1. Company Overview
Recursion Pharmaceuticals is a clinical-stage biotechnology company leveraging artificial intelligence (AI) and machine learning to discover and develop novel therapeutics. Their AI strategy centers on building the world's largest biological image dataset and using advanced algorithms to identify potential drug candidates. A robust and resilient supply chain, from raw compute power to cloud infrastructure and data acquisition, is essential for scaling their AI platform and accelerating drug discovery pipelines.
2. The Compute & Silicon Stack
Recursion's AI models require massive computational power. The company has diversified its compute partners, but is also investing in internal capabilities for specialized workload optimization.
| Company | Ticker | Role in Recursion Pharmaceuticals Stack | Competitive Moat |
|---|---|---|---|
| NVIDIA | NVDA | GPU provider for AI training and inference; DGX A100 systems | Dominant market share in high-performance GPUs for AI; CUDA ecosystem |
| AMD | AMD | GPU and CPU provider; EPYC processors for data processing and GPU acceleration | Increasing performance and price competitiveness in high-performance computing |
| Amazon Web Services (AWS) | AMZN | Cloud-based compute instances (e.g., P4d instances with NVIDIA A100 GPUs) | Largest cloud provider; wide range of services and global infrastructure |
| Groq | Private* | AI inference engine for real-time drug candidate prediction (likely via a SaaS model) | Novel Tensor Streaming Architecture (TSA) optimized for inference, lower latency |
| Recursion Pharmaceuticals | RXRX | Custom FPGA acceleration for image processing and feature extraction | Proprietary algorithms and custom hardware designs tuned to their specific data and models |
* Groq is expected to IPO by the end of 2026, but as of March 14th, it is still private.
3. The Software & Model Stack
The software stack is crucial for building and deploying AI models. Recursion relies on a combination of open-source tools and proprietary software. They are also increasingly partnering with specialized AI model providers.
| Company | Ticker | Role in Recursion Pharmaceuticals Stack | Competitive Moat |
|---|---|---|---|
| Databricks | Private* | Unified data analytics platform for data engineering, machine learning, and real-time analytics; Spark deployments | Strong data science platform, optimized for large-scale data processing |
| Weights & Biases | Private* | MLOps platform for tracking experiments, visualizing performance, and managing models | Best-in-class developer tools and active developer community |
| Schrödinger | SDGR | Physics-based modeling and simulation software for drug discovery; integration with AI models | Established position in computational chemistry; validated physics-based methods |
| Exscientia | EXAI | AI-powered drug discovery platform; potentially licensing specific AI models or algorithms for specific target classes | Specialized AI models focused on drug discovery; successful track record in clinical trials |
| Recursion Pharmaceuticals | RXRX | Proprietary AI models for image analysis, target identification, and drug candidate prediction | Largest biological image dataset; domain expertise in biology and drug discovery |
* Both Databricks and Weights & Biases are expected to be publicly traded companies within the next 12-18 months, but as of March 14, 2026, they are private entities.
4. The Data & Infrastructure Stack
Storing, processing, and managing large datasets requires robust infrastructure. Recursion leverages a hybrid cloud approach for flexibility and scalability.
| Company | Ticker | Role in Recursion Pharmaceuticals Stack | Competitive Moat |
|---|---|---|---|
| Amazon Web Services (AWS) | AMZN | Cloud infrastructure for data storage, processing, and compute (S3, EC2, SageMaker) | Largest cloud provider; mature services and global reach |
| Snowflake | SNOW | Cloud-based data warehousing and data lake solution | Scalable data warehousing solution optimized for analytics |
| Digital Realty Trust | DLR | Data center provider for on-premise infrastructure and colocation services | Global network of data centers; interconnection capabilities |
| Equinix | EQIX | Data center provider for on-premise infrastructure and colocation services | Global network of data centers; interconnection capabilities |
| Illumina | ILMN | Genomic sequencing and data generation technology, potentially providing raw genomic data for analysis (though this is becoming less critical as other data modalities grow) | Dominant market share in genomic sequencing (data source, not infrastructure, but relevant). Risk: ILMN is trying to become more of a software play, so it could eventually overlap more directly with RXRX. |
5. Manufacturing & Hardware Partners
While Recursion is primarily a software and data-driven company, they do have some hardware dependencies for their high-throughput screening and automated microscopy systems.
| Company | Ticker | Role in Recursion Pharmaceuticals Stack | Competitive Moat |
|---|---|---|---|
| Thermo Fisher Scientific | TMO | Supplier of laboratory equipment, reagents, and consumables for high-throughput screening | Broad portfolio of life science products; established relationships with research institutions |
| PerkinElmer | PKI | Supplier of automated microscopy systems and image analysis software (though Recursion increasingly builds its own) | Established player in laboratory automation; strong customer base |
| Hamilton Company | Private* | Robotics and automation solutions for liquid handling and sample preparation (integrated with imaging) | Expertise in precision liquid handling; custom automation solutions |
*Hamilton Company is not publicly traded.
6. The Moat Analysis
Recursion's supply chain exhibits several key strengths, but also faces some vulnerabilities.
- Key Concentration Risks: Reliance on NVIDIA for high-performance GPUs remains a concentration risk, though diversification into AMD and internal FPGA acceleration efforts are mitigating this. Dependence on AWS for cloud infrastructure is also a potential concern, but multi-cloud strategy provides some redundancy.
- Vertical Integration: Recursion is actively pursuing vertical integration by developing its own AI models, image analysis algorithms, and custom FPGA-based hardware acceleration. This reduces reliance on external vendors and creates a proprietary advantage. The company is also aggressively expanding its proprietary datasets, which act as a major barrier to entry.
- Geopolitical Risks: The reliance on TSMC for semiconductor manufacturing introduces geopolitical risks related to Taiwan/China relations. While Recursion does not directly purchase from TSMC, its reliance on NVIDIA and AMD indirectly exposes it to these risks. The rise of domestic US chip manufacturing, particularly from Intel, could mitigate this risk over time.
7. Investment Outlook
The Bull Case
Recursion's AI-driven drug discovery platform has the potential to significantly accelerate drug development and reduce costs. Their increasingly vertically integrated and robust supply chain is critical to realizing this potential. Continued growth in data generation, model performance, and drug candidate pipelines should drive significant shareholder value. Recursion's ability to generate its own proprietary data is a huge competitive moat, and they are aggressively defending this. The key will be demonstrating that their models can translate to clinical success.
The "Picks and Shovels" Play
NVIDIA (NVDA): Regardless of whether Recursion or its competitors dominate the AI-driven drug discovery market, demand for high-performance GPUs will continue to grow. NVIDIA's dominant position in the GPU market makes it a key beneficiary of the AI revolution in drug discovery. Amazon (AMZN) continues to benefit from companies growing their compute stacks in the cloud.
The Bear Case
Key risks include:
- Supplier Concentration: Over-reliance on NVIDIA for GPUs and AWS for cloud infrastructure.
- Data Acquisition Costs: Increasing costs associated with generating and acquiring biological data. The cost to train complex models is going up.
- Regulatory Threats: Potential regulatory hurdles related to the use of AI in drug development (e.g., data privacy concerns, model validation requirements). Increased regulatory scrutiny on AI models could slow down drug approvals.
- Commodity Risk: GPU prices are highly cyclical, with significant fluctuations based on supply and demand.