Vertex Backs Drylab AI’s $3 Mn Push to Take Agents Deeper into Scientific Research

The Singapore startup is betting that the next valuable AI agent will not simply assist scientists but will execute larger parts of computational research while keeping sensitive data inside an organisation’s infrastructure.

By Entrepreneur Staff | Oct 14, 2026
drylab

You're reading Entrepreneur Asia Pacific, an international franchise of Entrepreneur Media.

Singapore-based Drylab AI has raised $3 Mn in seed funding led by Vertex Ventures Southeast Asia & India, with participation from innovation intelligence company PatSnap, as it seeks to turn autonomous research agents into enterprise software for pharmaceutical companies and research institutions. The company announced the round on October 13.

The contest to commercialise AI agents is moving from office workflows into laboratories, where autonomy comes with a much higher burden of proof.

Drylab’s proposition goes beyond the familiar AI copilot. The company says its agents can take a research objective, formulate an analytical strategy, select tools, run computational experiments, interpret results and troubleshoot failures, rather than requiring a scientist to prompt the system at each step. Its platform brings together more than 800 integrated scientific tools and databases, according to the company.

That makes the funding round a bet on a more consequential stage of agentic AI: whether software can take responsibility for chunks of specialised work.

Earning Trust in the Lab

The distinction matters most in biopharma, where automation can deliver large value, but an unreliable result can be costly. Scientific workflows depend on reproducibility, methodological choices and proprietary research data, raising the bar for an AI system beyond many consumer or workplace applications.

Drylab has designed part of its pitch around that constraint. Its platform can run within a customer’s infrastructure or on local workstations, keeping data, workflows and intellectual property within the organisation’s security perimeter. It can also be deployed in isolated cloud environments, and customers keep ownership of the data and discoveries generated.

For pharmaceutical and research organisations, that architecture addresses a practical barrier to adopting generative AI. Sensitive scientific information doesn’t have to be transferred to an external platform for the software to work.

The company has also built in checks. A proprietary verification model, trained on real scientific analysis runs, flags methodological errors so that agents can correct their work, according to Vertex. Scientists can review critical decisions, such as data thresholds and computational methods, at set checkpoints.

Drylab has early evidence to support this. In the 2026 OpenADMET PXR challenge, an open drug-discovery modelling competition, the company said its system ran autonomously for seven days after the first-phase data was released. It ranked ninth on the Phase 1 public leaderboard and 23rd on the final private activity leaderboard. The competition drew roughly 400 submissions in the first round and 100 in the second.

The Enterprise Hurdle

Commercial traction will now determine whether the technical proposition translates into a scalable business.

Drylab commercially launched in February 2026 and has since signed paying customers among pharmaceutical companies, cancer centres, biomedical research institutes and universities across the US, Europe and Asia, according to the company and Vertex. Customer names and revenue figures were not disclosed.

The new capital will fund expanded engineering and biology teams, additional enterprise sales staff and the security and compliance certifications needed to sell to larger research organisations.

That spending plan points to the commercial challenge. Building increasingly capable agents is only one part of selling AI into life sciences. Startups also have to satisfy enterprise requirements around deployment, governance, security and integration before experimental technology can become part of everyday research infrastructure.

PatSnap’s participation adds another dimension to the round. Founded in Singapore, PatSnap runs an AI-powered platform spanning patents, research and life sciences, placing a strategic investor with an existing research-data ecosystem alongside venture investor Vertex.

A Singapore Base and a Vietnamese Team

Drylab, founded in 2025, operates across Singapore and San Francisco. Its Vietnamese founding team is led by co-founder and chief executive John Phan, who previously did AI research with ByteDance’s Seed team.

Chief technology officer Huy Phan was a visiting research engineer at Andrew Ng’s Landing AI. Chief science officer Nhi Nguyen is a PhD candidate in biomedical science at Singapore’s Nanyang Technological University and taught herself bioinformatics to analyse her own lab’s data.

Genping Liu, a general partner at Vertex Ventures Southeast Asia & India, described the startup as “global ambition coming out of Vietnam’s world-class tech talent pool”. Its emergence reflects a broader shift in AI entrepreneurship from horizontal assistants towards narrower systems built around specialised professional workflows.

For Drylab, the commercial opportunity is not simply to make scientists faster. The bigger bet is that research organisations will let AI agents execute more of the computational process, provided they can retain control of the data and the work remains inspectable.

The $3 Mn seed round gives the company more resources to test that proposition. Whether autonomous research agents become embedded infrastructure will depend less on how independently they can act in a demonstration than on whether scientists and regulated organisations can trust the work they produce repeatedly.

Singapore-based Drylab AI has raised $3 Mn in seed funding led by Vertex Ventures Southeast Asia & India, with participation from innovation intelligence company PatSnap, as it seeks to turn autonomous research agents into enterprise software for pharmaceutical companies and research institutions. The company announced the round on October 13.

The contest to commercialise AI agents is moving from office workflows into laboratories, where autonomy comes with a much higher burden of proof.

Drylab’s proposition goes beyond the familiar AI copilot. The company says its agents can take a research objective, formulate an analytical strategy, select tools, run computational experiments, interpret results and troubleshoot failures, rather than requiring a scientist to prompt the system at each step. Its platform brings together more than 800 integrated scientific tools and databases, according to the company.

Entrepreneur Staff • Editor

Entrepreneur Staff
For more than 30 years, Entrepreneur has set the course for success for millions of... Read more

Related Content