---
title: "Beyang's US$30M Series A shows AI drug discovery is still raising while apps stall"
date: 2026-11-13
category: Capital & Markets
site: NeuroAI
canonical: https://neuroai.site/a/na-capital-beyang-ai-drug-series-a
language: en
---

# Beyang's US$30M Series A shows AI drug discovery is still raising while apps stall

> China-based Beyang Therapeutics closed a nearly US$30M oversubscribed Series A, co-led by Legend Capital and Shanghai Healthcare Capital, to advance its ExCEED AI drug-design platform.

Most AI funding headlines in 2026 are about giants raising billions for data centers and frontier models. A smaller, quieter round in August suggests where patient capital is quietly moving instead. A Chinese biotech using AI to design drug molecules just closed a financing that says more about the sector's real priorities than the headline totals do.

## The round at a glance

- **Nearly US$30 million**, in an oversubscribed **Series A**, closed **12 August 2026**.

- **Co-led by Legend Capital and Shanghai Healthcare Capital**.

- Joined by China Medical System Holdings (CMS), Fenglei Capital, Shanghai Sci-Tech Innovation Center Capital and Harbor & Canton Capital; existing backers Root Venture Partners and YuanBio Venture Capital re-upped.

- Use of proceeds: iterate the ExCEED platform, push pipeline candidates through preclinical and clinical work, and build out international R&D and business-development teams.

Beyang describes its method as a dry-wet closed loop (干湿闭环): model-generated molecule proposals are tested in the lab, and the experimental results feed back into the next model iteration.

The size is modest next to the mega-rounds dominating the news cycle. The signal is not the size — it is who showed up, and what they are funding.

## What Beyang actually does

Beyang Therapeutics is a China-based AI-driven drug R&D company founded in 2021. Its engine, **ExCEED**, fuses proprietary databases with AI and computational simulation across three steps: molecular design, druggability optimization, and biological evaluation. The goal is to generate genuinely novel molecular scaffolds rather than tweak existing drug cores. The team behind it is not a fresh spinout: its scientists include alumni of Pfizer, Merck and Hengrui who, between them, have advanced more than ten projects into clinical trials and contributed to three globally launched small molecules.

The platform recently ranked **8th globally among 350-plus teams** in the OpenADMET blind prediction challenge — an independent test of how well models predict drug-behavior properties. That kind of third-party benchmark is rarer and more useful than a founder's slide deck, because the models were scored on data the teams had never seen.

The company frames its work as a "dry-wet closed-loop" (干湿闭环) approach: computation proposes molecules, the lab tests them, and the results feed back into the models. The loop is the point — it is what separates a prediction tool from a discovery engine.

## The pipeline it is funding

Two candidates anchor the raise:

- **BT01001**: a small-molecule eye drop for retinal diseases. In preclinical head-to-head work it showed best-in-class ocular exposure with efficacy comparable to injected Aflibercept. It has completed Phase I single- and multiple-ascending-dose studies and is moving toward first-in-patient trials.

- **BT01002**: a second-generation Menin inhibitor for acute leukemia tied to KMT2Ar or NPM1 mutations, designed to dodge the resistance that limits first-generation inhibitors. It is in IND-enabling studies.

Neither is a sure thing — drug development is inherently binary — but both are concrete, mechanistically defined programs rather than vague platform promises. The Menin target is a useful example of why computation-guided design earns attention: resistance mutations have blunted earlier inhibitors, and a molecule built from a novel scaffold to sidestep those mutations is exactly the kind of differentiated asset that licensing deals and later rounds reward. Beyang's pitch is that ExCEED can produce more of these per unit of lab spend than a conventional medicinal-chemistry team.

## Why this round matters beyond its size

AI Drug Discovery (AIDD) has cycled through hype before. What is different now, investors argue, is vertical integration: platforms that cover discovery-to-optimization under one roof, validated by lab data. Legend Capital's Fei Qi framed it as an "AIDD 2.0 era" defined by integration rather than isolated prediction tools — a promotional statement from the funding release, but one that captures the current investor thesis.

The contrast with the broader venture market is the real story. While application-layer AI startups struggle to raise, computationally grounded drug-discovery platforms are still attracting syndicates that mix biotech specialists with healthcare strategics. That divergence is exactly where disciplined capital is rotating.

## The funding climate behind the round

Beyang did not raise in a vacuum. Industry observers noted that through the second half of 2026, larger biopharma financings increasingly went to companies with platform capabilities and authentic pipelines rather than model claims alone. The pattern is visible across the AI-drug segment: rounds are narrowing to teams that can show experimental validation, not just architectural ambition. Beyang's oversubscription and its mix of specialist and strategic investors fit that script precisely.

That is also why the round's modest size matters less than its composition. A US$30M Series A co-led by a mainstream VC (Legend Capital) and a healthcare-focused investor (Shanghai Healthcare Capital), with a listed drug distributor (CMS) among the participants, is a vote of confidence in the commercial, not just the computational, path. For a sector that has burned through hype cycles, the signal is the syndicate.

## Honest limitations

- The US$30M figure is the company's and multiple outlets' stated round size ("nearly $30M"); the exact final close and post-money valuation were not disclosed.

- Clinical-stage claims (Phase I completion, IND-enabling status) come from the company's announcement, not an independent regulator; drug development carries inherent failure risk.

- The OpenADMET ranking is a single blind-challenge snapshot, not a guarantee of pipeline success.

- Investor quotes are promotional statements from the funding release and should be read as such.

## What readers can do now

- For healthcare investors, screen AIDD names by *validated* benchmarks — blind challenges, published data — not just model claims.

- Watch for strategic-pharma participation in rounds; it often signals pipeline credibility beyond the cash.

- Treat preclinical promise as optionality, not value: most candidates never reach market, so size AI-biotech bets as high-risk.

---

Published by NeuroAI (https://neuroai.site/) — https://neuroai.site/a/na-capital-beyang-ai-drug-series-a
Free to quote with attribution and a link to the original.
