Partner at Union Square Ventures (USV)
Check size: USV states that the firm writes $1 million to $30 million checks and invests from the earliest stages through the full company journey. This is a firm-wide range, not Albert's personal check authority; decisions are discussed collectively without a formal vote.
Albert is a thesis-driven generalist investing at the edge of large markets under technological and societal pressure. His current work emphasizes AI-enabled scientific discovery, robotics that can operate reliably in complex physical environments, resilient energy and civil infrastructure, and products that expand human agency. Across these areas he looks for compounding data or participation loops and systems that turn rapidly improving technical capabilities into durable real-world utility.
Expect a discussion-heavy full-partnership meeting. Frame the transformative pressure on a large market, the technical insight, and the compounding loop. For robotics, bring reliability, speed, deployment, safety, and customer evidence; for scientific AI, show the prediction-experiment-measurement loop and proprietary data advantage; for infrastructure, quantify the real deployment bottleneck and civilizational-scale upside.
Civilization-scale problems, strong technical founders, commercially deployed systems, measurable reliability, AI that discovers structure in high-dimensional domains, proprietary data flywheels generated by real experiments, robots moving beyond controlled factories, and products that expand knowledge, access, and human agency.
No personal pass list is published. His current writing implies caution toward robotics demos without reliability at speed, scientific AI without an experiment-to-data feedback loop, infrastructure products lacking evidence of deployment, and incremental automation that does not unlock a materially broader human or market capability. These are inferred filters.
For physical AI, measure success rate at commercially required cycle time under real operating variation; treat impressive capabilities without reliable throughput as pre-deployment evidence.
Trace how models propose candidates, physical experiments generate previously unavailable measurements, and proprietary results improve the next prediction round.
Connect a narrow first deployment to the broader construction or maintenance burden, proving present customer economics while preserving a path to a much larger infrastructure platform.
Evaluate whether automation merely lowers labor cost or meaningfully frees people from dangerous, repetitive, or backbreaking work for higher-value human activity.
Abundant intelligence makes products faster to build but raises the importance of resilient product experience, positioning, narrative, and uncertain new moats.
Robotics can address the civilizational challenge of building and maintaining infrastructure as machines become capable in complex outdoor environments, with grid-scale solar providing a concrete commercial wedge.
AI can search vast materials spaces, while synthesis and measurement create a proprietary experimental-data loop that compounds model quality and enables rare-earth-free magnets.
Robotics foundation models shorten task-specific training, but commercial readiness is determined by reliability at speed rather than model capability alone.
+ 207 more investments. View fund →