Latent-Y: AI Agent Designs New Drugs Autonomously from Text

LONDON & SAN FRANCISCO – March 23, 2026 – The race to discover new medicines has been dramatically accelerated with the launch of Latent-Y, an autonomous AI agent developed by Latent Labs. The company announced today that its new system can design novel therapeutic antibodies from a simple text prompt, compressing a process that typically takes expert teams weeks into just a few hours.

This new tool operates as a force multiplier for drug discovery, allowing a single researcher to run multiple design campaigns simultaneously. Powered by Latent-X2, the company's frontier generative model, Latent-Y promises to bring sophisticated structural drug design to any researcher, without the need for specialist infrastructure. The company is now opening access to the platform for selected partners.

A Quantum Leap in Discovery Speed

At the heart of Latent-Y's disruptive potential is its radical efficiency. In user studies, PhD-level experts working with the AI agent completed antibody design campaigns 56 times faster than independent estimates for manual workflows. This isn't merely an incremental improvement; it represents a fundamental shift in the pace of preclinical research.

The process begins with a simple input: a therapeutic goal specified in natural language, a research plan, or even a scientific publication. From there, the agent takes over. It analyzes target molecules, applies biological reasoning to identify viable binding sites (epitopes), designs a fleet of antibody candidates using the Latent-X2 engine, and computationally validates them. The system iterates on its designs until the specified goals are met, recording every decision and its underlying reasoning for scientific review.

To prove its capabilities, Latent Labs presented lab-validated results from three distinct, fully autonomous campaigns:
* Epitope Discovery: The agent successfully identified therapeutically relevant epitopes and designed VHH binders (nanobodies) with single-digit nanomolar affinities, a high-quality result for early-stage discovery.
* Cross-Species Design: Latent-Y generated antibodies that bind to homologous targets across different species, a critical step for ensuring preclinical models translate effectively to human trials. Notably, the AI autonomously extended its own capabilities by implementing a new generative method to solve this challenge.
* Design from Publication: Fed a peer-reviewed scientific paper, the agent autonomously designed antibodies targeting the human transferrin receptor (hTFR1), a key protein for delivering drugs across the blood-brain barrier.

"Latent-X2 gave us the breakthrough: antibodies designed computationally with drug-like developability," said Simon Kohl, CEO and founder of Latent Labs, in a statement. "Latent-Y builds on that foundation with an expert reasoning layer that handles the full workflow autonomously. The result is speed and scale that weren't possible before — a single researcher running dozens of campaigns in parallel."

Reshaping a Competitive Landscape

Latent Labs, which emerged from stealth in early 2025 with $50 million in funding, enters a fiercely competitive AI drug discovery market. The company, founded by Kohl, a former researcher on Google's groundbreaking AlphaFold team, joins a host of well-funded players like Generate Biomedicines, AbSci, and Insilico Medicine, all leveraging AI to reinvent therapeutics. Even Google's own Isomorphic Labs is a formidable competitor, building on DeepMind's protein-folding breakthroughs.

However, Latent Labs' unique selling proposition is the pronounced autonomy of its agent. While competitors offer powerful AI-powered design platforms, Latent-Y is positioned as a system that handles the entire workflow from start to finish. This "push-button" approach, moving from a text prompt to a list of lab-ready candidate sequences, aims to minimize the manual bioinformatics work that often surrounds AI design tools. The company’s focus on generating molecules with "drug-like developability" from the outset also aims to reduce the high rate of downstream failures that plague the industry.

The company's substantial backing, which includes investments from Google Chief Scientist Jeff Dean and Transformer architecture co-inventor Aidan Gomez, signals strong confidence from leaders in the AI field. This combination of top-tier talent and a differentiated product strategy gives Latent Labs a credible position in the race to digitize biology.

Democratizing the Fight for New Cures

Beyond sheer speed, Latent-Y's most significant long-term impact may be its potential to democratize drug discovery. Traditionally, designing novel antibodies has required large, specialized teams and expensive computational infrastructure, creating a high barrier to entry for smaller organizations.

By offering its powerful tools through a web browser or API, Latent Labs makes advanced structural design accessible to academic labs, small biotech startups, and researchers in resource-limited settings. This could level the playing field, allowing a much broader community of scientists to contribute to the development of new medicines. The ability to pursue more targets, test riskier but potentially more rewarding hypotheses, and explore a wider range of design strategies could unlock innovation across the entire biomedical ecosystem.

The current rollout is limited to "selected partners," a common strategy for platform companies to refine their product and build case studies with established industry players. However, the company's stated mission to transform drug discovery into "automated drug design" suggests a future where these tools become widely available, fundamentally changing who can create new drugs and how they do it.

The Human-AI Partnership and Its Hurdles

Despite its autonomy, Latent-Y is designed to work in partnership with scientists, not replace them. The system can be configured to pause at each stage, surfacing progress summaries and recommended next steps for expert review. Every design choice is logged with its rationale, creating an auditable trail that scientists can evaluate, challenge, and build upon. This "glass box" approach is a crucial feature for building trust and ensuring scientific rigor.

This human-in-the-loop model is also critical for navigating the complex ethical and regulatory landscape of AI-driven medicine. As AI agents take on more responsibility, questions of accountability become paramount. If an AI-designed drug causes unforeseen harm, where does the responsibility lie? Regulatory bodies like the U.S. Food and Drug Administration (FDA) and the European Medicines Agency (EMA) are actively developing frameworks to address these issues, focusing on algorithm transparency, data quality, and robust validation.

Latent Labs' emphasis on lab-validated results and auditable reasoning directly addresses these emerging regulatory demands. Even with these safeguards, experts caution that AI-designed antibodies are still a long way from the clinic. The path to an approved drug is fraught with challenges that extend far beyond initial design, including manufacturing, formulation, extensive safety trials, and clinical efficacy. While AI like Latent-Y can dramatically accelerate the first step of a marathon journey, the road to the finish line remains long and requires deep human expertise.