Pharma bets billions on AI labs to turbocharge drug discovery

The pharmaceutical landscape is witnessing a surge in strategic multi-partner collaborations that blend deep therapeutic expertise with cutting-edge AI technology to revolutionize drug discovery workflows. Nvidia and Eli Lilly’s $1 billion investment over five years to establish an AI lab exemplifies this trend, focusing on automating 24/7 experiments to accelerate scientific breakthroughs while easing the manual burden on researchers. This partnership reflects a broader industry movement where pharma giants like Lilly leverage AI infrastructure to drive innovation without straining their financials, mirroring similar initiatives such as Google’s DeepMind-driven agentic AI platforms.

Schrödinger’s collaboration with Bristol Myers Squibb (BMS) highlights how integrated AI workflows are scaling computational drug discovery to unprecedented levels, enabling exploration of trillions of molecular candidates with enhanced predictive accuracy. By jointly developing the Bunsen platform and RetroSynth synthesis planning tools, these partners combine physics-based models with large-scale molecular enumeration to improve drug candidate selectivity and streamline research operations. BMS’s planned large-scale deployment of Bunsen underscores the transformative impact of co-scientist AI platforms in accelerating pharmaceutical innovation and operational efficiency.

Eli Lilly’s strategic collaborations with Chinese AI pioneer Chai Discovery demonstrate a nuanced approach to integrating custom AI foundation models tailored to their proprietary data, enabling deeper embedding of AI throughout drug discovery pipelines from target identification to preclinical testing. This partnership not only aims to unlock previously inaccessible target spaces but also exemplifies how pharmaceutical companies are co-developing AI tools that become core to their workflows, enhancing both speed and scope of innovation.

The alliance between Bristol Myers Squibb and Chai Discovery to co-develop AI-driven antibody discovery platforms showcases the power of combining advanced AI models focused on molecular folding and biomolecular interaction prediction with deep therapeutic expertise. As Joshua Meier, CEO of Chai Discovery, notes, this collaboration is designed to rapidly shorten timelines from concept to viable therapeutics by creating a continuously learning AI system that accelerates discovery cycles and tackles historically challenging drug targets. This partnership further illustrates the growing trend of pharma companies integrating external AI capabilities to jointly innovate and enhance drug development efficiency.

AI is rapidly evolving from a mere experimental tool into foundational infrastructure within pharmaceutical R&D, with custom foundation models and agentic AI platforms becoming central to innovation. These agentic systems function like coordinated research teams, orchestrating multiple AI models and workflows to accelerate molecular discovery and clinical development by understanding objectives, planning tasks, and evaluating results autonomously. This shift enables seamless integration across diverse specialized tools, reducing manual hand-offs and speeding the transition from hypothesis generation to candidate prioritization.

Collaborations such as Chai Discovery’s partnership with Eli Lilly exemplify the strategic deployment of custom AI foundation models tailored to proprietary historical data, leveraging advanced Chinese AI technology to unlock previously inaccessible target spaces and accelerate drug discovery from inception through late preclinical stages. Similarly, platforms like NYB.AI’s Vecura demonstrate how agentic AI infrastructure manages fragmented specialized tools—including Drug-Target Interaction Graph Neural Networks and LigoSPACE models—to enhance operational efficiency and innovation in molecular research.

Agentic AI platforms are advancing toward closed-loop discovery systems that integrate information synthesis, experiment design, and data generation in automated cycles, significantly shortening drug discovery timelines. The integration of AI with emerging automated wet labs and experimental platforms, such as cell-on-a-chip technologies, enables rapid, cost-effective validation of in silico research, bridging computational predictions with real-world biological testing and enhancing the reliability and speed of scientific workflows.

Faro AI’s agentic platform illustrates how structured clinical development ontologies and custom foundation models convert complex scientific and operational concepts into machine-readable data, empowering AI agents to reason across clinical workflows rather than merely generate documents. Adopted by six of the world’s top ten pharmaceutical companies, Faro’s infrastructure automates interconnected decisions—from study design to risk identification—while preserving human oversight, underscoring AI’s emerging role as indispensable core infrastructure in clinical development.

Bullfrog AI exemplifies how integrating multimodal data—ranging from genomics to environmental factors—can dramatically enhance precision medicine by identifying patient subgroups with significantly improved outcomes, such as tripling pancreatic cancer survival from two to six months. This approach reflects a broader industry trend where even smaller biotech firms are embedding AI early in drug development to mitigate traditionally low success rates, underscoring AI’s transformative role in clinical trial optimization.

Pharmaceutical companies are evolving from merely partnering with AI technology providers to acquiring and owning proprietary AI models and infrastructure, as seen with Eli Lilly’s investment in supercomputing capabilities. This strategic shift aims to boost operational efficiency and innovation in clinical trials, enabling more tailored and agile drug development processes that leverage AI’s full potential.

AI-powered digital twins are revolutionizing clinical trial design by creating personalized placebo controls that predict disease progression with high accuracy, thereby addressing patient variability and reducing false positives or negatives in efficacy assessments. Springbach Analytics’ use of machine learning on whole-body imaging to forecast MRI outcomes exemplifies this innovation, which is gaining regulatory acceptance in Europe and enabling smaller, faster trials through improved patient matching.

Despite AI-driven advances in drug design and trial optimization, clinical trial enrollment remains a critical bottleneck, with only about 5% of cancer patients participating. AI’s growing role in accelerating patient recruitment and reducing reliance on animal testing in preclinical studies highlights its potential to streamline drug development timelines, yet overcoming human participation challenges remains essential for realizing AI’s full impact in precision medicine.

The partnership between Nvidia and Eli Lilly, marked by a $1 billion investment over five years, exemplifies how leading pharmaceutical companies leverage robust AI infrastructure to accelerate drug discovery. Eli Lilly’s substantial cash flow enables sustained investment in automating lab experiments around the clock, translating scientific discoveries into market-ready drugs more rapidly and underscoring computing infrastructure as a critical competitive differentiator.

While Nvidia GPUs provide essential computational horsepower, Generate Biomedicines CTO Gevorg Grigoryan stresses that compute alone is insufficient to secure a competitive edge in AI-driven drug discovery. Instead, proprietary biological data combined with continuous experimental feedback loops—where AI-generated molecules are experimentally validated and results fed back into the models—create a dynamic, self-improving platform that enhances prediction accuracy and operational efficiency over time.

Beyond capital and hardware, Nvidia’s strategic contribution to partners like Generate Biomedicines includes deep technical expertise in optimizing complex biological workloads on advanced GPUs. This technical exchange is pivotal for scaling AI platforms efficiently and extracting maximal value from computing resources, reinforcing how sophisticated infrastructure management complements proprietary data to sustain long-term competitive advantage.