The Strategic Path Forward: Leveraging Industry Partnerships for Nonprofit Innovation


 

INFORMED PROSTATE CANCER SUPPORT GROUP NEWSLETTER

How UCSD's relationships with pharma leaders could accelerate open-source AI tools for non-patentable drug development

April 10, 2026

Bottom Line Up Front

UCSD already has established partnerships with major pharmaceutical companies including Pfizer, AstraZeneca, and others. Rather than asking big pharma to support nonprofit drug development directly (which misaligns with their profit incentives), universities could collaborate with industry partners to develop and refine AI drug discovery and trial tools—then release these tools as open-source or nonprofit-accessible resources. Big pharma gets validated, cutting-edge technology; nonprofits get access to sophisticated tools; universities cement leadership in translational science. This isn't charity. It's mutually beneficial innovation infrastructure.

UCSD's Existing Pharma Relationships: The Foundation

UCSD is not a passive academic bystander to pharmaceutical innovation. The institution has deep, active partnerships with industry:

Pfizer-UCSD Collaboration: Pfizer provided an initial investment of $5 million to support the establishment of the Center for Microbiome Innovation at UCSD, with potential for additional milestone payments based on successful development of new antibiotics. This partnership demonstrates Pfizer's willingness to fund early-stage discovery at UCSD when the science aligns with their strategic interests.

Skaggs School of Pharmacy Structure: The Translational Research Alliance, housed within UCSD's Skaggs School of Pharmacy and Pharmaceutical Sciences, actively develops pathways for researchers to collaborate with pharmaceutical companies, healthcare providers, and biotech firms. The school's strategic plan explicitly emphasizes "strong partnerships beyond our school" with "industry partners, community organizations and others."

Drug Discovery Institute: UCSD's Drug Discovery Institute (DDI) brings together faculty from across campus to translate fundamental insights into new drug technologies. The institute explicitly facilitates collaboration with pharmaceutical partners through regular brainstorming sessions and is cataloging a massive library of marine natural products—exactly the kind of resource-sharing infrastructure that could support nonprofit trials.

Regulatory and Clinical Infrastructure: Through its affiliated Moores Cancer Center and extended studies division, UCSD has regulatory expertise, clinical trial experience, and FDA knowledge that pharmaceutical companies value. The institution teaches clinical trial design, regulatory affairs, and pharmaceutical development—expertise that could be packaged into tools.

The point: UCSD has credibility with big pharma. The question is how to redirect this relationship to serve both commercial and humanitarian purposes.

The Strategic Model: Co-Developing Tools for Dual Use

The key insight is that UCSD and its pharma partners don't need to choose between commercial and nonprofit goals. They can develop AI tools that serve both markets—with different licensing arrangements for each.

How it works:

UCSD proposes a joint research project with a pharma partner (say, Pfizer or AstraZeneca) to develop AI tools for clinical trial optimization. The project would be structured as:

  • Industry Goal: Pfizer needs tools to design and conduct more efficient trials. They want to compress timelines and reduce costs for their Phase I/II programs.
  • Academic Goal: UCSD develops and validates the tools through Pfizer's trial infrastructure and data (with proper privacy protections).
  • Nonprofit Release: Once the tools are mature and validated on Pfizer trials, UCSD releases them as open-source or through academic licenses for nonprofit use.

Pharma Benefits:

  • Cutting-edge AI tools developed through rigorous academic collaboration
  • Validated on real trials with proven efficacy data
  • Proprietary version with additional features for commercial trials
  • Patent protection if desired on novel components
  • PR value: "Pfizer advances AI tools for rare disease research" is positive corporate messaging

UCSD Benefits:

  • Industry funding for tool development (shared R&D costs)
  • Access to pharma trial data for model training (anonymized/de-identified)
  • Publications and visibility in AI-pharma field
  • Demonstration that academic-industry partnerships can serve public health
  • Capability to lead nonprofit trials with sophisticated tools

Nonprofit Benefits:

  • Access to validated, industry-grade tools at no cost
  • Tools developed with pharma-scale rigor (no second-rate academic alternatives)
  • Community of support around tools (documentation, conferences)
  • Ability to run sophisticated trials (dandelion root) that would otherwise be impossible

Precedent: This Model Already Works

The strategic model of pharma-funded development followed by nonprofit release is not hypothetical. Multiple examples demonstrate it functions well:

1. Structural Genomics Consortium (SGC)

Since 2003, the SGC has been funded by pharmaceutical companies (GSK, Bayer, J&J, AbbVie) and governments to solve protein structures and develop chemical probes for difficult targets. All outputs—crystal structures, assays, inhibitors—are released without IP restrictions. Industry backers leverage these tools freely for their own drug discovery. This proves that pharma companies will fund tool development when tools are released openly if the company maintains freedom-to-operate and first-mover advantage.

2. AstraZeneca's Open-Source AI Tools

AstraZeneca has released multiple AI drug discovery tools as open-source on GitHub, including REINVENT4 (molecular design), aizynthfinder (retrosynthetic planning), and GraphINVENT (graph neural networks). These are industry-grade tools developed by AstraZeneca and now freely available to academics and nonprofits. AstraZeneca maintains proprietary versions with enhanced features for internal use.

3. OpenFold Consortium

OpenFold is a nonprofit AI research consortium developing free, open-source protein folding tools for biology and drug discovery. Members include biotech startups, academics, and major pharma companies (Biogen, others). Industry partners fund development in exchange for early access and freedom-to-operate. The tools are released openly while companies maintain competitive advantages through implementation and proprietary datasets.

The Pattern: Pharma companies fund or contribute to tool development when: (1) the tools are foundational (useful but not themselves drugs), (2) open release preserves competitive advantage (proprietary data, implementation know-how), and (3) pharma gets visibility and first-mover benefit.

Concrete UCSD Proposal: The Three-Tool Initiative

UCSD could propose to a pharma partner (e.g., Pfizer) a joint program to develop three AI tools:

Tool 1: Trial Design Optimizer – Uses machine learning on historical trial data to recommend optimized protocols, eligibility criteria, and sample sizes. Pharma uses proprietary version on their trials; open version released for nonprofit use.

Tool 2: Patient Matching Engine – NLP-based system for identifying trial-eligible patients from EHR data. Pharma uses on their sites; nonprofits get access for dandelion-type trials.

Tool 3: Real-Time Trial Monitor – Analyzes trial data continuously, flags safety signals, supports adaptive designs. Pharma implements; nonprofits get academic version.

Funding Structure: Pfizer funds development ($2-3M over 3 years). UCSD provides faculty time, clinical site access, and validation. Upon completion, UCSD releases tools under open-source or nonprofit license. Pfizer retains proprietary implementation and enhancements.

Timeline: Year 1-2: Development on Pfizer trials. Year 2-3: Validation and open release. By year 4: Nonprofits using tools on dandelion, lemongrass, and similar candidates.

Addressing Big Pharma's Competitive Concern: Dual Licensing for Nonprofit-Only Use

Your concern is valid and critical: why would Pfizer fund development of tools that might be used by their competitors?

The answer is dual licensing—a well-established model that allows the same software to be released under different licenses for different users. The key is that UCSD releases the tools under an AGPL (Affero General Public License) with a commercial license option, explicitly restricting free use to 501(c)3 nonprofits only.

How Nonprofit-Restricted Dual Licensing Works

License Structure:

  • Free License (UCSD-maintained): AGPL with nonprofit restriction. Code is publicly available, but commercial use requires a paid commercial license.
  • Commercial License: A traditional proprietary software agreement that Pfizer (and other pharma companies) must purchase to use the tools in their commercial trials.

In Practice:

  • A nonprofit like Medicines for Malaria Venture or a dandelion root charity: uses the tool for free under AGPL with nonprofit clause
  • Pfizer wanting to deploy the tool in their trials: must purchase a commercial license from UCSD (price negotiated)
  • A competing biotech company wanting to use the tool: must also purchase a commercial license
  • An academic researcher at a university: typically falls under nonprofit clause (university is a 501(c)3), free use

The Nonprofit Restriction Clause: The key legal language restricts free use to organizations that are tax-exempt under Section 501(c)(3) and don't generate commercial revenue from the tool. Any for-profit company—competitor or not—must license commercially.

Why This Protects Pfizer's Interest:

  • Competitors cannot use the free version without paying for a commercial license
  • UCSD controls the commercial licensing terms (can include non-disclosure agreements, usage restrictions, pricing tiers)
  • Pfizer likely negotiates a favorable rate as a co-developer (partner discount)
  • Pfizer maintains proprietary enhancements for its own version (not shared in the free release)

Precedent: Nonprofit-Restricted Licensing Already Exists

This model is not hypothetical. Multiple major software projects use nonprofit-restricted licensing:

1. Business Source License (BSL)

Companies like MariaDB and others use BSL to restrict free use to nonprofits, academia, and small developers, while requiring commercial licenses from businesses. BSL is source-available (code is visible) but usage-restricted. After a defined "Change Date" (e.g., 3 years), the code converts to a fully open license. This model allows companies to fund development while protecting against competitor free-riding.

2. Qt Framework (Dual GPL + Commercial)

Qt, a major GUI framework, is released under AGPL for open-source projects (including nonprofits and academic users) and under a commercial license for companies building proprietary software. Hobbyists and academic researchers use Qt free under AGPL. Companies pay for commercial licenses. Qt maintains a clear distinction: nonprofits don't pay, commercial entities do.

3. MySQL (Oracle)

Oracle MySQL comes in multiple editions: Community Edition (GPL, free for everyone including nonprofits), Standard Edition (commercial subscription), and Enterprise Edition (commercial subscription with support). Nonprofits and academics use the community edition free. Companies must purchase subscriptions.

The Pattern: Major software companies successfully use dual licensing with nonprofit carve-outs. It's legally enforceable (courts have upheld both open-source and commercial licenses), well-understood in the industry, and creates revenue streams while supporting nonprofits.

Why This Model Works for UCSD-Pharma Partnership

For Pfizer (the funding partner):

  • Commercial tool development is funded upfront by both Pfizer and UCSD
  • Tools are validated on Pfizer's trials (improving their ROI)
  • Competitors must pay for commercial licenses—UCSD negotiates pricing
  • Pfizer gets a partner discount or even a cut of commercial licensing revenue
  • Nonprofits get free access (good PR for Pfizer: "We funded tools for rare disease research")

For UCSD:

  • Receives co-development funding from Pfizer
  • Owns intellectual property and can license commercially to generate revenue
  • Establishes itself as leader in nonprofit drug development infrastructure
  • Generates licensing revenue from commercial pharma users

For nonprofits:

  • Access to industry-grade tools at zero cost (nonprofit clause)
  • Can run sophisticated trials (dandelion, lemongrass) that would otherwise be impossible
  • No concern about license costs—nonprofit status grants free access

For academic researchers:

  • Universities are 501(c)3 nonprofits—faculty and students get free use
  • Encourages next generation of researchers to use the tools

How to Structure the License Agreement

Step 1: Establish Copyright Ownership – UCSD (as copyright holder) or a designated nonprofit foundation retains full rights to the software and can relicense it under any terms.

Step 2: Define "Nonprofit Use" – Software is free only for:

  • Organizations with 501(c)(3) status under U.S. tax law
  • Educational institutions (universities, schools)
  • Government agencies (NIH, FDA, etc.)
  • Individual researchers with no commercial application

Step 3: Define Commercial Use (Requires Paid License) – Any for-profit company, biotech, pharma, or CRO must pay for a commercial license, including:

  • Use in commercial drug development pipelines
  • Incorporation into proprietary products
  • Use where revenue is generated from the tool or tools incorporating it

Step 4: License Language – Use AGPL 3.0 with a nonprofit amendment (Non-Profit Open Software License 3.0 is an approved option, or a custom clause added to AGPL stating: "This software is free for nonprofit use. For-profit organizations must obtain a commercial license from [UCSD]")

Step 5: Commercial License Agreement – UCSD negotiates commercial licenses individually or offers tiered pricing (startup: $50K/year, mid-size pharma: $200K/year, large pharma: $500K+ /year with enterprise support)

The Pfizer Pitch Revised

With this structure, UCSD's pitch to Pfizer:

"We propose a co-development partnership where Pfizer funds 40% of AI drug discovery tool development ($2M over 3 years). UCSD funds 30% from research reserves; external foundations contribute 30%. Tools are validated on Pfizer trials. Upon completion, UCSD releases tools under AGPL with nonprofit carve-out—nonprofits and academic researchers use free; for-profit companies including Pfizer purchase commercial licenses at a negotiated partner rate (approximately 30% of standard commercial pricing as development partners). Pfizer gets the competitive advantage of early proprietary versions; we generate funding for further nonprofit capacity; nonprofits get access to industry-grade infrastructure."

Pfizer Benefits:

  • Validated AI tools at below-market development cost (shared R&D)
  • Preferential pricing on commercial licenses (partner discount)
  • Proprietary enhancements not shared in free version
  • Competitors must pay full commercial license rates
  • Strong PR: "Pfizer develops tools for rare disease research" (nonprofit clause visible)
  • No risk of competitor free-riding—all commercial use is licensed and paid

Why This Model Works Better Than Purely Open Source

Pfizer wouldn't want to fund tools available for free to competitors. The nonprofit-restricted dual licensing model solves this:

  • Pure open source (unrestricted): Any competitor can use for free → Pfizer objects
  • Nonprofit-restricted open source: Nonprofits use free; competitors pay → Pfizer agrees because competitors are still paying
  • Closed proprietary: Pfizer keeps tool proprietary → Nonprofits get nothing

The nonprofit-restricted model balances all incentives: Pfizer gets protected from competitor free-riding, UCSD generates licensing revenue, and nonprofits get access to sophisticated tools.

Why Pharma Would Say Yes (Revised)

With nonprofit-restricted licensing in place, the pharma pitch is much stronger:

Strategic Positioning: "Pfizer advances AI for rare disease research" is valuable positioning, but now the narrative is even stronger: "Pfizer partners with UCSD to develop AI tools. These tools are freely available to nonprofits and academic researchers; for-profit organizations license commercially." This addresses pharma criticism about access while protecting profits.

Competitor Pricing Control: Competitors cannot use tools for free. They must purchase commercial licenses, and UCSD can set aggressive pricing ($500K–$1M/year for large pharma companies). Pfizer, as a partner, gets a discount—reducing its effective tool cost below what competitors pay.

Revenue Sharing (Optional): If UCSD generates commercial licensing revenue (say, $2M/year from 4–5 pharma competitors), Pfizer could negotiate a revenue-sharing agreement: Pfizer gets 10–20% of commercial licensing revenue as a partner dividend. This transforms Pfizer's $2M co-development investment into an ongoing revenue stream.

Nonprofits as Proof-of-Concept: Pfizer gets to see the tools validated on nonprofit trials (dandelion root, etc.) before deploying them internally. Nonprofits effectively become beta testers with regulatory oversight.

Talent and Visibility: Partnership with UCSD on AI tools raises company profile. The nonprofit carve-out becomes a recruiting talking point ("We invest in tools for patients with rare diseases").

Risk Mitigation: Pfizer's R&D investment is protected—tools are not freely available to all competitors, only to nonprofits. Commercial entities must pay.

The Bottom Line: Pfizer funds tool development ($2M), gets validated tools, negotiates preferential pricing, maintains competitive advantage, and can even earn revenue sharing. Competitors must pay full price. Nonprofits get free access. Everyone wins, and Pfizer's financial interest is protected.

Implementation Pathway

Step 1 (Months 1-2): UCSD identifies which pharma partner is most aligned. Based on existing relationships, Pfizer or AstraZeneca are natural candidates.

Step 2 (Months 2-3): UCSD and pharma partner identify 2-3 specific tools to develop together. Tools should solve pharma's current pain points (trial cost, enrollment delays, data analysis).

Step 3 (Months 3-4): Negotiate joint development agreement with clear intellectual property terms: Pharma gets proprietary enhancements; UCSD gets rights to release academic/open-source version.

Step 4 (Months 4+): Co-develop tools on pharma trials. UCSD validates using Moores Cancer Center data for prostate/blood cancer applications.

Step 5 (Year 2-3): Release open-source versions. UCSD announces capability to lead nonprofit trials using industry-grade tools.

Step 6 (Year 3-4): First nonprofit trial (dandelion root, lemongrass, or similar) using UCSD-developed tools.

The Elegant Aspect: This doesn't ask pharma to sacrifice anything. It aligns pharma's profit incentive with public health. Big pharma gets validation and visibility. UCSD gets capability and prestige. Nonprofits get access. The dandelion root trial, which stalled after 5 years with 5 enrolled patients, could have been completed in 12-18 months with these tools.

Why This Works Better Than Direct Nonprofit Funding Requests

You might ask: why not just ask Pfizer to fund UCSD's nonprofit trial initiative directly?

Because that's not aligned with pharma's business model. Pfizer's goal is to develop profitable drugs, not to enable nonprofit competition. But if UCSD frames the proposal as "co-develop AI tools that improve Pfizer's trials," the ask is not threatening—it's valuable.

The clever part is what happens after: once the tools are developed on pharma trials and proven to work, releasing them to nonprofits costs pharma nothing (development is already funded) but generates goodwill and positioning. Pharma can honestly say "we funded this, and now it's available to rare disease research."

This is how open-source works in other industries. Red Hat built a business on Linux. Pharma can do the same with AI tools: fund development, release openly, compete on implementation and proprietary features.

Sources and Citations
[1] Skaggs School of Pharmacy and Pharmaceutical Sciences, UCSD. "Innovative Pharmaceutical Sciences Research."
https://pharmacy.ucsd.edu/research
[2] Skaggs School of Pharmacy and Pharmaceutical Sciences, UCSD. "Affiliate Research Centers, Institutes, Resources, and Training Programs."
https://pharmacy.ucsd.edu/research/affiliate-research-centers-institutes-resources-and-training-programs
[3] DrugBank Blog. (August 15, 2024). "The Collaboration Between Industry and Academia in Drug Development."
https://blog.drugbank.com/the-collaboration-between-industry-and-academia-in-drug-development/
[4] Skaggs School of Pharmacy and Pharmaceutical Sciences, UCSD. "Message from the Dean – 2025 Strategic Plan."
https://pharmacy.ucsd.edu/message-dean
[5] IntuitionLabs. (January 11, 2026). "Top 10 Open-Source Software Tools in the Pharmaceutical Industry (2026)."
https://intuitionlabs.ai/articles/top-10-opensource-software-pharma-industry
[6] EPAM. (November 6, 2025). "Open Source Advantage: The Tools Transforming Pharma Research & Development."
https://www.epam.com/insights/blogs/open-source-advantage-the-tools-transforming-pharma-research-and-development
[7] OpenFold Consortium. "Advancing Protein Structure Prediction and Drug Discovery."
https://openfold.io/
[8] NVIDIA BioNeMo. "Drug Discovery with Generative AI."
https://www.nvidia.com/en-us/industries/healthcare-life-sciences/biopharma/
[9] PMC/NIH. "Open-source approaches for the repurposing of existing or failed candidate drugs."
https://pmc.ncbi.nlm.nih.gov/articles/PMC3743608/
[10] GitHub. "Top Open Source Software from the Top 50 Pharmaceutical Companies."
https://github.com/servierhub/top-pharma50
[11] IntuitionLabs. "Open Source Pharma: Tools & Trends in Drug Development."
https://intuitionlabs.ai/articles/open-source-pharma-trends
[12] Open Source Initiative (OSI). "Frequently Asked Questions."
https://opensource.org/faq
[13] TermsFeed Blog. (November 10, 2025). "Dual Licensing Explained: How to Balance Open Source Principles with Commercial Profit."
https://www.termsfeed.com/blog/dual-license-open-source-commercial/
[14] Wikipedia. "Multi-licensing" (March 5, 2026).
https://en.wikipedia.org/wiki/Multi-licensing
[15] FOSSA Blog. "Dual-Licensing Models Explained, Featuring Heather Meeker" (December 13, 2023).
https://fossa.com/blog/dual-licensing-models-explained/
[16] Open Source Guides. "The Legal Side of Open Source."
https://opensource.guide/legal/
[17] SPDX License List. "Non-Profit Open Software License 3.0 (NPOSL-3.0)."
https://spdx.org/licenses/NPOSL-3.0.html

Comments

Popular posts from this blog

Dr. Christopher Kane of UCSD Health Appointed Chairman of the American Board of Urology

PSMA-Targeted Therapies for Prostate Cancer: Move Treatment Earlier in Disease Course

ASCO 2025: Non-Androgen-Receptor–Driven Prostate Cancer: Updates in Biology, Classification, and Management