Our featured investment: Lightberry

Lightberry takes the spotlight as Singularity Capital’s most recent featured investment.

Lightberry is building the missing software layer for humanoid robots: a social, context‑aware “brain” that makes hardware commercially deployable rather than just demo‑ready. Their software enables humanoid robots to listen, speak, and act autonomously, making them contextually aware, customizable, and emotionally intelligent.

Lightberry was co-founded in 2025 by Ali Attar and Stephan Koenigstorfer, both seasoned entrepreneurs with deep technical and entrepreneurial pedigrees. Ali Attar is also the founder of SigmaOS (YC S21), a high-performance macOS browser, and has prior startup experience with Loop and Dromadis Health. He’s a product engineer with a track record in AI-driven consumer tech and human-robot interaction. Stephan Koenigstorfer holds a PhD, MBA, and MSci, and brings expertise from his time at CERN, as well as robotics and software engineering roles at startups like Bucket Robotics and Stack AV. Their combined backgrounds in AI, robotics, and consumer product development position Lightberry at the forefront of human-robot interaction innovation.

The Problem: Impressive Hardware, Socially Broken Software

The current state of humanoid robotics reveals a striking asymmetry. Hardware manufacturers have cracked the physics problem. Unitree's G1, a $35,000+ platform, can walk autonomously, grasp objects with precision, and execute complex motor sequences. Booster's T1, priced in the $20,000–$30,000 range, offers similar kinematics in a more affordable package. These are no longer laboratory curiosities, they are deployable platforms.

Yet in practice, humanoid robots deployed in public environments (conferences, retail locations, offices) consistently demonstrate the same behavioral deficiency: they are not socially autonomous. They either require a remote operator to drive behavior in real time, or they suffer from a brittle chatbot problem. When generic large language models are simply bolted onto robot hardware, the result is reliably poor:

Lack of situational awareness: Robots respond to every ambient noise, every stray voice command, breaking the flow of human conversation and creating the impression of incompetence.

Absence of personality: The robot speaks in the flat, undifferentiated voice of a generic API call. There is no sense that the machine has preferences, boundaries, or an internal life.

Dependency on teleoperation: When the pre-programmed behaviors fail (which they do, frequently, in unstructured human environments), the robot defaults to a remote operator. This destroys the sense of autonomy and makes deployment uneconomical at scale.

Rigidity outside scripted scenarios: The robot cannot adapt to contextual variation. It freezes when confronted with behavior that doesn't match the narrow band of training scenarios.

Why Lightberry?

Lightberry’s core insight is that the missing piece is not more metal, but a robust social operating system for robots: a brain that fuses perception, navigation, dialogue, and personality into a single, configurable layer. By abstracting the complexity of multimodal perception into a software platform that works across multiple manufacturers, Lightberry aims to become for humanoids what MS‑DOS or Windows were for early PCs, a standard interface that unlocks a mass market around a fragmented hardware ecosystem.

Lightberry implements what they call an "internal monologue" system, where the robot maintains an ongoing reasoning process about context, intent, and appropriate action. Instead of a reactive chatbot that responds to every input stimulus, the robot decides when to engage, what to do, and how to respond, based on a coherent model of the situation.

Why This Matters Now

Humanoid platforms are finally approaching a cost‑performance curve where broad deployment looks plausible, but there is no clear default software layer for social, people‑facing behavior. Lightberry is explicitly going after that horizontal “brain” layer, with hardware‑agnostic integrations, emotionally intelligent interaction, and voice‑driven configuration that map well to the needs of facilities, event operators, and consumer environments that lack in‑house robotics talent.

If humanoids follow PCs and smartphones, the dominant value may accrue not to any single hardware form factor but to the software stack that makes them safe, personable, and commercially useful in the real world.

Lightberry’s mission: robots with soul, deployed at scale into human‑dense environments positions it as a compelling candidate to own that layer in a category that is likely to compound over the coming decade.

#BackedBySingularityCapital #portfolio #portco #investment #vc #venturecapital