A Major Bet on AI-Powered Drug Discovery

The race to use artificial intelligence to accelerate pharmaceutical drug discovery has attracted enormous capital from investors who believe that the traditional drug development process — which takes an average of 12 years and $2.6 billion to bring a single drug from discovery to approval — is ripe for compression through intelligent automation. Into this landscape steps Earendil Labs, a China-rooted AI biotech that has just closed a $787 million pre-IPO funding round, one of the largest single raises in the AI drug discovery space and a validation of investor appetite for companies that are making credible progress toward AI-designed therapeutics.

The funding round, which values Earendil at a figure that places it among the most highly capitalized private biotechs in the world, comes as the company prepares for what sources describe as a possible initial public offering in the United States or Hong Kong, or potentially both. The specific timing and venue have not been confirmed, but the pre-IPO label on the round suggests that the company and its investors believe public market conditions in the next 12 to 18 months are favorable for a listing.

What Earendil Does

Earendil Labs was founded by researchers with backgrounds at leading Chinese academic institutions and pharmaceutical companies, with early development work conducted primarily in China and subsequent global expansion into the United States and Europe. The company's core platform uses generative AI models trained on protein structure, molecular interaction, and clinical outcome data to identify and optimize drug candidates in target classes that have historically been difficult for conventional drug discovery approaches.

The company has been particularly active in the development of treatments for cancers and neurodegenerative diseases — two of the largest unmet need areas in medicine and two categories where the complexity of the underlying biology has defeated many conventional drug discovery programs. Earendil's AI platform, according to the company, has generated several clinical-stage candidates in these areas that are progressing through Phase 1 and Phase 2 trials, representing a pipeline that investors are betting will produce commercially significant approvals.

Geopolitical Complexity

The China-rooted label on Earendil is not incidental. The company was founded and initially developed in China, and a significant portion of its research operations, data assets, and leadership team are based there. This creates a set of geopolitical complications that have become increasingly relevant for biotech investment as US-China relations have deteriorated and as regulatory and legislative scrutiny of Chinese-connected companies in sensitive sectors has intensified.

The BIOSECURE Act, which has been under congressional consideration in various forms, would restrict US federal agencies and certain US-connected entities from doing business with Chinese biotechs deemed to pose national security concerns. While Earendil has not been specifically named in these discussions, the category of AI-enabled drug discovery companies with Chinese origins is one that lawmakers and regulators are watching carefully. The company's pre-IPO investors and its potential IPO underwriters will need to navigate these regulatory currents carefully if they are pursuing a US listing.

Earendil has reportedly taken steps to establish an organizational structure that separates its US and Chinese operations, creating a corporate firewall designed to address regulatory concerns about data access and technology transfer. Whether this structure is sufficient to satisfy US regulators and potential institutional investors who are subject to restrictions on China-connected investments remains to be determined.

The Broader AI Drug Discovery Landscape

Earendil competes in a rapidly expanding sector that includes Isomorphic Labs (Google DeepMind's drug discovery spinout), Recursion Pharmaceuticals, Insilico Medicine, Exscientia, and a growing number of well-capitalized startups. The sector received a significant credibility boost from the success of AlphaFold, DeepMind's protein structure prediction system, which solved a problem that had challenged structural biologists for decades and which has now been used in the early stages of drug discovery programs at essentially every major pharmaceutical company.

The current generation of AI drug discovery platforms is attempting to go beyond structure prediction to address the harder problem of predicting how a molecule will interact with a target in a way that produces the desired therapeutic effect without intolerable side effects — essentially, predicting clinical success rather than just molecular structure. This is a substantially harder problem, and the field has not yet produced a landmark AI-discovered drug that has achieved regulatory approval and commercial success at scale. Earendil and its peers are betting — and $787 million in investors are also betting — that they will be among the first to reach that milestone.

This article is based on reporting by endpoints.news. Read the original article.