Behind the hundred billion dollars, you can’t just look at headline value.

In the first half of 2026, the potential total amount disclosed for China’s innovative drug license-out deals will approach $99.7 billion, with Chinese companies occupying eight spots among the top ten global pharmaceutical transaction volumes. China’s innovative drugs going global are moving from sporadic breakthroughs to large-scale and normalized approaches.

But first, a professional distinction is needed: the so-called “total transaction amount” is usually the sum of the down payment, development milestone, registration milestone, sales milestone, and even future sales revenue sharing. It does not mean the company has already received it or will ultimately realize the revenue. The down payment determines current certainty, milestones define how risk is shared, sales revenue sharing determines long-term returns, and IP equity arrangements determine which value the licensee can continue to keep, which value may be absorbed by the other party during cooperation, and whether assets can be fully returned to square one after project failure or termination. Particularly noteworthy is that more and more transactions are no longer satisfied with the traditional “license-to-collect” model and are adopting Co-Co models based on co-development, co-commercialization, or profit-sharing. The transaction parties have shifted from a licensing relationship to a long-term partnership, and the IP issue has escalated from “which patents to license” to “who owns, who can use it, and who decides the value increment over the next decade.”

01 From “selling young seedlings” to “growing together”: The transaction model is undergoing structural changes

Over the past decade, license-out transactions for Chinese innovative drugs have mainly adopted traditional models such as “one-time buyout” or “down payment + milestone + sales revenue sharing.” As licensors, Chinese pharmaceutical companies grant development and commercialization rights outside China to multinational pharmaceutical companies, while retaining their own interests in China. The advantage of this model is its simple structure and high maturity, but the downside is that Chinese pharmaceutical companies are relatively weak in profit distribution in overseas markets.

Trading data from the first half of 2026 reveals a structural shift: the “Co-Co” model is emerging.

The so-called “Co-Co” refers to Chinese pharmaceutical companies and multinational pharmaceutical companies not only cooperating on a transactional level but also deeply binding in R&D and commercialization—jointly advancing clinical trials, sharing R&D costs, and sharing profits. This is not a simple “license fee” model, but a true “partnership” model. Chinese pharmaceutical companies are no longer just “selling” an early-stage asset, but are trying to retain their right to participate in the asset’s growth process and a higher share of value.

This change is partly due to improvements in the asset quality and bargaining power of China’s innovative drugs, and partly from the accumulation of international development and BD capabilities among Chinese pharmaceutical companies. But Co-Co is not simply about “holding on to some more rights.” It requires both parties to establish a long-term balance between R&D investment, profit sharing, regional responsibility, and decision-making power, and also transforms intellectual property from a transactional attachment into the underlying rules of cooperative governance.

Traditional License-out focuses more on the boundaries of rights on the signing date; Co-Co focuses more on the ongoing results generated during the cooperation period and whether both parties can independently and long-term implement their respective reserved interests. This structural change sounds beautiful, but when it comes to IP terms, it is full of pitfalls. On July 24, we and Dr. Tang Huadong prepared three real transaction case studies, analyzing everything from the Term Sheet to the final agreement terms step by step. First, clarify the logic on paper, then see how to sign in real games.

02 Why are the three IP issues in the Co-Co model more complicated than they appear?

First, jointly developed IPs: you can’t choose between “solely owned” or “jointly owned.”

The first question raised in the draft is how to attribute incremental intellectual property generated from joint development: according to the principle of “who invents, owns,” is it owned by the leading party and licensed to the other party, or jointly owned by both parties? This is indeed the core of the Co-Co deal, but simply agreeing on ownership is far from enough. First, different jurisdictions differ in the rules regarding joint inventions, the exercise of co-ownership, and whether co-owners can license them separately. The phrase “jointly owned by both parties” may make it difficult for either party to handle the matter independently; In some countries, it may also result in one party implementing or permitting the other without the consent of the other. Second, the results of the cooperation cannot be broadly classified as “Foreground IP.” At least three categories should be distinguished: product-specific results formed around licensed products; Universal platform results applicable to multiple projects; Independent results not directly related to the collaborative project. For example, new indications, new dosage regimens, and product-specific processes are usually product-specific achievements; New connectors, delivery systems, expression platforms, or analysis platforms may affect other pipelines of the licensor and cannot be fully transferred due to this project collaboration.

Again, ownership does not equal enforceability. Even if the licensor retains its Chinese interests, if overseas clinical data, CMC improvements, dosing regimens, or regulatory information cannot be used, its China rights may only be nominal rights. Mature programs should simultaneously arrange outcomes attribution, cross-licensing, data usage, sublicensing, and continued use rights after termination.

Therefore, to judge whether the distribution of results is reasonable, one cannot only ask “who owns the patents,” but must ask: whether both parties can fully develop, apply, produce, and continuously iterate products within their respective regions and fields.

Second, the boundaries of the background IP: Is the license being granted to a single product, or is it unintentionally licensing the entire platform?

Each license-out distinguishes between background IP and prospective IP generated after cooperation. In the Co-Co model, the technologies of both sides continue to cross paths, making the boundaries of background IP more easily blurred. Especially in platform technologies such as ADCs, bispecific antibodies, cell therapies, and RNA delivery, a candidate drug may rely on universal linkers, payloads, delivery systems, expression systems, or production platforms. If a licensed IP is defined as “all intellectual property related to or useful to the product,” without specifying necessity, product, domain, or use, a single product license may be factual extended into a platform license, even affecting the licensor’s next-generation products and other pipelines. Special examination is also needed to define “Control.” If the licensor is considered controlled simply by granting a certain right, technology acquired from universities, partners, or acquisition targets in the future may automatically enter the licensing pool; Conversely, if upstream licensing, joint ownership, or government funding conditions do not allow for sub-licensing, the master protocol may lack a complete authorization chain.

Therefore, background IP should not be made into a single patent list; instead, boundaries should be defined by “product—target—indication—region—platform technology,” distinguishing assets that require licensing, selective licensing, and explicitly retained assets, while simultaneously covering unpublished applications, family members, divisional/continuing applications, trade secrets, materials, and data.

Third, maintenance and operation of patent portfolios: control must match economic interests

Under the Co-Co model, who is responsible for patent applications, geographic layout, maintenance, rights protection, and defense is superficially a procedural issue, but in essence, it is a matter of controlling the product lifecycle and commercial value. A review response, a divisional selection, a waiver in a key country, or a litigation settlement can all change the future market space for both parties. The licensor wishes to retain patent ownership and control over the application; Licensees responsible for global development investments hope to lead layout and rights protection. Both parties have commercial rationality; the key is to separate routine enforcement from major decisions: routine applications and maintenance can be led by one party, but when core claims are narrowed, key states waive, division/continued applications, invalidation, litigation, and settlement, the other party should have the right to know, opinion, or consent. If one party is unwilling to continue defending a patent in a certain country, the other party should have the right to take over; If interests conflict in different regions, a mechanism for avoidance of interests and final decision-making should be agreed upon in advance; The costs of rights protection, litigation gains, adverse judgments, and the impact of third-party settlements on the reserved area should also be allocated in parallel. Joint Patent Committees can serve as Co-Co governance tools, but only when authority, voting, deadlock resolution, and emergency procedures are clearly defined. Otherwise, it is just a meeting mechanism.

03 Two “Second-Stage Risks” That Are Easily Overlooked in the Co-Co Model

01 Co-development doesn’t mean waiting indefinitely: How to avoid “buying but not doing”

Co-Co emphasizes long-term cooperation, but pipeline priorities cannot always be aligned. After obtaining exclusivity, licensees may lower project priorities due to clinical outcomes, budget changes, competing products, or internal strategic adjustments. Although licensors nominally share future profits, they may miss the opportunity to retrade during the most valuable window period.

Therefore, development obligations cannot remain merely “commercially reasonable efforts.” Development plans, key milestones, minimum investment or resource commitments, periodic reporting, and remedial deadlines should be combined, and if standards are consistently not met, the rights to reclaim areas, convert monopoly to non-monopoly, and adjust or terminate joint development rights should be established. Co-Co offers opportunities for mutual growth, not the right to lock assets indefinitely.

02 After the Cooperation Terminates, What Should Be Returned Should Be ‘Assets That Can Continue to Be Developed’

Innovative drug development has a high failure rate; Co-Co contracts must be designed both for success and for ending collaboration. A nominal patent reversal does not mean the project can be restarted. After termination, it is necessary to simultaneously handle trademarks, regulatory filings, clinical and pharmacovigilance data, samples, cell lines, process packages, third-party CRO/CDMO contracts, supply arrangements, and product-specific patents formed through collaborations. It should also be clearly defined whether the transfer is or licensing, how fees are calculated, who supplies the drug during the transition period, and who maintains clinical trial and safety obligations.

If the licensee is acquired by a competitor, or if there are more priority projects with the same target internally, considerations must also be made regarding information isolation, ongoing development commitments, disposal of competitive assets, and equity recovery under specific circumstances. The exit mechanism is not a standard clause at the end of the contract, but rather a design that gives the License-out asset a “second life.”

04 From IP to BD: Five Practical Recommendations for Implementation

First, pre-check IP due diligence into valuation and term sheets. Due diligence should not only verify ownership and validity but also evaluate protection periods, geographic coverage, FTO, upstream licensing restrictions, service inventions, data rights, and trade secret management, and ensure the results are fed back into down payments, milestones, warranty representations, indemnification, and delivery terms.

Second, emphasize the “rights system” rather than a single patent. A tradable asset is typically backed by a compound/sequence, use, formulation, process, materials, data, and technical secrets. The scope of licensing should cover the assets necessary for implementation, while explicitly excluding generic platforms and non-trading pipelines. Third, clarify the incremental path for results early in the trading phase. Anticipate outcomes that may result from clinical, CMC, formulation, combination therapies, companion diagnostics, and next-generation products, distinguish in advance between product outcomes, platform outcomes, and independent outcomes, and configure ownership, cross-licensing, and data usage rights. Fourth, ensure that control is matched with capital contributions, regional interests, and profit sharing. Establish tiered decision-making, takeover rights, and deadlock mechanisms for major matters such as patent layout, development plans, budgets, sublicensing, and litigation settlements, to avoid one party bearing the main economic risk without necessary control, or one party’s decisions harming the other’s reserved territory. Fifth, conduct a “failure scenario stress test” before signing. Assuming clinical failure, halted investment, competitor acquisitions, invalid key patents, supply interruptions, and agreement termination, each item is examined to see if rights can be recovered, documentation complete, and whether the project can continue development and re-license. Each of these five points alone is worth spending half a day carefully examining. Actual transactions often involve complex issues such as cross-jurisdictional rule alignment, the integrity of sublicensing chains, and the timing coordination between regulatory data rights and patent exclusivity. These will be discussed at the July 24 “From IP to BD: China’s Innovative Drug Trading Power Special Session,” where Dr. Tang Huadong and several frontline traders will analyze real cases one by one and analyze the design logic of difficult clauses on site.

The rise of the Co-Co model shows that Chinese innovative drug companies are no longer satisfied with selling overseas interests all at once, but instead seek to participate in global development, share long-term profits, and retain more strategic options. This is progress from “selling young seedlings” to “growing together,” but it also means that transactions must move from simple licensing to long-term joint governance.

For licensors, the best IP arrangement is not to keep all rights to themselves—that is neither realistic nor beneficial for partner investment; Instead, it is about reasonably supporting global development while ensuring the platform is not accidentally transferred, the reserved area is not reverse-blocked, the incremental output can be reasonably shared, patent control matches economic interests, and a truly sustainable asset can be reclaimed if cooperation stalls.

Ultimately, IP rights design protects not just a patent certificate, but the ability for companies to continue development, participate in global value distribution, re-trade, and reselect partners in the future.

About Us

MeritsIP: Your Global Intellectual Property Partner

At MeritsIP, we provide comprehensive IP services in biomedicine, medical devices, manufacturing, semiconductors, and AI. We support 200+ top clients globally with strategic, high-quality IP services.

Stay Connected

• • Website: https://www.meritsipglobal.com

• • LinkedIn: Follow us on LinkedIn

• • Email: meritsip@meritsandtree.com

© 2025 MeritsIP. All Rights Reserved.