In 2025, AI is embedded across the drug development pipeline, from early-stage target discovery to clinical trial optimization and regulatory submissions, reshaping how therapies are discovered, developed, and deployed.
Unlocking biological complexity
One of AI’s most profound impacts is in decoding complex biological systems. AI-powered knowledge graphs link disparate data across genomics, proteomics, and clinical records, enabling connections that were previously impractical.
For example, researchers at the Oxford Drug Discovery Institute used AI to evaluate 54 immune-related genes as potential Alzheimer’s disease targets — a process that once took weeks, now completed in days.
Deep learning models are also advancing biomarker discovery by identifying molecular signatures that predict disease progression or therapeutic response. This is especially important in diseases like Alzheimer’s, cancer, and autoimmune disorders, where traditional models often fall short.
Designing drugs from scratch
Perhaps the most headline-grabbing application is de novo drug design. Generative models, such as variational autoencoders and generative adversarial networks (GANs), are creating novel molecules with optimized binding, pharmacokinetics, and toxicity profiles.
Advances like DeepMind’s AlphaFold 3 have improved protein structure predictions, critical for rational drug design.
Companies such as Insilico Medicine have already brought AI-designed drugs into the clinic. Its fibrosis candidate, ISM001-055, reached human trials in under 18 months — compared to the four years typical for traditional approaches. Other AI discovered drugs like DSP-0038 are advancing in trials, with early data suggesting higher-than-average success rates.
Smarter preclinical and clinical pipelines
In preclinical work, AI models are increasingly used to predict absorption, distribution, metabolism, excretion, toxicity, (ADMET) properties through in silico simulations. This approach not only speeds up the drug development process but also significantly reduces reliance on animal testing. In peptide development, platforms that combine AlphaFold with design tools like ProteinMPNN have dramatically accelerated the creation of stable andpotent therapeutics.
In clinical trials, AI is now improving patient selection by identifying individuals most likely to benefit from a therapy. This approach increases trial efficiency, reduces dropout rates, and improves the likelihood of demonstrating drug efficacy. Additionally, sophisticated simulation tools are being used to model various clinical scenarios to predict outcomes such as disease progression, treatment response, and adverse events. These simulations enable optimized trial designs by refining inclusion criteria, dosing regimens, and endpoint selection.
A 2024 industry analysis found that AI-assisted drug candidates achieved Phase I success rates of nearly 90 percent, compared with industry averages of 40–65 percent. These gains stem from integrating real-world data, predictive analytics, and rigorous in silico validation.
Rediscovering old drugs
AI is also breathing new life into shelved or off-patent drugs — an approach that can dramatically shorten development timelines and reduce costs.
By integrating molecular mechanism data, disease biology, and clinical outcomes into large-scale AI models, researchers can match existing compounds to new therapeutic indications that may have been overlooked in initial development.
Startups like Ignota Labs are using AI to mine public and proprietary datasets for such matches, aiming to reduce repurposing timelines to less than two years and costs to under $1 million.
Investment and regulatory momentum
The AI drug discovery sector drew $3.3 billion in venture funding in 2024 and shows no sign of slowing. Major deals include Generate:Biomedicines’ $1 billion partnership with Novartis and Isomorphic Labs’ $600+ million expansion to integrate AlphaFold into drug design.
This rapid growth and increasing industry adoption have prompted regulatory bodies to develop frameworks to ensure the safe and effective use of AI in drug development. Reflecting this trend, in 2025, the FDA released draft guidance on the use of AI to support regulatory decision-making for drug and biological products, introducing a risk-based framework for model credibility, emphasizing transparency, validation, and data governance.
Challenges ahead
AI’s potential comes with caveats. Models are only as reliable as the quality and diversity of their training data, and biases within these datasets can produce misleading or inaccurate results, especially in underrepresented populations or rare diseases. Intellectual property disputes, explainability, and data privacy concerns also remain unresolved.
And while AI accelerates workflows, human expertise is still central. As one Massachussets Institute of Technology researcher noted, AI is “a collaborator, not a replacement” — scientists still drive hypothesis generation, candidate evaluation, and strategic decisions.
The road to 2030
By the end of the decade, experts anticipate broader adoption of foundation models trained on multimodal biomedical data, autonomous AI lab agents, and continuous design-test cycles that dramatically accelerate discovery. These advances will require robust regulatory frameworks, better data infrastructure, and a new generation of interdisciplinary talent fluent in both biology and machine learning.
Early clinical successes like ISM001-055 offer proof of concept, and a glimpse into a shift from trial-and-error pharmacology toward rational, data-driven medicine. This technology is poised to fundamentally reshape how new medicines are discovered, developed, and delivered.
Frequently Asked Questions
(generated with AI assistance)
What is the current market size and projected growth for AI in drug discovery?
The global AI in drug discovery market was valued at $1.6 billion in 2023 and is projected to grow to $9.1 billion by 2032, reflecting a CAGR of about 21 percent.
What are the key applications of AI in drug discovery?
- Target discovery & validation: Integrating genomics, proteomics, and clinical data to uncover novel disease targets.
- De novo drug design: Using generative AI to create new molecules with optimized properties.
- Biomarker discovery: Applying deep learning to identify predictive molecular signatures.
- Clinical trial optimization: Improving patient selection, simulating outcomes, and integrating real-world data.
- Drug repurposing: Matching existing molecules to new therapeutic indications.
What notable breakthroughs have been achieved?
- AlphaFold 3 (DeepMind): Provides high-accuracy protein structure predictions.
- ISM001-055 (Insilico Medicine): An AI-discovered fibrosis candidate that advanced to human trials in under 18 months Nature Biotechnology.
- MIT Jameel Clinic antibiotic discovery: Identified novel compounds that evade traditional screening methods.
How much investment is going into AI for drug discovery?
In 2024, global venture funding reached $3.3 billion. Notable deals include:
- Generate:Biomedicines & Novartis: $1 billion partnership (2024).
- Isomorphic Labs: Secured $600M+ for AlphaFold-integrated drug design.
What are the latest regulatory developments?
In 2025, the US FDA issued draft guidance on AI in drug development, focusing on:
- Model validation
- Transparency
- Data governance
What challenges does the field face?
- Data bias & quality limitations
- Intellectual property (IP) and legal issues surrounding AI-generated molecules
- Need for explainability and human oversight to ensure safe, ethical adoption
