Viral Data Goes In — Do Vaccines Come Out? Taiwan's Gap in the PABS Negotiations

The World Health Assembly adopted the WHO Pandemic Agreement on 20 May 2025. That political and legal milestone is real: the Agreement covers surveillance, health systems, research and development, manufacturing, technology transfer, supply chains, financing, and international cooperation. ^[6][7]

But adoption does not mean the Agreement immediately opens for signature, still less that it has entered into force. WHA78.1 and Article 31 of the Agreement are explicit: only once the World Health Assembly has adopted the PABS annex referred to in Article 12 does the Agreement open for signature; after that, sixty instruments of ratification, acceptance, approval, formal confirmation, or accession must still be deposited before it enters into force after a set period. ^[5]

The A79/8 document submitted to WHA79 in 2026 is likewise not some final annex quietly completed behind closed doors. The document itself is labeled a screen text as of 30 April, not an agreed final text; different colors mark provisional agreement, provisional convergence, Bureau text, and Secretariat drafting, and the body still carries extensive bracketed language and alternative options. ^[4]

For example, the draft contains different formulations under which manufacturers would provide WHO with a share of real-time production in an emergency, part of it free of charge; it also contains options for annual monetary contributions, licensing, WHO-recognized sequence databases, and material access agreements. ^[4] This content can be used to map where the fault lines lie — it cannot be written up as a fixed percentage or legal obligation that is binding on companies today.

The most accurate way to state the timeline is this: the Pandemic Agreement has been adopted by the World Health Assembly; the PABS annex is still under negotiation; the Agreement has not yet opened for signature because the annex is unfinished, and it has certainly not entered into force, which requires sixty ratifications. ^[5][6]

The knowledge chain of a new outbreak typically begins with cases, samples, and sequences. A laboratory isolates the pathogen and completes sequencing; public health authorities and research teams use the data to judge transmissibility and danger; companies then develop diagnostics, vaccines, or treatments. The earlier the sharing, the greater the world's chance to respond ahead of the curve.

But the party providing the samples and sequences does not necessarily control manufacturing, patents, procurement, or logistics. Data can flow quickly into the research-and-development system, while the resulting products are delayed in returning to where they are most needed, held up by manufacturing capacity, pricing, advance purchase agreements, and political priority. PABS therefore asks not only "is the data open?" but also "how do the benefits generated from that data flow back?" ^[5][7]

Article 12 of the Agreement places both sides on the same footing, requiring the annex to define terms, modalities, legal nature, conditions, operations, and WHO's coordinating role, and requiring all elements to operate simultaneously under the annex. ^[5] That word "simultaneously" is the crux: if a data pipeline that only takes in is built first, with benefit-sharing left to future political goodwill, providers will not trust the system; but if every research use first has to clear a complicated benefit-negotiation process, the world may also lose the speed of early warning.

The April draft attempts to connect the two through three kinds of infrastructure: a WHO-coordinated laboratory network handling physical materials; WHO-recognized sequence databases handling sequences and provenance tracking; and manufacturers and other users bearing benefit-sharing obligations through contracts or terms. ^[4] The formal report from the seventh round in July shows negotiators are still comparing federated and hybrid architectures, and have pushed further work on the scope of pathogens covered and the conceptual framework to informal sessions ahead of the eighth round. ^[2]

That confirms the hardest part is not the phrase "share fairly" — it is who is included, who signs a contract, what use triggers an obligation, how data is tracked continuously, and who has the authority to allocate in an emergency.

Discussion of Taiwan tends to fall into two overstatements. The first is "Taiwan is completely excluded, so nothing can be done." The second is "Taiwan's scientific capacity is strong and data can still be shared, so none of this matters." Both compress a multilayered institutional question into a single question of WHO membership.

Taiwan's Centers for Disease Control (CDC) said plainly in 2024: Taiwan is not a WHO member, cannot take part in the relevant meetings and discussions, and the government and experts can only analyze publicly available information while seeking meaningful participation. ^[8] The WHO's seventh-round report records that article-by-article work and next steps on the annex are taken up in closed sessions attended by IGWG members — that is, member states, associate members, and regional economic integration organizations — with certain observer delegations also listed, while relevant stakeholders may hear updates and offer views in dedicated sessions. ^[2]

So "no formal seat at the negotiating table" is a confirmed present fact. But that alone is not enough to support the absolute conclusion that "none of Taiwan's laboratories, research institutions, databases, or companies can participate through any channel whatsoever." PABS's actual network and contracts are not yet finished, and technical institutions may end up connecting through cooperative networks, databases, companies, or third-party arrangements in different ways.

Articles 31 and 32 of the Agreement describe eligibility for signature, ratification, and accession in terms of "all States" and regional economic integration organizations. ^[5] Whether and how Taiwan can be handled by depositary authorities and the WHO system involves international legal and political questions that this article does not have enough primary documentation to resolve. The responsible way to write this is not to hand down a verdict on the future, but to acknowledge: the gap in formal negotiating participation already exists; final treaty and operational status remains to be confirmed.

The real operational risk breaks down into at least five doors: who can enter the negotiations; which laboratories can join the WHO-coordinated network; which databases can be recognized; who signs contracts when Taiwanese institutions use or provide data; and through what eligibility, buyer, or partner Taiwan obtains vaccines, diagnostics, and treatments when WHO allocates them. Opening one of these doors does not mean the whole path is open.

The advantage of sequence data is speed and reproducibility. It also brings a tracking problem: once data is downloaded, recombined, derived, and passed along, the link between the original provider, the sample's source, the conditions of use, and any resulting product can break.

The April draft therefore proposes a range of options — persistent identifiers, source and laboratory information, data governance, user notification, data access agreements, downstream transfer, and handling of misuse. ^[4] But the brackets have not disappeared: some versions lean toward requiring users to accept binding conditions before access, while other language places more weight on open access and notification. The final balance is still being negotiated.

Taiwanese laboratories do not need to wait for final terminology to improve their internal controls now. At minimum, every sample and sequence can retain a record of the source institution, the basis for sampling and consent, the de-identification method, the sequence version, the upload time, the database identifier, who it was shared with, any restrictions, and any later corrections. These records cannot guarantee that benefits flow back, but they can prevent a future in which no one can even reconstruct where a piece of data came from and who used it when.

Above all, "publicly uploaded" must not be written up as an automatic voucher for a vaccine. Whether a database is folded into PABS, whether a download triggers a contract, how non-commercial research is distinguished from commercial development, and whether derivative data carries forward the original conditions are all still negotiation questions. ^[2][4]

Article 12.6 of the Agreement adopted by the WHO in 2025 already writes in a "20% target" for participating manufacturers to provide WHO with a share of their real-time production, with at least 10% as a donation floor; an announcement the same year likewise describes a 20% target. ^[5][7] But the Agreement has not entered into force, so this is not a delivery right that can be directly invoked today. The April annex screen text still places side by side options of at least 20%, 8% or 10% provided free of charge, 12% or 10% provided at non-profit prices, plus bracketed language on licensing, pricing, and other benefit-sharing schemes — reflecting that the contractual and operational design has not converged, not that any rate is already in force. ^[4]

These numbers make for easy headlines, but the most important thing for Taiwan's decision-making is not to calculate how many doses it might get first — it is to acknowledge three layers of uncertainty.

First, although the 20% target and the at-least-10% donation floor have already entered the text of the Agreement, how the final annex will operationalize the contracts, whether the percentages will rise, and the operating conditions for each product category remain unfinished, and the Agreement itself has not entered into force. Second, even once WHO obtains products, they are to be allocated according to public-health risk and need; through what status and channel Taiwan would be included has not been guaranteed by any public document. Third, a product moving from a global pool to a patient still has to pass through regulatory review, liability, procurement, cold-chain logistics, and clinical guidance; a global commitment is not a local delivery note.

None of this means PABS has no value for Taiwan. On the contrary, a fairer, more predictable global mechanism could increase overall supply, and might also make it easier for partners to fold Taiwan into regional arrangements. But that is a reasonable inference, not a right that exists today.

For that reason, the state cannot pin all its hopes on a single global allocation gateway. Bilateral advance purchase agreements, regional procurement cooperation, stockpiles of raw materials and key consumables, domestic fill-finish and manufacturing capacity, emergency-use review, and contracts with private suppliers all need to advance in parallel with PABS diplomacy — not wait for it.

From the vantage point of the Global South and pathogen-providing countries, the deepest distrust during a pandemic comes from an asymmetric exchange. The places that provide cases, samples, and sequences bear the biosecurity, stigma, and economic risk; the side that holds research and development, patents, manufacturing, and advance-purchase capital may end up getting priority access to the resulting products and revenue.

If PABS ends up merely "encouraging" donations, or lets companies freely decide which uses count as commercial and which data escapes institutional tracking, then pathogen sovereignty gets diluted the moment the data is uploaded. Open databases could also be used as a back door around material-transfer agreements.

This position carries a double warning for Taiwan. Taiwan is a high-capacity research and manufacturing node, and cannot simply demand that others share quickly; Taiwan is also a party whose formal governance standing is fragile, and must insist that its own data contributions be identifiable and that benefit-sharing and emergency supply come with enforceable arrangements — not reliance on verbal goodwill.

From the standpoint of open science and rapid response, if PABS ties every sequence use to complex contracts, unlimited liability, or prior approval, researchers and databases may simply opt out. The value of an early outbreak often comes from many teams analyzing data simultaneously, spotting risk signals first, and only then deciding which development paths are worth pursuing.

If provenance tracking and benefit-sharing are designed too finely, they can also create a false sense of security: the system spends enormous effort logging who downloaded what, without adding any testing capacity, clinical trials, factories, or distribution. For Taiwan, pouring limited resources entirely into legal drafting cannot substitute for domestic manufacturing and procurement fallback capacity either.

This counter-argument is not opposition to fairness — it is a demand that the system place its triggers in the right place. Non-commercial public-health risk assessment should stay fast; once a use moves into commercial development, bulk acquisition of materials, manufacturing, or sales, it should carry a stronger, enforceable benefit-sharing obligation.

The genuine common ground between the two sides is this: without speed, products will not appear in time; without fairness, providers will not keep trusting the sharing system. The most workable direction is to tier use cases — giving emergency surveillance and non-commercial risk assessment a standard fast track while still retaining identifiers and provenance, while commercial development, manufacturing, and sales carry stronger obligations through standard contracts, monetary or non-monetary benefit-sharing, and product availability commitments.

The April draft does contain some components that could develop in this direction, but whether the final system adopts them, and where the trigger lines are actually drawn, remains to be confirmed. ^[4] This article cannot dress up an ideal design as a negotiated consensus.

What would help Taiwan most is not betting on which bracketed option wins, but building a domestic process that can map onto whichever version prevails: if a database requires users to accept a contract, who signs on the institution's behalf? If manufacturers must provide benefits based on having used PABS materials or sequences, how does a company trace its upstream source? If non-party entities cannot directly obtain certain benefits, can eligible partners, regional procurement, or commercial contracts fill the gap?

The first task is cross-ministry, article-by-article tracking. Health and welfare, foreign affairs, economic affairs, agriculture, science and technology, justice, and procurement agencies cannot each watch only their own slice; they need to jointly maintain a shadow text, tabulating the scope, databases, laboratories, contracts, benefits, allocation, and Taiwan interface of every new version, so differences can be tracked.

The second task is to translate diplomatic demands into operational requirements. Rather than simply saying "Taiwan should participate," the more concrete questions are: which of Taiwan's laboratories could be included in a cooperative or recognized network? Can databases retain Taiwan's provenance and contact points? Can Taiwanese companies sign standard PABS contracts? Which qualified entity would handle product allocation and emergency procurement? These testable questions are more likely to yield trackable commitments.

The third task is to rehearse a broken link. Assuming Taiwan cannot directly become a formal node at some point, the government should pre-designate cooperating laboratories, procurement agents, regulatory-recognition arrangements, and transport and liability interfaces — while making sure any alternative arrangement does not erase Taiwan's data provenance and rights.

Hospitals, research institutions, and biotech companies should first draw up a "pathogen data life-cycle map." From sampling, storage, isolation, sequencing, de-identification, upload, download, and model analysis through to material transfer, joint research, and product development, every step should note who controls the data, the sample, consent, liability, and rights.

Companies should also model three scenarios: using physical materials, using only sequence information, and using models or targets derived from a sequence. Whether each scenario triggers registration, a contract, a monetary contribution, product reservation, licensing, or reporting is not yet settled — but internally, a company should at minimum know what it used, how its origin can be proven, and how downstream contracts will pick it up.

Small and medium-sized diagnostics and biotech companies especially need trade associations to provide standard contract terms and advisory support. If every company has to interpret the global system on its own, compliance costs become an advantage for large firms; if a company keeps no record of provenance at all, it may only discover after a product succeeds that it cannot prove its chain of rights.

The eighth round in September needs to be watched for three things: whether the scope of pathogens covered converges; how a federated or hybrid architecture would handle states and databases; and whether access and benefit-sharing produce an integrated text with fewer brackets. November should be watched for whether an agreed text emerges sufficient to convene a special World Health Assembly. ^[2][3]

Taiwan also needs to watch whether its own interface gets written into any public arrangement: named technical participation, cooperative networks, retention of data provenance, a manufacturer-contract pathway, and product-access fallback capacity. A single forum or a friendly statement can be progress, but it is not proof of completion.

Finally, two narratives need to be kept from closing the case prematurely. Supporters cannot claim "the Agreement has already made vaccine access fair for every country." Opponents cannot claim "PABS will necessarily strangle open science." Both have to wait for the final text and how it actually plays out.

The question PABS is trying to answer is actually a simple one: when the world asks a place to quickly share a pathogen and information that could save lives, can it just as quickly and fairly send back the vaccines, diagnostics, treatments, technology, and resources that result, to the people who need them?

As of 24 August, the answer is still being negotiated. The seventh round did not complete the annex, and the eighth round has not yet met; although the Agreement's text already contains a 20% target and a minimum donation floor, it has not entered into force, and the additional percentages, database, and contract options in the April annex draft cannot be treated as current rules either. ^[1][2][4][5]

Taiwan's difficulty runs one layer deeper: a formal negotiating seat is absent, but pandemic risk, data contributions, research cooperation, and product needs are not. That means Taiwan cannot simply wait for a political answer, nor can it use technical cooperation to paper over uncertainty about its rights.

The most practical approach is to close the chain segment by segment: data with provenance and identification, use with notification and terms, companies with contracts and benefit-sharing, government with procurement and manufacturing fallback, and diplomacy with concrete interfaces and verifiable commitments. Wherever a segment can only rely on goodwill, it should be flagged as a risk — not written up as a guarantee.

The next virus will not check whether Taiwan is a member first. Nor should Taiwan wait for the virus to arrive before asking, for the first time: what we share outward — how exactly does it come back as the capacity to protect our people?