In the high-stakes world of pharmaceutical research, where the average cost to bring a new drug to market now exceeds $2.6 billion, efficiency is the ultimate currency. For industry giants like AstraZeneca, the challenge lies not in a lack of data, but in the inability to distill meaningful, actionable insights from the massive, fragmented repositories of information housed within their biobanks. Enter Immunai, a New York-based startup that is fundamentally altering this landscape by positioning itself as the "high-end plumbing" of clinical research.
Immunai recently announced the third expansion of its strategic collaboration with AstraZeneca, a move that could see the startup net up to $37.5 million through 2027. This partnership marks a deepening of their relationship, integrating Immunai’s proprietary AMICA-OS platform into the very fabric of AstraZeneca’s vast clinical development pipeline.
The Evolution of a Strategic Partnership
The relationship between the two entities is not a recent phenomenon; it is the culmination of a half-decade of shared scientific endeavor. According to Immunai CEO Noam Solomon, the roots of the collaboration stretch back to the height of the pandemic, evolving from initial, targeted discussions into a multi-faceted operational alliance.
"We’ve known the AstraZeneca team for about five years," Solomon noted in a recent interview. The partnership began in late 2022 with a focus on oncology clinical programs. However, as the utility of Immunai’s platform became apparent, the scope expanded significantly. By October 2025, the collaboration had pivoted into Inflammatory Bowel Disease (IBD), signaling a broader shift for the startup.
The trajectory has been rapid: starting in immune-oncology, the partners expanded into broader oncology areas, moved into immunology and inflammation, and are now tackling complex therapeutic areas including cardiovascular inflammation, neuroinflammation, and even obesity and diabetes. The common denominator, Solomon asserts, is the immune system—the central regulatory hub of human health that Immunai aims to map through its foundational AI models.
Solving the Infrastructure Bottleneck
For an organization of AstraZeneca’s scale—boasting approximately 95,000 employees and managing over 100 Phase 3 clinical trials—integrating a startup’s technology is an exercise in logistical complexity. The collaboration involves dozens of researchers, data scientists, and clinicians on both sides of the table.
Solomon often refers to his company’s role as that of a "high-end plumber." While the metaphor may seem modest, the reality is a sophisticated intervention into the "plumbing issues" that cause multi-billion dollar drug programs to stall. These bottlenecks often stem from an inability to stratify patients effectively, an inability to predict toxic events, or a failure to optimize dosing schedules because the underlying immunological mechanisms remain obscured.
By providing a platform that can interpret the immune system’s state at a single-cell level, Immunai offers pharmaceutical partners the ability to look "under the hood" of their own clinical trials. They are not merely applying AI to existing data; they are transforming inert biobank samples into high-resolution, digital-ready insights.
The "Immune MRI" Process: Turning Samples into Data
Many AI-in-pharma companies focus on analyzing existing, often low-resolution, public datasets. Immunai, however, takes a "wet-lab-first" approach. Every collaboration begins at their laboratory at 430 East 29th Street in New York, where clinical patient samples are shipped for deep processing.
The process is akin to a high-resolution "immune MRI." When a sample is processed, the platform performs single-cell, multi-omic profiling, capturing a massive amount of information for every cell:
- Gene Expression: Approximately 37,000 measurements per cell.
- Surface Proteins: Around 75 specific markers.
- VDJ Sequencing: Analyzing the genetic diversity of immune receptors.
With a matrix of 10,000 cells per profile, the amount of data generated is gargantuan. By correlating these immune snapshots with clinical endpoints—such as progression-free survival or drug-induced toxicity—Immunai can help developers understand exactly why a drug worked in some patients but failed in others.
The Power of Resolution vs. Scale
A recurring theme in the critique of modern health-tech is the "big data trap"—the belief that having millions of records is sufficient, even if those records are shallow. Solomon is highly critical of this approach. He compares it to attempting to scale up a blurry, low-resolution photograph; no matter how much you zoom in, the image remains pixelated and useless.
"A lot of big numbers in this field don’t actually lead to better decisions or better insights because the data was collected without depth," Solomon explains. "If you need to distinguish between green and blue, but your data is black and white, you’re stuck."
Immunai’s AMICA database, which contains over 300,000 samples, stands out because 50,000 of them are captured at the "single-cell resolution" required to make clinical sense of the immune system’s nuances. This depth allows their foundation model to perform with precision, even when provided with small cohorts—sometimes as few as 20 patients—that would typically be statistically insignificant in traditional clinical analysis.
Broadening the Industry Footprint
The partnership with AstraZeneca is not an isolated case. In April 2025, Immunai joined forces with the Parker Institute for Cancer Immunotherapy to build one of the world’s largest single-cell datasets for immunotherapy research, utilizing 3,700 blood samples from over 1,000 patients.
Furthermore, in January 2026, Bristol Myers Squibb signed a multi-year partnership with the startup, focused on identifying patient subgroups and clarifying mechanisms of action for their pipeline drugs. These deals underscore a broader industry trend: Big Pharma is increasingly willing to outsource the "plumbing" of complex data integration to specialized, AI-driven firms that have mastered the art of biological data translation.
Implications for the Future of Medicine
The implications of this technology are profound. If drug developers can accurately predict how a patient’s immune system will react to a specific compound before a full-scale Phase 3 trial, the rate of clinical failure could drop significantly.
- Precision Stratification: Instead of treating a broad population, developers can identify the specific patient subgroups most likely to respond, leading to higher efficacy rates and fewer side effects.
- Rational Dosing: By monitoring the immune system’s reaction at a granular level, researchers can identify the "Goldilocks zone" for drug dosage, avoiding both under-dosing and toxicity.
- Combination Therapy Optimization: As monotherapies struggle to achieve the desired outcomes, Immunai’s ability to map immune pathways provides a roadmap for designing rational, effective combination treatments.
A New Era of "Biological Engineering"
As the partnership between Immunai and AstraZeneca enters this new phase, it serves as a bellwether for the future of the pharmaceutical industry. The era of "trial and error" in drug development is slowly giving way to an era of "biological engineering," where the immune system is treated as a programmable, interpretable, and highly readable biological computer.
While the $37.5 million price tag is significant, the real value lies in the time saved. For a company like AstraZeneca, shaving months or even years off the clinical development lifecycle by correctly identifying a failing program early—or, conversely, accelerating a successful one—is worth billions.
As Noam Solomon continues to "fix the plumbing" of the industry, the broader hope is that these digital tools will eventually make life-saving treatments not only more effective but also more accessible by lowering the astronomical barriers to entry that have defined the pharma industry for decades. The immune system, long viewed as a mysterious, chaotic frontier, is finally beginning to reveal its secrets to those with the right tools to observe them.
About the Author
