Published: · Last updated: with original publication · Region: Global · Category: cyber

Source-reported · Independent verification status is not recorded for this article; claims remain attributed to their sources.

Alibaba’s Reported $4 Billion Nvidia Chip Route via Singapore and Malaysia Raises Export Control Questions

Alibaba has reportedly accessed more than $4 billion in advanced Nvidia chips through a Singapore-based firm and Malaysian data centers, despite U.S. export restrictions, in a case that highlights how easily high-end hardware can move through third countries and cloud facilities.

A report that Alibaba obtained billions of dollars’ worth of Nvidia chips through intermediaries in Singapore and Malaysia is fuelling doubts about how tightly U.S. export controls can limit China’s access to advanced AI hardware.

According to an investigation attributed to the New York Times, Alibaba has accessed more than $4 billion worth of Nvidia’s advanced chips through a Singapore‑based company called Megaspeed and data centers in Malaysia. The report suggests this channel may have allowed Alibaba to keep expanding its AI computing capacity despite U.S. efforts to restrict direct exports of high‑end chips to China.

The same reporting says Alibaba co‑founded Zhejiang Lab, a research institution with links to China’s military that works on artificial intelligence projects involving cyberattacks and war simulations. This connection has prompted concern that U.S.‑origin chips could be supporting not only commercial AI but also research with potential military applications. The investigation notes, however, that it’s unclear whether Megaspeed’s activities or the Malaysian deployments directly supported Zhejiang Lab’s specific programs.

For U.S. officials, the case points to a core enforcement problem. Export controls usually focus on shipments from U.S. companies to named entities in restricted countries. Once chips are sold into global markets, they can be routed through third‑country firms or installed in foreign data centers, where access is then sold as a cloud service. That makes it harder to track who is effectively using the computing power, even if the hardware itself never physically re‑enters a restricted jurisdiction.

Cloud infrastructure compounds the issue. By placing Nvidia chips in data centers outside China and offering remote access, intermediaries can keep the equipment within jurisdictions that appear to comply with formal export rules while still making high‑performance computing available to Chinese users.

For engineers and AI developers in China, such arrangements provide continued access—albeit with more legal and financial complexity—to the chips used to train large models. For U.S. and allied governments, they represent a pathway for powerful hardware to support organizations that sit close to China’s defense sector, blurring already thin lines between civilian and military use in the tech ecosystem.

There are commercial risks as well. If U.S. authorities decide that certain intermediaries or data centers are undermining export policy, they could face sanctions or restrictions. That would disrupt Chinese customers but also other clients sharing the same facilities.

Strategically, the reported Alibaba–Megaspeed–Malaysia route illustrates how hard it is for export controls to contain widely used, high‑value components once they enter complex global supply chains. The central challenge shifts from blocking shipments at the border to tracing ownership structures, service contracts, and usage patterns across jurisdictions.

Key developments to watch include any formal moves by U.S. regulators to investigate Megaspeed or related entities, public statements from Malaysia or Singapore on their export and re‑export rules, and changes in Nvidia’s compliance policies in response to the reporting. New U.S. measures aimed at regulating access to cloud‑based AI accelerators, in addition to physical chip exports, would indicate an attempt to close some of the gaps exposed by this case.

Sources