Skip to main content

涌生智能

Category: AI in Biotech / Drug Discovery

An AI-driven self-evolving laboratory company focused on AI for Science, building dry-wet experiment closed-loop infrastructure for life sciences, and a subsidiary of MGI Tech (华大智造). 涌生智能 was founded in 2026. The company is led by 杨梦 (Yang Meng). Based in Shenzhen, China. Latest round: Undisclosed. Key investors include 华大智造 (MGI Tech) - parent company, holds 87.5% equity.

Founded
2026
Headquarters
Shenzhen, China

Value proposition

Building the world's first complete dry-wet lab closed-loop for life sciences — enabling AI to not just read papers but actually perform real experiments in physical laboratories, bridging the gap between AI-generated protocols and wet lab execution.

Products and solutions

ProtoPilot (self-evolving multi-agent system for wet lab automation), BioLab Bench (first full-process Agent evaluation system for life sciences from user intent to device execution), αLab Brain (AI lab assistant/companion), AlphaTool (intelligent liquid handling platform), PrepALL (automated sample preparation), AIO (all-in-one intelligent lab device), E25 Flash (AI-integrated flash sequencer)

Unique value

Only company that has achieved full-chain Physical AI in life sciences — from natural language experiment intent to wet lab physical execution, with real experimental feedback and self-evolution. ProtoPilot scored 52.38% on ProtocolQA (vs GPT-5.6 Sol's 43.5%) and 85.18% on non-open QA (exceeding human expert level).

Target customer

Life science research institutions, pharmaceutical companies, clinical diagnostic labs, synthetic biology companies, academic research labs, and biotech enterprises needing automated lab workflows

Industries served

Life Sciences, Biotechnology, Drug Discovery, Clinical Diagnostics, Synthetic Biology, Genomics

Technology advantage

Physical AI advantage — built from real lab equipment and wet lab experience (not just model training). Multi-agent orchestration (Orchestrator + Protocol Expert + Coding Agent). Protocol2Code translation across 4+ major lab automation platforms (MGI AlphaTool, Hamilton STAR, OpenTrons OT-2, Tecan EVO) with 96.6% gate pass rate. Self-evolving through real experimental feedback. Backed by MGI Tech's 3800+ global users and decade of life science equipment engineering.

How they differentiate

Unlike Silicon Valley AI companies (OpenAI, Anthropic, Google) that approach life sciences from the model side (scale compute), 涌生智能 approaches from the equipment/lab side — growing AI from inside the laboratory. This gives them access to real physical constraints, real experimental failures, and real wet lab feedback that pure AI companies cannot replicate. ProtoPilot outperformed GPT-5.6 Sol on ProtocolQA benchmark (52.38% vs 43.5%).

Main competitors

百图生科 (BioMap) - AI life science foundation model company, 晶泰科技 (XtalPi) - AI-powered lab automation and drug discovery, OpenAI GPT-Rosalind / Anthropic Claude Science - frontier AI models applied to life sciences

Key partnerships

上海人工智能实验室 (Shanghai AI Laboratory) - co-developed ProtoPilot and BioLab Bench, 华大智造 (MGI Tech) - parent company providing equipment, distribution, and user base

Major milestones

2026-03-19: Company officially registered (工商注册日), 2026-07-01: Joint release of ProtoPilot and BioLab Bench with Shanghai AI Laboratory, ProtoPilot outperformed OpenAI GPT-5.6 Sol on ProtocolQA benchmark, Published EvoPlay and PrimeGen in Nature sub-journals (predecessor work by team), Developed E25 Flash AI-integrated sequencer

Market positioning

First-mover in Physical AI for life sciences — positioned as the bridge between AI models and real laboratory execution. Competes against both AI giants (OpenAI, Anthropic) entering bio and traditional lab automation companies, with the unique advantage of owning both the equipment layer and the AI layer.

Geographic focus

China (primary), with global ambitions through MGI Tech's international presence

About 杨梦 (Yang Meng)

华大智造 (MGI Tech) Chief AI Officer & Senior VP; Led X-team, self-luminescent sequencer R&D, and GLI (Generative Lab Intelligence) business; Published in Nature sub-journals (EvoPlay, PrimeGen); PhD supervised by Yuan Yingjin (academician in synthetic biology); Cross-disciplinary background in biology, genomics, AI, and automation.

Official website: