atlas are the leading humanoid robot products at present.
AI has safety and alignment restrictions, and legitimate requests are sometimes mistakenly intercepted.
The hardware cost of humanoid robots is generally in the hundreds of thousands of yuan range, which is one of the cost challenges for the commercial deployment of embodied intelligence.
Timeline (last 90 days)
Boston Dynamics' Atlas and Tesla's Optimus both rely on reinforcement learning training to achieve balance and recovery.
Airlock, Atlas, and Passport are currently in the preview stage and still some distance from production availability, requiring waiting for open trials or official release.
The simultaneous launch of Airlock, Atlas, and Passport shows that WorkOS is moving toward a more complete enterprise infrastructure platform, not content to be just an SSO vendor.
WorkOS's official recap did not provide detailed feature lists, APIs, pricing, or technical architecture information for the three products Airlock, Atlas, and Passport.
At the init() 2026 conference, WorkOS showcased three products—Airlock, Atlas, and Passport—in the form of live previews.
Boston Dynamics activated an Atlas robot training center at Hyundai's US factory campus to train parts logistics, sorting, and placement tasks.
World Labs released a new model called Atlas earlier this month that can predict the view after a camera position change and reconstruct an interactive 3D scene from only a small number of 2D images.
Boston Dynamics' Atlas is a representative product in the field of bipedal robots.
Boston Dynamics platforms such as Atlas and Spot are typically priced in the hundreds of thousands of dollars range and are primarily aimed at specialized scenarios such as defense and industrial inspection.
Airlock originated from an internal experimental project called Atlas.
40 more timeline events
All Facts (20)
Tesla's Optimus, Figure AI's Figure 02, and Boston Dynamics' Atlas are the leading humanoid robot products at present.
80%VerifiedAI has safety and alignment restrictions, and legitimate requests are sometimes mistakenly intercepted.
65%VerifiedThe hardware cost of humanoid robots is generally in the hundreds of thousands of yuan range, which is one of the cost challenges for the commercial deployment of embodied intelligence.
65%VerifiedThe context silos between AI platforms are more the result of business model choices than a technical impossibility.
65%UnverifiedAtlas is a world model project oriented toward Spatial Intelligence.
70%UnverifiedAtlas can accept multi-modal inputs including text, images, video, and 3D data, and uniformly convert them into controllable high-definition video
60%UnverifiedBoston Dynamics' Atlas and Tesla's Optimus both rely on reinforcement learning training to achieve balance and recovery.
50%UnverifiedWorkOS's official recap did not provide detailed feature lists, APIs, pricing, or technical architecture information for the three products Airlock, Atlas, and Passport.
50%UnverifiedThe simultaneous launch of Airlock, Atlas, and Passport shows that WorkOS is moving toward a more complete enterprise infrastructure platform, not content to be just an SSO vendor.
50%UnverifiedBoston Dynamics' Atlas is a representative product in the field of bipedal robots.
50%UnverifiedBoston Dynamics platforms such as Atlas and Spot are typically priced in the hundreds of thousands of dollars range and are primarily aimed at specialized scenarios such as defense and industrial inspection.
50%UnverifiedAirlock originated from an internal experimental project called Atlas.
50%UnverifiedThe unified generation and reconstruction architecture allows the model to share the same underlying representations and prediction capabilities, eliminating the need to design separate mechanisms for each.
50%UnverifiedWhen input observations are sufficient, Atlas tends toward precise reconstruction; when observations are sparse or missing, Atlas invokes generative priors to complete the visual content.
50%UnverifiedTeam members involved in the development of Atlas include @jcjohnss, @BenMildenhall, and @martin_casado, among others.
50%UnverifiedAtlas reduces both generation and reconstruction to the problem of predicting the next frame/next view pixel content given existing visual context.
50%UnverifiedThe Atlas world model positions "Next View Prediction" as the core hub of its model architecture.
50%UnverifiedAtlas unifies pixel-level generation and reconstruction within a single framework.
50%UnverifiedThird-party markets in the MCP ecosystem can connect to cloud services such as Atlas and Cloudflare, as well as integrate local CLI tools such as ffmpeg.
50%UnverifiedAtlas uses Slack as its sole carrier.
50%