Article
Two announcements landed within a week, and together they show exactly where humanoid robots stand. Tesla says it has started building Optimus on its high-volume lines in Fremont. GXO, which calls itself the world's largest pure-play contract logistics provider, says it's putting Agility Robotics' Digit humanoids into two more U.S. warehouses. Robots are shipping. What isn't shipping yet is proof that they pay for themselves.
What Tesla actually said
In its third-quarter update, filed with the SEC on October 21, Tesla wrote that Optimus production began in Fremont on its high-volume lines and that these are the first production-quality bots it has built. Its capacity table lists Optimus in California as "Early Ramp" and the Texas line as under construction. Elon Musk said he expects output to reach 20,000 robots a week in about a year, with deliveries starting late next year.
That's a company goal, and the track record matters. Musk once hoped Optimus would be production-ready by 2023. In 2024 he projected limited production in 2025, with more than 1,000 units working inside Tesla. Tesla also hasn't said which generation of Optimus the Fremont line is building. In the same quarter, net income fell 43% to $787 million as spending on AI and robotics rose.
What GXO and Agility actually said
On October 26, GXO and Agility announced Digit deployments at GXO sites in Statesville, North Carolina, and Lebanon, Tennessee, supporting e-commerce fulfillment. The rollouts start in early 2027 with Digit 4 and move to Digit 5 once it's commercially available.
The track record is real but modest. Since a 2024 multi-year robots-as-a-service deal, Digit has moved more than 100,000 totes at GXO's Flowery Branch, Georgia site, mostly picking up totes and placing them on conveyors. Agility says Digit has logged more than 65,000 hours of real-world work across customers, and its Oregon factory is designed for up to 10,000 robots a year. Agility is also going public through a merger with Churchill Capital Corp XI, expected to close this quarter under the ticker AGLT.
The skeptical counterpoint
GXO's own CEO supplied it. On the company's second-quarter earnings call, Patrick Kelleher said GXO had run 45 humanoid pilots and hadn't achieved ROI on humanoids yet. Asked how many of the nearly 20,000 robots GXO plans to run this year would be humanoids, he said none would be in production this year. Conventional warehouse robots, he noted, already earn their keep at scale.
So why expand? Because the contract limits the downside. Under robots-as-a-service, the customer pays a recurring fee instead of buying the machine. An Agility investor presentation, as reported by analyst Rajesh Beri, gives an illustrative price of about $8,500 a month per robot, roughly $100,000 a year, against about $200,000 a year in fully loaded labor it says one robot replaces. That's Agility's estimate, not a published price or a proven saving.
Why it matters to builders
The humanoid race is moving from demo videos to procurement contracts, and regulators are adjusting. South Korea, for example, cleared its first factory-floor test of 20 humanoids in October, after years of case-by-case sandbox approvals. The nearest opportunities aren't in building the robot. They're around it: systems integration, task training, fleet software, safety certification, maintenance, and financing models that let a warehouse try a humanoid without betting the balance sheet.
Your takeaways
- Copy the business model, not the hype. Robots-as-a-service turns a capital purchase into an operating expense the customer can cancel. That logic works for equipment, software and AI agents too.
- Judge every humanoid claim by one number: cost per completed task against the person or conventional robot doing that task today.
- Watch the first quarter of 2027. That's when GXO's new sites go live and when Tesla's delivery promise starts meeting the calendar.
Leave a comment