Greetings Stellar Community!
We hope you’re enjoying the summer. At Stellar, we spend the quieter months going deep on the sectors we believe will define the next decade—and right now, none has our attention quite like robotics and physical AI.
Our team recently spent a day at Citi’s Robotics & Physical AI Leadership Conference with the founders, operators, and investors building the industry. We came away more convinced than ever that the sector is moving from demo to deployment in real time.
A quick note on terminology: physical AI refers to the intelligence layer—the models that allow machines to perceive, reason, and act. Robotics describes the broader field of machines performing physical work, while humanoids are the human-shaped subset.
Where We’ve Been Spending Time
Capital and talent are pouring into physical AI at a pace we haven’t seen since the early LLM wave—roughly $20 billion over the past 24 months. The investment thesis rests on the belief that LLM-style scaling can transfer to robots, supported by autonomy proof points from platforms such as Waymo and FSD, more than $500 million of DoD demand, and a falling hardware cost curve.
This cycle is already generating revenue, with early traction concentrated in applications where the return on investment is clearest.
Why We’re Excited
• Metrics, not hype. The conversation has moved to ROI, payback periods, and operating metrics such as the work-to-charge ratio. One humanoid platform improved from 4:1 to 10:1, unlocking true three-shift operation.
• Adoption runs on RaaS, rental, and leasing. These models allow fleets to deploy with limited upfront capex while creating deep workflow integration, recurring revenue, and low churn. Locus, Bear, and Botrista are all scaling through this approach.
• Two positioning axes are emerging. Companies can be assessed across general hardware versus purpose-built software, and “engineering” platforms that build robots versus “enablement” platforms that make automation more accessible. Where a company sits helps define its moat.
• Scalable applications outperform custom skills. Reusable applications that can be deployed across an entire fleet compound more quickly than bespoke dexterity solutions because reusability drives the economics.
• Humanoids are a distinct category. They are designed for social, multisensory engagement across entertainment, education, and eventually the home, following their own adoption curve.
• Safety and cost leadership will determine the winners. Safety certification is becoming both a competitive moat and the license to operate outside the cage, while cost leadership will determine which platforms can scale.
Where It’s Already Working
Unlike many emerging AI applications, robotics and physical AI are already deployed and generating revenue:
• Warehousing and logistics: Autonomous mobile robots and picking systems are operating at scale. Locus runs approximately 17,000 robots, while Dexterity’s systems load trucks for FedEx, UPS, and GXO.
• Defense and security: Drones and autonomous systems support more than $500 million of DoD programs, with national-security priorities pulling additional investment and production onshore.
• Manufacturing: Browser-designed automation cells from Vention and AI-formed metal components from Machina Labs are compressing tooling and deployment timelines from months to days.
• Hospitality and service: Bear Robotics and Botrista have deployed service and beverage robots across thousands of locations.
• The home: Consumer deployment remains the highest bar and will likely come after humanoid platforms have cleared industrial operating, reliability, safety, and cost hurdles.
Later-Stage Companies We’re Following
The following are select examples provided for industry context only:
• Dexterity: Develops physical AI systems for warehouses and cargo operations. Its dual-armed Mech “superhumanoid” loads trucks and palletizes items weighing approximately 130 pounds. Last disclosed round: $95 million at a $1.65 billion valuation in March 2025.
• Vention: A browser-based, full-stack automation platform now adding physical AI capabilities such as AI-driven bin picking across more than 3,000 factories, including Toyota, GE, and Tesla. Last disclosed round: $95 million Series C in 2022; valuation undisclosed.
• Machina Labs: Develops AI-powered “robotic blacksmiths” whose Roboforming technology shapes large, complex metal components without conventional tooling. Last disclosed round: $32 million Series B in 2023; valuation undisclosed. Founded by former SpaceX employees.
• Dyna Robotics: Builds foundation models for dexterous robotic manipulation. Its DYNA-1 system operates continuously at greater than 99% task accuracy. Last disclosed round: $120 million Series A at a valuation exceeding $600 million in September 2025.
What’s Next
We expect to remain active in evaluating category leaders while closely monitoring two significant risks: commoditization—particularly at the hardware layer—and intensifying international competition.
If any of these themes resonate, reply and let’s discuss. We value the investment community that we are surrounded by and love the dialogue. We would be glad to share our deeper work on the sector or any of the companies discussed above.
Warmly,
Stellar GCP Team
Important Disclosure
This communication is provided by Stellar Growth Capital Partners and its affiliates solely for informational and discussion purposes. It does not constitute, and should not be construed as, an offer to sell, a solicitation of an offer to purchase, or a recommendation regarding any security, investment product, or investment strategy. References to companies are illustrative only and do not indicate any current or prospective investment. The views expressed are as of the date of this communication and may change without notice. Certain information has been obtained from third-party sources believed to be reliable, but its accuracy or completeness has not been independently verified, and no representation or warranty is made regarding such information. Forward-looking statements are inherently uncertain, and actual outcomes may differ materially. Recipients should conduct their own diligence and consult their legal, tax, and financial advisers before making any investment decision.