Recommended Articles
Fundraising

Drylab AI Raises $3M Seed to Build Autonomous Research Agents for Biopharma

By Angeled · October 13, 2026 · Fundraising
Drylab AI Raises $3M Seed to Build Autonomous Research Agents for Biopharma
Drylab AI Raises $3M Seed to Build Autonomous Research Agents for Biopharma

Singapore-based Drylab AI has raised a $3 million seed round led by Vertex Ventures, with participation from PatSnap, the patent-intelligence unicorn. The funding, announced October 12, 2026, will be used to grow Drylab's engineering and biology teams, hire sales staff, and pursue the security and compliance certifications required to serve large research organizations.

Drylab is building AI agents that conduct scientific research independently of human supervision β€” defining research strategies, selecting from more than 800 scientific tools and databases, interpreting results, and troubleshooting experimental failures on their own.

Who Is Behind Drylab AI

The company was founded by CEO John Phan, CTO Huy Phan, and Chief Science Officer Nhi Nguyen β€” a trio that pairs engineering leadership with scientific depth, a combination investors increasingly demand in AI-for-science startups. The seed round structure also tells a story: Vertex Ventures is a global VC with strong life-sciences credentials, while PatSnap's participation as a strategic investor is notable. PatSnap operates one of the world's largest patent and R&D intelligence platforms, meaning Drylab gains not just capital but a potential data and distribution partner deeply embedded in the biopharma research workflow.

What "Autonomous Research Agents" Actually Means

To understand Drylab's ambition, it helps to distinguish three tiers of AI in scientific research. The first tier is search and summarization β€” tools that retrieve papers or answer questions, essentially research copilots. The second tier is workflow automation β€” systems that execute multi-step pipelines a human has designed. Drylab is aiming for the third tier: autonomous agents that take a research goal, decompose it into a strategy, pick the right tools (literature databases, chemical libraries, assay protocols, patent filings), run analyses, read the results, and decide what to try next when something fails.

The claim of access to 800+ scientific tools and databases is the company's core technical bet. Breadth of tooling matters because scientific research is fractured across hundreds of specialized databases β€” genomic repositories, compound libraries, patent databases, clinical trial registries. An agent that can only query one of them is a demo; an agent that can orchestrate across many of them starts to look like a digital research scientist.

The AI-for-Science Funding Wave

Drylab is riding one of venture capital's hottest 2025–2026 themes: AI systems that compress the scientific method itself. R&D productivity in pharma has been stagnant for decades β€” the "Eroom's Law" observation that drug discovery gets more expensive even as technology improves. Anything that plausibly accelerates the design–test–analyze loop commands investor attention, because a single blockbuster drug is worth billions.

But the space is crowded and technically unforgiving. Scientific agents must be reliable in a domain where a hallucinated citation or a misread assay can waste weeks of lab time and millions of dollars. The certification and compliance investments Drylab flagged in its use-of-funds plan are a direct response to this: enterprise biopharma buyers will not hand autonomous agents to regulated research pipelines without audited security, data handling, and traceability.

What Investors Should Watch

Three questions will determine whether Drylab's seed becomes a breakout Series A: benchmarking β€” can it demonstrate measured improvements on real research tasks versus human or copilot baselines; enterprise adoption β€” will large pharma and research institutions trust autonomous agents inside regulated workflows; and defensibility β€” whether its agent architecture and tool integrations constitute a moat, or whether foundation-model providers will absorb this capability.

The PatSnap relationship is the wildcard. Strategic investors from the patent and R&D intelligence world can provide proprietary data access and credibility with enterprise research teams β€” exactly the two hardest things for a seed-stage AI startup to build from scratch.

Why It Matters for Founders and Investors

For founders, Drylab validates a fundraising pattern: in AI-for-science, a deep founding team plus a strategic investor from the target industry's data layer can carry a seed round even pre-revenue. For investors, the round is another data point that capital is flowing toward agentic AI rather than chatbots β€” systems that do work, not just answer questions. The biopharma vertical remains one of the few where AI startups can command real enterprise budgets rather than experimentation dollars.

Angeled.com connects founders with investors.

Read Next