In mid-September 2025, U.S. and Chinese officials convened in Madrid for a high-level negotiation that yielded a breakthrough on the TikTok deal: a “basic framework consensus” for resolving the long dispute over the platform.
This meeting marked the fourth round of China–U.S. trade and technology negotiations since Trump announced new tariffs in April. The possibility of a Xi–Trump meeting at the APEC summit in Seoul, expected by late October, added momentum to both sides’ efforts to secure a stronger position. Within this broader bilateral contest for the upper hand, the TikTok deal became one of several key bargaining points that reflected the shifting dynamics of the bilateral relationship, along with other heated topics such as tariffs, semiconductors, and rare earths.
TikTok as bargaining chip and export control regime showcase
Per the White House’s order, the consensus represents a triumph that guarantees the core concerns upheld by PAFACA: American ownership and data control. According to Trump’s September 25 Executive Order, a new TikTok US entity will be “majority-owned and controlled by United States persons,” with ByteDance and any Chinese affiliates holding less than a 20% stake. Noticeably, clauses (B) and (D) mentioned that the “the divestiture puts the operation of the algorithms and code……under the control of the new joint venture”, and “all recommendation models, including algorithms, that use United States user data to be retrained and monitored by trusted security partners”.
IGP has critically assessed the extent to which the deal achieved its original purposes. As the Chinese perspective has been largely absent in the English media, in this article, I will go through the official statements from the Chinese government to grasp their major concerns.
Wang Jingtao, the deputy chief of Cyberspace Administration of China (CAC), the national Internet regulator, also participated in the Madrid negotiation and stated that, “[Both sides] have reached basic consensus on problems including American user data and content safety hosting operations, and intellectual property authorizations on algorithms”, and Chinese government will continue to “review and screen technology export and intellectual property authorization relevant to the TikTok case”. As Financial Times quoted from an Asian investor, Beijing’s bottom line is a “licensing deal”.
For Beijing, the biggest value of a TikTok deal other than showing defence to Chinese corporate interests and a bargain chip in the US-China trade negotiation, is a showcase of its nascent export control rules over TikTok’s advanced recommendation algorithm, which ultimately serves to strengthen China’s capacity of technology export control. Covering many products and technologies, the Chinese export list of controlled products and technologies included “personalized information recommendation service technology based on data analytics” (基于数据分析的个性化 信息推送服务技术) in 2020, when Trump launched the first wave of TikTok ban threats.
This is another example of how Beijing has been actively learning trade policy instruments and tools previously developed and mastered by the U.S. government to defend its advantages over critical technologies, raw materials, and products. Such a policy learning process is evident in many places. For example, Beijing seems to have imitated American concepts and tools such as “de minimis” principle and entity/control list in the new Chinese dual-use products export control regulations proposed in late 2024. China’s recently announced export controls over rare earth materials and technologies also serves as such an example. Ironically, the outraged reaction of American trade and foreign policy analysts to the rare earth controls seemed entirely oblivious to the fact that it was the U.S. who initiated “weaponized interdependence” and that the Chinese side are simply mirroring the American tactics.
Possible playbook and hinging questions for the TikTok deal
Back to the TikTok deal, thinking logically based on the statements from Washington and Beijing, we can only reach one possible solution that aligns with both the White House Executive Order and Wang’s statement: the Chinese government reviews and potentially authorizes ByteDance’s export of its recommendation algorithm intellectual property to the newly established TikTok U.S. joint venture. This new entity would have absolute control over the operation of the algorithms, code, content decisions, and U.S. user data, all “retrained and monitored by trusted security partners.”
Singaporean Chinese-language media outlet Lianhe Zaobao cited three sources and multiple Chinese media reports, depicting a more concrete blueprint of how ByteDance will separate the new joint venture with its U.S. branch: the new TikTok U.S. joint venture would manage back-end technological operations, utilizing the algorithms from ByteDance and licensed by the Chinese government, while ByteDance would establish a new subsidiary in the U.S. responsible solely for advertising and e-commerce operations, obtaining revenue flows from the new joint venture. This type of joint-venture solution provides a possible template for future Chinese companies operating in the U.S.
However, such an arrangement hinges on several unsettled assumptions:
- How will such an “algorithm export” be implemented and retrained? Does it mark a shift from internal technology sharing within TikTok to a segregated, packaged algorithm transfer?
- Does the algorithm export authorization include maintenance or support services from China-based ByteDance engineers — and if so, would that constitute access to U.S. users’ data?
- Finally, can the new joint venture maintain its level of service relying solely on the authorized recommendation algorithms, without additional strategic, talent, or technological support from ByteDance?
With all these unclear assumptions, current comments on the deal blended into a chorus of optimism and speculation. Some Chinese finance media reporting on the TikTok deal recalled their articles to avoid potential factual error. But one thing is sure: a consensus framework is only the beginning of a roadmap to a final deal, and the bumpy broader China-U.S. trade and technology negotiations will continuingly shape it.
Chinese companies striving for new strategies of redesigning country-of-origin
As I have mentioned above, defending ByteDance/TikTok stands for the Chinese government as both a bargain stake and a case of showcasing its technology export regime. But how will this case impact Chinese technology companies that are undergoing a massive trend of “sailing out” (出海), i.e., the internationalization of platforms?
In the past 8 years, TikTok emerged as the pioneer flagship of globally expanding Chinese platform companies, and its business models, organizational structurals, and responses to geopolitical pressures provide a reference point for other Chinese firms seeking international expansion.
The core conundrum in front of TikTok and all other China-originated companies, is their country-of-origin identity. As PAFACA and other U.S. restrictive policies on Chinese firms suggest, such identity is interpreted as denoting potential Chinese government control over the firms. We don’t have space here to get into the long debate over the existence and extent of such control, but what is noticeable is that Chinese technological firms have been proactively redesigning their country of origin through core team relocation, data segregation, and target market discretion, which ultimately challenges the simple “country of origin” assumptions.
As the most prominent flagship of “sailing out” Chinese tech firms, ByteDance and TikTok represents a classic and mainstream “parallel platformization” model to differentiate their products across markets and jurisdictions: ByteDance created two parallel domestic and international apps, Douyin and TikTok. During the course of its international expansion and constant regulatory pressures from many governments, the two apps have gradually diverged through different data governance structures and teams. Many other short video and e-commerce platforms such as Temu/Pinduoduo and Kuaishou/Kwai also followed this path, segregating their target markets and making discretional platform products and data structures.
However, such parallel structure seemed not sufficient to address the mounting geopolitical pressures: the long TikTok saga including “Project Texas” and the current severance solution sends an alarming signal to many other Chinese platform companies: even if you can construct an internally segregated parallel systems like TikTok, the U.S. government still won’t be satisfied until it has absolute control over the algorithm and data.
Such distrust and regulatory pressures demonstrated through the TikTok case and previous chip embargos drive many Chinese tech platform firms, especially new AI start-ups, to develop new approaches of internationalizing their businesses by setting headquarters outside China (such as Zilliz, HeyGen), utilizing local cloud service providers, localizing data storage, engaging with local stakeholders, and even completely distancing from the mainland Chinese market.
In this sense, except for the founding teams and talent flows from mainland China, some new companies start to design their identity and presence as completely international, even circumventing their home country. This is a rather unthinkable phenomenon for the U.S.-originated companies: in light of heightened geopolitical pressures, firms have to avoid the founding team’s home country to get the ticket into broader international markets and access to advanced computing chips. However, such “Original Sin” identity issue is still existential for Chinese companies, even if the validity and feasibility of the “country of origin” principle is seriously undermined.
