What the Best Cell Therapy Partnerships Have in Common

Lessons from 25 Years of Cross-Border Deal-Making

This is the first article in a ten-part series on cell therapy partnership strategy.

In December 2017, Legend Biotech and Janssen signed a partnership that would produce Carvykti—approaching $1 billion in annual sales by 2024 and establishing the benchmark for BCMA-targeted CAR-T therapy. The deal structure was elegant: $350 million upfront, milestone potential of $1.35 billion, and a 50/50 profit-sharing arrangement that aligned both parties around commercial success.

That same year, multiple other CAR-T partnerships were signed. Most have not reached commercialization. Some were quietly terminated. A few produced assets that now sit in indefinite clinical hold.

After twenty-five years of structuring cross-border partnerships—between big pharma and big pharma, French and Japanese companies, American and Swiss corporations, small biotechs with fewer than 200 employees and enterprises with tens of thousands—I've observed patterns that separate partnerships that create lasting value from those that do not. The cell therapy sector, with its unique manufacturing complexity and rapid technological evolution, amplifies these patterns.

Five characteristics consistently appear in the partnerships that succeed.

1. Manufacturing Mastery from Day One

The most successful cell therapy partnerships share a characteristic that distinguishes them from traditional pharmaceutical licensing: the licensor brings production capability, not just a construct.

When Legend partnered with Janssen, they didn't offer a CAR design and a technology transfer package. They brought an integrated manufacturing operation that could produce clinical-grade product. When AstraZeneca acquired Gracell for $1.2 billion, they weren't buying a pipeline—they were acquiring FasT CAR-T, a manufacturing platform that reduces production time from weeks to days.

This matters because cell therapy manufacturing is not a commodity. Vein-to-vein time—the duration from patient apheresis to infused product—directly impacts clinical outcomes. Standard autologous CAR-T requires 4-6 weeks. Some platforms have reduced this to 7-14 days. The fastest validated processes achieve production in under 48 hours.

Data from multiple clinical programs now demonstrates that manufacturing time correlates with T-cell quality. Shorter production preserves naive and stem cell memory T-cell populations (Tscm)—the subset most strongly associated with durable responses. In one analysis of 50 patients across multiple trials, CAR-T persistence was the only factor significantly correlated with long-term event-free survival. Age, dose, prior transplant, tumor burden—none reached significance. Only persistence mattered, and persistence correlates with T-cell phenotype at infusion.

"Partners who control both design and production capture advantages that licensing arrangements cannot replicate."

The parallel to semiconductor manufacturing in the 1980s is instructive. During that period, Japanese companies dominated global chip production—not through superior design, but through integrated design-and-manufacturing operations. Japanese yields reached 70-80%, while American competitors topped out at 50-60%. The products were both better and cheaper. By 1989, six of the top ten global chip makers were Japanese.

American companies like Texas Instruments and Motorola had separated design from production. When yields matter, that separation creates structural disadvantage. The companies that eventually regained competitiveness did so by re-integrating—Intel's "copy exact" manufacturing philosophy emerged directly from this crisis.

Cell therapy is at a similar inflection point. Partners who control both CAR design and production capture advantages that pure licensing arrangements cannot replicate. The design decisions that optimize manufacturing—shorter expansion protocols, simplified activation, reduced reagent dependency—emerge from teams that see both sides of the process.

2. Platform Value Beyond Single Assets

The Legend-Janssen partnership works because it isn't about one product. The underlying platform—the manufacturing process, the CAR architecture, the quality systems—can generate multiple clinical candidates. The partnership economics reflect this: profit-sharing rather than royalties, joint governance rather than milestone-driven handoffs.

Experienced acquirers distinguish between product companies and platform companies. A product company offers a single asset—valuable if that asset succeeds, worthless if it fails. A platform company offers a reusable manufacturing and design capability that can produce multiple candidates across targets and indications.

The economics differ substantially. In the China-Japan licensing benchmarks from 2020-2025, upfront payments typically represent 8-13% of total deal value. For single assets, the remaining 87-92% depends entirely on clinical success. For platforms, the remaining value includes option rights on future candidates—dramatically different risk profiles.

Consider the manufacturing platform dimension specifically. Companies that have validated rapid autologous production possess capabilities that extend beyond current products. The lentiviral vector expertise required for 48-hour CAR-T manufacturing is the same expertise required for in vivo CAR-T delivery. The quality systems that enable bedside production are the quality systems that will enable hospital-based next-generation therapies.

This is why platform partnerships command premium structures. When a partner brings both a validated manufacturing process and the engineering capability to extend that process to new modalities, they bring strategic optionality that single-asset licensors cannot match.

3. The Cost Trajectory: Learning from Monoclonal Antibodies

In 1986, when the first therapeutic monoclonal antibody reached market, manufacturing costs exceeded $1,000 per gram. Today, best-practice production achieves $50-100 per gram—a 10-20x reduction over four decades. New continuous processing platforms are targeting $10 per gram, with pilot data showing 50-80% cost reductions versus batch processing.

The drivers of this cost curve are now well understood: improved cell line productivity, continuous processing, economies of scale, and accumulated process optimization. The same trajectory is beginning in cell therapy.

Current CAR-T manufacturing costs range from $50,000 to $150,000 per dose. But the platforms emerging now—rapid autologous processes, allogeneic approaches, in vivo delivery—each represent different points on the cost curve. Rapid autologous reduces reagent consumption and facility time. Allogeneic enables batch production across patients. In vivo delivery converts a cell manufacturing problem into a vector manufacturing problem.

The strategic implication for partnerships is significant. The right partnership today is not necessarily the one that optimizes current manufacturing—it's the one that positions both parties for where manufacturing will be in five years. Partners with demonstrated cost reduction in autologous processing have the operational DNA to drive cost reduction in whatever modality ultimately dominates.

In a recent France-Japan partnership I advised, we structured option rights specifically around manufacturing milestones rather than clinical milestones. The logic: clinical outcomes depend on patient selection and trial design, but manufacturing improvements compound. A partner who reduces vein-to-vein time from 14 days to 7 days will likely reduce it further. That capability is worth more than a single Phase 2 readout.

4. Stage-Appropriate Deal Structures

The most successful partnerships match deal structure to asset and company stage. Analysis of Chinese biotech partnerships with Western and Japanese companies from 2020-2025 reveals a pattern: upfront payments average 8-13% of total deal value, with the remainder in development, regulatory, and commercial milestones.

The BeiGene-Novartis tislelizumab partnership illustrates optimal timing. BeiGene partnered after proof-of-concept, with robust Phase 3 China data in hand. The result: $650 million upfront—one of the largest in the sector—plus $1.55 billion in milestones and mid-to-high-teen royalties. Partnering post-POC with strong data commands premiums.

Contrast this with partnerships that failed. The Innovent-Eli Lilly collaboration for sintilimab struggled partly because the partnership preceded global regulatory strategy alignment. China-only data proved insufficient for FDA approval, and the partnership for the US market did not progress.

For early-stage cell therapy assets, the "Bridging Option" structure has emerged as particularly effective. The partner funds the bridging study—accepting development risk—and holds an option to acquire commercial rights at pre-agreed terms upon positive data. This de-risks both parties: the licensor doesn't lock in low valuation too early; the partner doesn't commit capital until clinical signal exists in the target population.

The Kite-Daiichi Sankyo partnership for Yescarta in Japan set the CAR-T-specific benchmark: $50 million upfront for Japan-only rights, with low-to-mid double-digit royalties and $200 million in total milestone potential. This remains the reference point for single-asset, single-territory CAR-T deals.

5. Geographic and Cultural Complementarity

The Legend-Janssen partnership works in part because each party brings geographic capabilities the other lacks. Legend provides China market access and manufacturing; Janssen provides global regulatory expertise and commercial infrastructure. Neither could execute as effectively alone.

This complementarity extends beyond market access to regulatory strategy. Cell therapy regulatory pathways differ substantially across jurisdictions. NMPA (China), FDA (US), PMDA (Japan), and EMA (Europe) each have distinct requirements for CMC documentation, clinical trial design, and manufacturing site qualification. Partners who can navigate multiple pathways simultaneously—with genuine expertise rather than contracted consultants—move faster.

Cultural alignment matters more in cell therapy than in small molecule licensing because the ongoing manufacturing relationship is continuous. A traditional license transfers knowledge and then operates independently. Cell therapy partnerships often require ongoing technology transfer, manufacturing oversight, and joint problem-solving as processes are adapted to new facilities.

In my experience, partnerships that span geographies succeed when they establish direct relationships between operating teams rather than routing everything through business development. The JW Therapeutics-Takeda collaboration for relma-cel worked partly because Japanese BD teams engaged directly with the Chinese technical organization—not through US intermediaries. Direct communication compressed timelines and built the mutual understanding required for manufacturing transfer.

The ASLAN Pharmaceuticals partnership with Zenyaku Kogyo in Japan demonstrates another pattern: Singapore-based intermediary structures can facilitate deals that direct China-Japan relationships cannot. Sometimes geographic complementarity requires a bridge.

Applying the Framework

When evaluating cell therapy partnerships, these five characteristics provide a practical framework:

Manufacturing integration: Does the partner bring production capability, or only a construct and data package? Can they demonstrate validated manufacturing with quantified quality metrics?

Platform extensibility: Is this a single-asset transaction or a platform partnership? What option rights exist for future candidates? Does the manufacturing capability extend to emerging modalities?

Cost trajectory positioning: Where is the partner on the manufacturing cost curve? Do they have demonstrated capability to drive cost reduction? Is the deal structured to capture manufacturing improvements?

Stage-appropriate structure: Does the deal structure match the asset's development stage? Are the economics aligned with actual risk distribution?

Geographic complementarity: Does each party bring capabilities the other lacks? Is there a mechanism for direct operational engagement, not just BD coordination?

The partnerships that generate returns—for acquirers, for licensors, and ultimately for patients—share these characteristics. The cell therapy sector is maturing rapidly, and the partnerships formed now will shape which companies lead the next generation of therapies.

The opportunity is substantial. The question is whether the partnership structures match the technical and commercial complexity of the assets involved.

Next in series: The Integration Edge

For advisory on cell therapy partnership strategy, contact Kerlann Advisory.