Eli Lilly is taking a page from the software world and applying it to one of the most expensive, uncertain parts of modern medicine: finding new drug candidates. Instead of treating innovation as something that happens only inside large pharmaceutical labs, the company is increasingly positioning itself as a platform partner—funding and enabling smaller biotechs and research teams to plug into shared tools and workflows, with artificial intelligence at the center of the effort.
The move, described in recent reporting as a kind of “cash on the App Store” strategy for scientists, reflects a broader shift in how big pharma is approaching discovery. For years, the industry’s model has been relatively linear: identify targets, run experiments, screen compounds, optimize leads, and then scale up clinical development. But the cost and time required to move from early hypotheses to approved therapies have only grown. At the same time, the amount of biological data available—genomics, proteomics, imaging, patient records, and lab-generated datasets—has exploded. The bottleneck is no longer simply collecting information; it’s turning that information into decisions quickly enough to matter.
That is where AI and other advanced technologies come in. Lilly’s collaboration approach aims to accelerate drug discovery by giving researchers access to computational methods and innovation pathways that can shorten the distance between an idea and a testable candidate. The emphasis is not just on building models, but on creating an ecosystem where smaller teams can contribute and iterate—potentially faster than traditional discovery pipelines allow.
What makes this approach notable is the framing. “App Store” is a metaphor, but it points to a real structural change: Lilly is not only investing in internal R&D, it is also investing in the surrounding infrastructure that helps others build. In practical terms, that means working with small biotechs rather than relying exclusively on acquisitions or long, closed collaborations. It also means supporting the kinds of tools and platforms that can be used across multiple programs—so that learning accumulates rather than resetting with each new target.
Why weight-loss money matters here
Lilly’s weight-loss franchise—anchored by Mounjaro—has generated substantial cash flow, and that financial strength is now being deployed toward discovery capabilities. This is not unusual in principle; successful products often fund future pipelines. What is different is the direction of the investment. Instead of funneling resources solely into late-stage trials or incremental lab expansion, Lilly is using part of its momentum to underwrite a discovery ecosystem that can produce more candidates earlier.
In the current environment, speed is a competitive advantage. Drug discovery is not only about scientific creativity; it’s also about operational tempo. Teams that can rapidly generate hypotheses, prioritize experiments, and learn from results can outpace competitors who are still waiting for slower cycles of screening and analysis. By collaborating with smaller biotechs and enabling AI-driven workflows, Lilly is effectively trying to compress the timeline from “we think this might work” to “we have evidence it does.”
The “cash on the App Store” concept also suggests a willingness to fund innovation in a modular way. In software, developers don’t need to build an entire operating system to create value; they can build apps that plug into a larger platform. Translating that logic to drug discovery is ambitious, because biology is far messier than code. Still, there are parallels: if Lilly can provide access to data, compute, validation pathways, and integration support, then external teams can focus on specific components—such as target identification, molecule generation, property prediction, or experiment design—without having to recreate everything from scratch.
How AI fits into the discovery workflow
AI in drug discovery is often discussed as if it were a single tool that replaces wet-lab work. In reality, the most promising uses tend to be complementary. AI can help interpret complex datasets, predict molecular properties, identify patterns in biological systems, and propose candidate structures that are more likely to succeed. But those predictions still need experimental validation, and the quality of the AI output depends heavily on the data used to train and evaluate models.
Lilly’s collaboration strategy implies a pragmatic view: AI is valuable when it is embedded into a workflow that connects computation to experimentation. That means AI isn’t just generating ideas; it’s helping teams decide what to test next, how to design experiments, and how to interpret results. When done well, this reduces wasted effort—testing too many weak candidates, repeating experiments that could have been avoided, or failing to recognize signals early.
Small biotechs can be particularly effective in this context because they often move quickly and specialize in narrow technical areas. Some may excel at model development; others may focus on data curation, assay design, or specific therapeutic domains. By partnering with these teams, Lilly can diversify the sources of innovation while still steering the overall discovery agenda toward its strategic priorities.
The ecosystem approach: more than one-off partnerships
Large pharmaceutical companies have long collaborated with startups, universities, and research institutes. The difference here is the apparent intent to create a repeatable pathway—an “innovation marketplace” rather than a series of isolated projects.
A platform-like approach can offer several advantages:
First, it can reduce friction. If Lilly provides standardized ways to integrate tools, share data, or align on evaluation criteria, external teams can spend less time negotiating bespoke arrangements and more time building and iterating.
Second, it can improve learning loops. When multiple teams work within a shared framework, the organization can compare approaches, identify which methods perform best, and refine the platform over time. That creates compounding returns: each collaboration doesn’t just produce one result; it improves the system that produces results.
Third, it can broaden the talent and idea pool. Big pharma has deep expertise, but it can also become insulated by its own processes. Smaller biotechs often bring different perspectives and technical approaches, especially in fast-moving areas like machine learning, automation, and data engineering.
This is where the “App Store” metaphor becomes more than marketing. In a true platform model, the platform owner benefits from a growing catalog of capabilities. Lilly’s goal appears to be to cultivate that catalog—supporting multiple “apps” (tools, models, and workflows) that can be used across discovery programs.
What “accelerating drug discovery” really means
Accelerating drug discovery can sound vague, but it usually translates into concrete operational changes. In practice, it can mean:
1) Faster target-to-candidate cycles
AI can help prioritize targets and generate candidate molecules more quickly, reducing the time spent on low-probability paths.
2) Better candidate triage
Instead of running expensive assays on a large number of compounds, teams can use predictive models to rank candidates and focus resources on the most promising ones.
3) Improved experiment design
AI can suggest which experiments will yield the most informative results, helping teams learn efficiently from each round of testing.
4) More robust interpretation of results
Biology is noisy. AI can help detect subtle patterns and reduce the risk of missing signals due to variability in experimental outcomes.
5) Earlier detection of failure modes
Models can flag likely issues—such as poor predicted properties or potential safety concerns—before candidates advance too far.
The key is that acceleration is not only about speed; it’s about maintaining quality while moving faster. A platform that enables rapid iteration must also ensure that outputs are reliable enough to justify experimental follow-through. That is why collaboration with biotechs matters: it can bring specialized expertise in model validation, data handling, and experimental integration.
A unique take: Lilly as a “discovery integrator”
Many observers expect big pharma to either build AI internally or buy companies that already have AI capabilities. Lilly’s reported approach suggests a third path: acting as a discovery integrator. Rather than treating AI as a standalone product, the company is treating it as a set of capabilities that must be integrated into a broader discovery system.
This integrator role is often overlooked, but it may be the most important. AI models can be impressive in isolation, yet fail to deliver value if they cannot be connected to real-world lab workflows. Data formats may not match. Assays may differ. Validation criteria may be inconsistent. Teams may not have the right feedback loops to retrain models based on new results.
By collaborating with small biotechs and supporting their work within a shared framework, Lilly can potentially standardize the integration layer—making it easier for AI tools to translate into actionable decisions. In other words, the “platform” is not just compute or software; it’s the operational glue that turns predictions into progress.
This also helps explain why the strategy is described as “cash on the App Store.” It implies that Lilly is funding the creation and deployment of tools that can be used by scientists, not merely funding research papers or prototypes. The emphasis is on enabling adoption—getting AI into the hands of teams who can apply it to discovery tasks.
The broader industry context
Lilly’s move fits into a wider trend across healthcare and life sciences. Companies are increasingly recognizing that AI’s value depends on data access, workflow integration, and iterative validation. Meanwhile, smaller biotechs are looking for partners who can provide not only funding but also pathways to apply their technology to real drug discovery programs.
For Lilly, partnering with small biotechs can also mitigate risk. Discovery is inherently uncertain. If Lilly relies entirely on internal models and internal pipelines, it concentrates risk in one approach. By diversifying across multiple external teams and technical strategies, Lilly can hedge against the possibility that any single method underperforms.
There is also a strategic signaling effect. When a major pharma company invests in an ecosystem, it attracts more talent and more startups. That can create a virtuous cycle: more participation leads to more innovation, which leads to better outcomes, which leads to more participation.
What to watch next
While the reporting highlights the collaboration and the AI focus, the details of how these partnerships are structured will determine whether the strategy delivers sustained impact. Several questions are likely to matter:
– How will Lilly evaluate success across different AI approaches?
If evaluation metrics are unclear or inconsistent, it becomes hard to compare tools and scale what works.