President Donald Trump-backed cryptocurrency venture World Liberty Financial is collaborating with Hong Kong-based AI platform WorldClaw, which provides access to dozens of artificial intelligence models developed by Chinese technology companies that the U.S. government has flagged over national security and technology concerns.
The relationship is drawing scrutiny because it places a business linked financially to the Trump family alongside Chinese AI developers at a time when the administration is seeking to strengthen the United States’ technological position against China and has imposed restrictions on several Chinese technology companies.
A Reuters review found that 43 of the 90 AI models available through WorldClaw were developed by Alibaba, Baidu, Z.ai and other Chinese companies. The platform also provides access to models from U.S. companies including OpenAI and Anthropic.
Register for the next Tekedia Mini-MBA.
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
There is no indication that the arrangement itself violates U.S. law. Chinese AI models are generally available for use by individuals and businesses in the United States, even when their developers face government restrictions. The issue is instead the apparent tension between the Trump administration’s national security policies toward Chinese technology and the commercial relationship involving a venture in which the Trump family holds a substantial ownership interest.
Sam Bresnick, a fellow at Georgetown University’s Center for Security and Emerging Technology, described the arrangement as potentially hypocritical given Washington’s efforts to counter China’s AI industry, according to Reuters.
The White House rejected the suggestion of a conflict.
“There are no conflicts of interest” in the relationship between World Liberty and WorldClaw, White House spokesperson Anna Kelly said, adding that Trump “only acts in the best interests of the American public.”
David Wachsman, a spokesman for World Liberty, said WorldClaw is independent and argued that major U.S. companies also provide access to AI models developed in different countries.
A WorldClaw spokesperson said the company helps U.S. AI companies reach international users and stressed that making an AI model available does not amount to endorsing its developer.
Chinese AI Models Dominate A Large Part Of WorldClaw’s Offering
The scale of Chinese participation on WorldClaw is central to the controversy. Of the 90 models listed on the platform, almost half were developed by Chinese companies. They include models from Alibaba and Baidu, as well as Z.ai, DeepSeek, and Moonshot.
The Trump administration has taken action against several of these companies. The U.S. Department of Defense has designated Alibaba and Baidu as Chinese military companies, a designation that restricts the Pentagon from doing business with them.
Z.ai, formerly known as Zhipu AI, is on the Commerce Department’s entity list, which imposes significant restrictions on access to U.S. technology. The U.S. government has said Z.ai contributes to China’s military modernization through AI research. The companies have disputed those allegations. Alibaba, for example, said it is not a Chinese military company and described its inclusion on the Pentagon’s list as “arbitrary and capricious.”
DeepSeek and Moonshot have also come under scrutiny from U.S. officials over allegations involving intellectual property. The companies have disputed such accusations.
The situation illustrates one of the complications in Washington’s effort to contain China’s technological rise. Government restrictions may limit certain commercial relationships, but they do not necessarily prevent companies and consumers from accessing Chinese-developed software and AI models.
World Liberty Has A Financial Interest In The Relationship
The commercial connection becomes more significant because World Liberty’s tokens can be used to pay for WorldClaw’s services. WorldClaw accepts World Liberty’s USD1 stablecoin, among other cryptocurrencies, as payment for access to AI models.
The Trump family owns 38% of World Liberty, according to Reuters. The family therefore has a financial interest in the broader use of the venture’s tokens.
USD1 is designed to maintain a value of $1 and is backed by traditional assets such as U.S. Treasury securities. World Liberty generates revenue from the stablecoin, including interest earned on the assets backing it, with the Trump family entitled to a share of that income.
This matters because the relationship does not appear to be simply a technology partnership. The use of World Liberty’s stablecoin on a platform offering Chinese AI models potentially creates a commercial link between the Trump family’s crypto interests and a business whose services include technology developed by companies targeted by the administration on national security grounds.
WorldClaw says it is independently owned and operated and is not managed or controlled by World Liberty or its affiliates. There are, however, personnel and promotional links between the two ventures.
Ryan Fang, World Liberty’s head of growth, serves as an external adviser to WorldClaw. The company said his work has focused on USD1 adoption, partnerships and expanding access to AI services.
Trump’s sons have also publicly promoted WorldClaw.
Donald Trump Jr. said in May that he would meet winners of a WorldClaw contest at Mar-a-Lago, while Eric Trump described the collaboration as “the future of finance.”
The Trump family has already generated more than $1.4 billion from World Liberty token sales, according to a Reuters report cited in the material, forming the largest component of the family’s reported $2.3 billion in cryptocurrency earnings.
Trump retains ownership of his interests through a trust while his sons, Donald Trump Jr. and Eric Trump, operate the Trump Organization.
The Security Question Extends Beyond Politics
The relationship also raises a separate question about the security implications of giving users access to Chinese AI models through a single aggregation platform. WorldClaw’s WorldRouter service acts as an intermediary between users and different AI models. The company’s website says the platform has more than 10,000 users and handles more than 50 million requested tasks a day.
WorldClaw is also developing AI-agent products capable of performing tasks such as ordering food and summarizing emails. That creates a potentially greater security concern than simply allowing users to ask a chatbot questions.
AI agents can be given access to applications, data, and digital services in order to perform tasks on a user’s behalf. If an AI system is compromised, manipulated or exposed to malicious instructions, the consequences can extend beyond an inaccurate answer to actions being carried out within connected systems.
Daniel Remler, a senior fellow at the Center for a New American Security and a former State Department policy adviser, said Chinese AI models could present risks including government monitoring, censorship and malicious code that could potentially compromise AI agents.
WorldClaw says on its website that it may share user inputs with the companies providing the underlying models. That raises an additional privacy consideration for businesses or individuals using the platform to process sensitive information.
The company said it applies privacy and security safeguards and takes security seriously.
The Broader U.S.-China AI Race Complicates The Picture
The WorldClaw relationship comes at a particularly sensitive moment in the global AI competition.
Washington has sought to limit China’s access to advanced semiconductors and other technologies while encouraging U.S. companies to maintain leadership in artificial intelligence. At the same time, Chinese developers have produced competitive models, often at lower costs, making them difficult to exclude entirely from the global AI ecosystem.
That creates a tension between national security policy and commercial reality.
Peter Jeydel, an attorney specializing in sanctions and trade controls at Troutman Pepper Locke, said the relationship could be viewed as inconsistent with a China-hawk approach but also fit with a more business-oriented interpretation of Trump’s policies.
The contradiction is likely to become more pronounced as Chinese AI models become more capable and cheaper to operate.
For U.S. companies, investors and consumers, the question is not simply whether a Chinese AI model can be accessed, but what data is sent to it, where that data is processed, who controls the underlying infrastructure, and what an AI system can do once it is connected to other applications.
WorldClaw’s business model puts those questions directly alongside the political and financial interests surrounding World Liberty.
Meanwhile, the arrangement highlights a difficult policy dilemma for the Trump administration. Washington can restrict Chinese companies, limit technology transfers and warn about national security risks, but Chinese AI models can still gain international users when they are competitive, inexpensive and distributed through global platforms.
The World Liberty and WorldClaw relationship therefore sits at the intersection of three rapidly expanding industries: cryptocurrency, artificial intelligence and cloud-based digital services.