AI-Powered Drug Discovery Startup Raises $200M to Accelerate Pipeline
London-based biotech startup EnzymeAI has secured $200 million in a Series B funding round led by Andreessen Horowitz, with participation from existing investors Flagship Pioneering and SoftBank Vision Fund. The financing values the company at $1.6 billion, according to sources familiar with the deal. The funds will be used to advance EnzymeAI’s lead candidate, EAI-309, into Phase II clinical trials and expand its proprietary enzyme-targeting platform across oncology and rare metabolic disorders.
Founded in 2020 by a team of computational biologists from Oxford and Imperial College London, EnzymeAI combines deep learning models trained on 3D protein structures with high-throughput robotic screening to identify novel enzyme modulators. Its platform analyzes over 500,000 structural variants per week, prioritizing candidates with optimal binding affinity and minimal off-target effects. The company currently has six programs in preclinical development, three in partnership with major pharmaceutical firms: AstraZeneca, Merck KGaA, and Takeda.
EAI-309, the most advanced asset, targets a mutated form of isocitrate dehydrogenase (IDH1) present in approximately 20% of gliomas and 10% of acute myeloid leukemia (AML) cases. In Phase I trials involving 48 patients, the drug demonstrated a 37% partial response rate and manageable side effects, primarily fatigue and mild liver enzyme elevation. The upcoming Phase II study will focus on newly diagnosed AML patients and is expected to begin enrolling in Q2 2025, with results anticipated by late 2026.
How EnzymeAI’s Platform Differs from Traditional Approaches
Traditional drug discovery relies on screening millions of chemical compounds in vitro, a process that can take years and cost over $2 billion per approved drug, according to a 2023 JAMA study. EnzymeAI cuts that timeline by using generative AI to simulate molecular interactions before physical testing. Its neural networks are trained on public databases like the Protein Data Bank and proprietary datasets from collaborations with academic labs at Cambridge and the Francis Crick Institute.
The company’s AI system predicts how small molecules will bind to enzyme active sites with greater accuracy than earlier machine learning models. It factors in dynamic behaviors such as protein folding shifts and allosteric regulation—features often missed by static structural analysis. This enables EnzymeAI to design compounds that modulate enzyme function without fully inhibiting it, an approach that may reduce toxicity.
For example, in early testing, the platform identified a non-competitive inhibitor of mutant IDH1 that reduced production of the oncometabolite D-2-hydroxyglutarate by 70% in cell cultures, compared to 50% reduction seen with legacy inhibitors like ivosidenib. The finding, published in Nature Chemical Biology in 2023, suggests a potential mechanism for deeper disease suppression.
Competitive Landscape in AI-Driven Drug Discovery
EnzymeAI operates in a crowded field of AI biotechs racing to validate their platforms. Recursion Pharmaceuticals, based in Utah, went public via SPAC in 2022 and has over 40 programs in its pipeline, including partnerships with Roche and Bayer. Its AI platform uses high-content cellular imaging to map disease phenotypes and drug responses, differing from EnzymeAI’s structure-based approach. Recursion’s lead oncology candidate, REC-994, entered Phase II trials in 2024 for glioblastoma.
In the UK, Isar Imaging and Exscientia also compete in AI-driven discovery. Exscientia’s collaboration with Bristol Myers Squibb yielded DSP-1181, an AI-designed molecule for obsessive-compulsive disorder that entered clinical testing in 2021. The drug was later discontinued due to lack of efficacy, underscoring the technical risks in the field. Despite setbacks, Exscientia has maintained partnerships with Sanofi and Sumitomo Dainippon Pharma, with total collaboration funding exceeding $5 billion since 2019.
Beyond startups, large pharma companies are investing heavily in internal AI capabilities. Pfizer launched an AI research center in Boston in 2022, allocating $100 million over five years. Its platform, dubbed “PfizerWorks,” integrates real-world patient data with molecular modeling to prioritize clinical trial designs. Novartis has partnered with Microsoft to use Azure-based machine learning for target identification, while GSK has built an in-house AI team that now includes over 150 data scientists.
Despite growing competition, EnzymeAI’s focus on enzyme-specific modulation gives it a niche advantage. Enzymes represent about 30% of known drug targets, according to the IUPHAR/BPS Guide to Pharmacology, and are involved in nearly all metabolic pathways. Success in this domain could open access to diseases with limited treatment options, including urea cycle disorders and lysosomal storage diseases.
Regulatory and Ethical Dimensions of AI in Drug Development
As AI becomes more integrated into drug discovery, regulatory agencies are adapting their frameworks. The U.S. FDA issued draft guidance in early 2024 on the use of artificial intelligence in clinical investigations, emphasizing transparency in algorithm design and validation. The agency requires sponsors to document training data sources, model performance metrics, and steps taken to mitigate bias—especially in datasets that underrepresent certain populations.
EnzymeAI has engaged with the FDA’s Emerging Technology Program since 2023 to align its development process with evolving expectations. The company maintains detailed logs of every model iteration and conducts external audits to verify reproducibility. It also uses synthetic data augmentation to improve representation of genetic variants more common in non-European populations, addressing a known limitation in genomic databases.
Outside the U.S. the European Medicines Agency (EMA) has established an AI Task Force to evaluate regulatory needs. The UK’s Medicines and Healthcare products Regulatory Agency (MHRA) published its “AI in Regulated Research” framework in 2023, calling for standardized validation protocols. These efforts aim to prevent scenarios where AI-generated compounds fail in late stages due to unanticipated biological interactions.
Ethical concerns also persist. Some bioethicists warn that overreliance on predictive models may reduce emphasis on mechanistic understanding, potentially leading to drugs with unclear modes of action. Others caution that intellectual property rules may struggle to keep pace with AI-generated inventions. In 2022, the U.S. Patent and Trademark Office ruled that AI systems cannot be listed as inventors, but questions remain about who owns discoveries made by AI trained on public data.
The Bigger Picture: Why AI Drug Discovery Matters Now
The pharmaceutical industry faces mounting pressure to deliver new treatments faster and at lower cost. R&D spending across the top 20 pharma companies exceeded $130 billion in 2023, yet the number of new molecular entities approved by the FDA has remained relatively flat—averaging around 35 per year over the past five years. Productivity, measured as approved drugs per billion dollars spent, has declined by over 75% since the 1950s, a trend known as Eroom’s Law (Moore’s Law spelled backward).
AI offers a way to reverse this trend. A 2023 analysis by McKinsey estimated that AI could shorten drug discovery timelines by 20–40% and reduce costs by up to 30% across preclinical and clinical phases. Successes like EnzymeAI’s progress with EAI-309 suggest these gains are becoming tangible. If Phase II trials confirm efficacy, the drug could reach market by 2028, potentially cutting five years off the typical development cycle.
The broader implications extend beyond individual companies. AI-driven platforms could democratize access to drug development, enabling smaller biotechs to compete with large pharma. Open-source tools like AlphaFold, developed by DeepMind, have already made high-accuracy protein structure prediction available to thousands of researchers worldwide. When combined with automated labs and cloud computing, these tools lower the barrier to entry.
Still, challenges remain. Many AI-generated compounds fail in later stages due to poor pharmacokinetics or unforeseen toxicity. No AI-designed drug has yet achieved blockbuster status. But as models improve and validation pipelines mature, the odds of success are rising. EnzymeAI’s latest funding round reflects investor confidence that the long-promised era of computational drug discovery may finally be arriving.
Next Steps for EnzymeAI
With the new capital, EnzymeAI plans to double its workforce to 180 employees, primarily hiring computational biologists and medicinal chemists. It will also expand its robotic screening facility in London’s White City Innovation District by 40%, adding three new high-throughput systems from Thermo Fisher Scientific. The company intends to file IND applications for two additional candidates by the end of 2025—one targeting phenylketonuria, the other a fibrosis-related kinase.
CEO Dr. Amara Singh said in a statement, “This round positions us to move beyond target discovery and into clinical validation at scale.” The company has not ruled out an IPO but expects to remain private through at least 2027, focusing on data generation rather than short-term market expectations.
Partnerships will remain central to EnzymeAI’s strategy. The collaboration with Takeda, announced in late 2023, includes $85 million in upfront and milestone payments for the development of enzyme modulators in neurodegenerative diseases. AstraZeneca’s involvement focuses on combining EAI-309 with existing immunotherapies in solid tumors, a combination that showed synergistic effects in mouse models.
If EAI-309 succeeds in Phase II, it would not only validate EnzymeAI’s platform but also strengthen the case for AI as a core component of modern drug development. For patients with cancers driven by metabolic enzyme mutations, it could mean access to more effective, less toxic therapies years ahead of traditional timelines.
