A long-trailed resolution to Washington’s stand-off with TikTok appears closer, with US President Donald Trump and China’s President Xi Jinping expected to discuss terms on Friday after negotiators outlined a “framework” agreement this week.
Reports suggest TikTok’s US operations could be sold to a consortium of American firms, potentially including Oracle, Andreessen Horowitz and Silver Lake. The arrangement would affect the app’s 170 million US users and mark what some analysts describe as a rare breakthrough in US-China economic relations.
At the core of the talks is TikTok’s recommendation algorithm — the closely guarded system that drives the app’s virality by serving content aligned to user behaviour.
ByteDance, TikTok’s Chinese owner, has long refused to sell or transfer this technology outright. In a notable shift, however, China’s top cyber regulator has signalled that ByteDance may license the algorithm and other intellectual property to a US owner, rather than relinquish it. That would be a departure from Beijing’s previous red lines on the export of sensitive code.
A licensed, rather than transferred, algorithm would likely result in a US-specific TikTok built on a “stripped-down” software stack, according to computing experts. Limited access could still illuminate how engagement, moderation and ad targeting are engineered — the mechanics that underpin TikTok’s commercial heft — but would preserve what ByteDance sees as its crown jewels. The trade-off could be visible to users: a more domestically ring-fenced feed and potentially less diverse content discovery than on the global app.
US Treasury Secretary Scott Bessent, who is leading the American side, has said the user experience would remain unchanged, albeit with “Chinese characteristics” — a phrase associated with the Chinese Communist Party’s preferred framing of policy. That formulation goes to the heart of US concerns: who ultimately influences the product and who can access American user data. Those issues drove legislation signed last year requiring TikTok to cede control of its US arm or face a ban.
Politics complicate the path to closure. Mr Trump, who previously backed a ban, has since credited the platform with energising younger voters during the 2024 election and now favours a deal. Even so, any sale would need to satisfy Congress. Lawmakers in both parties have warned that a licence arrangement may not meet the statute’s requirement for “full separation” from a foreign adversary’s control. Representative John Moolenaar has voiced concern that the current framework could leave scope for Chinese state influence. Legal analysts similarly note that a licence, by definition, falls short of an outright divestment.
Practical hurdles are substantial. It remains unclear how a US-owned TikTok would interoperate with the rest of the platform still controlled by ByteDance — from code maintenance and feature roll-outs to trust-and-safety policies and advertising marketplaces. ByteDance, a private company, would also need board approval. Beijing’s export controls on advanced technologies, including algorithms, add another layer of authorisation. Large cross-border technology deals typically take months, if not years, to complete.
For Beijing, the prospective arrangement offers several advantages. First, licensing keeps the algorithm out of US ownership while preventing an American ban that would cut the firm off from its most lucrative market. Industry estimates suggest average revenue per user in the US is multiple times that of other regions, and the US could account for a substantial share of ByteDance’s turnover. Tech publication The Information has estimated ByteDance’s 2024 global revenue at about $39bn, with TikTok contributing roughly $30bn.
Secondly, a TikTok resolution could establish a broader template for Chinese firms to deploy critical technologies in the US through licensing and joint structures. Analysts point to sectors such as batteries and rare earth supply chains, where Chinese capabilities remain central to global manufacturing. A workable “TikTok model” could, in theory, ease the entry of companies such as BYD or CATL into US markets under arrangements that address security concerns without ceding core intellectual property.
Thirdly, progress on TikTok may create negotiating space in wider US-China trade talks. China remains a major buyer of American agricultural goods; the US is a key destination for Chinese exports. Both sides maintain high tariffs and export controls, including on inputs where China dominates processing capacity. A contained solution on a politically salient platform could reduce friction, at least tactically, while larger disputes over semiconductors, data governance and investment screening continue.
For Washington, the gains would be tangible but conditional. A domesticated TikTok, firewalled data, and verifiable oversight could answer immediate national-security objections, preserve a platform popular with American users and creators, and avoid the political and legal costs of an outright ban. Yet any licence-based architecture would require sustained monitoring and clear compliance triggers if obligations are breached.
The outcome now depends on whether negotiators can translate a high-level framework into a legally robust, technically workable separation that satisfies lawmakers, regulators and the two governments. Even if agreed, such a deal would operate under persistent scrutiny. The app may look the same to users; behind the scenes it would run on borrowed code, segregated data and a layer of political trust that could be tested by events well beyond TikTok’s control.
