US blacklists Chinese firm tied to engineer who stole Google secrets

A US blacklist or sanction places named companies and investors under government restrictions. The aim is to limit their ability to use US-linked property, financial services, or business relationships. In this case, the measures targeted an AI startup and the fund accused of financing it.

The article says the actions could freeze any US-based assets held by the entities. They could also impose severe international banking restrictions. MiraclePlus and three corporate affiliates were targeted because they supplied seed capital to Zhisuan.

The practical effect is wider than losing one investment opportunity. Banks and companies may avoid dealings with the listed entities, making fundraising, payments, and cross-border business more difficult. The article does not detail every legal restriction, but it presents the measures as potentially serious for the organizations’ international operations.

The main target was Shanghai Zhisuan Technology Company, an AI startup founded in 2023 by Ding Linwei. Ding was a former Google software engineer convicted of taking confidential AI trade secrets. The US Department of State announced enforcement measures against Zhisuan.

The action also covered MiraclePlus, Lu Qi’s incubator fund, and three corporate affiliates. MiraclePlus supplied seed capital to Zhisuan. The fund followed Y Combinator’s former China division, which Lu helped establish after Sam Altman recruited him to lead YC China.

Ding applied to MiraclePlus’s startup incubation pipeline on Zhisuan’s behalf. According to the State Department account reported by Bloomberg, he promoted and leveraged his inside knowledge of Google’s proprietary technology to secure backing. Thus, the connection involved both the alleged theft and the financing relationship.

Federal investigators said Ding Linwei exfiltrated more than 2,000 pages of Google’s confidential technical trade secrets. The material concerned proprietary artificial intelligence technology. The scale matters because it shows the alleged transfer was extensive rather than a single document or isolated incident.

Ding first moved the files to a personal cloud storage account. Investigators also found that he continued transferring internal files directly onto his personal computer. Those transfers happened less than two weeks before he formally resigned from Google in December 2023.

The case later produced a criminal conviction. Ding was convicted in January and sentenced in September to one year in prison. The article does not give the exact dates of each transfer, but it clearly places the final computer transfers shortly before his December 2023 resignation.

Ding’s use of Google’s confidential AI technology mattered because it connected Zhisuan’s funding effort to alleged theft from his former employer. Proprietary technology can give a company an advantage, so presenting inside knowledge may make a new venture appear more capable or advanced to investors.

According to the State Department account reported by Bloomberg, Ding applied to MiraclePlus’s incubation pipeline for Zhisuan. He actively promoted and leveraged his knowledge of Google’s proprietary technology while seeking backing. MiraclePlus then supplied seed capital to the startup.

That connection explains why the US action included not only Zhisuan but also its investor. The issue was not simply that Ding had worked at Google. It was that confidential AI information allegedly helped him attract financing for a competing technology venture, creating a direct link between the theft and the investment.

Blacklisting can damage a company’s ability to operate across borders. The article specifically says the targeted entities could lose access to US-based assets and face severe international banking restrictions. Those measures can affect how money is held, moved, and received.

The targets include Zhisuan, MiraclePlus, and three corporate affiliates. Because MiraclePlus provided seed capital to Zhisuan, restrictions may reach beyond the startup itself. The article does not identify particular banks or assets, but it says US-based holdings could be frozen.

International business could therefore become harder and more costly. Financial institutions and potential partners may avoid transactions involving listed entities, while fundraising and expansion may be disrupted. The article frames these as possible consequences, not confirmed outcomes for every target. It gives no further details about the duration or exact scope of the restrictions.

Trade-secret laws generally protect business information that has commercial value because it is not public and that its owner takes reasonable steps to keep secret. Protection can cover technical designs, software methods, and internal research. The rules aim to preserve fair competition and reward investment in confidential know-how.

An employee may access information for a job without owning it. Copying confidential files for personal use, or transferring them to another company, can violate duties to the employer and trade-secret law. In Ding’s case, investigators said he moved more than 2,000 pages of Google material to personal storage and a computer.

The alleged conduct became a criminal matter because the information was confidential and connected to a technology venture. Ding was convicted in January and sentenced in September to one year in prison. The article does not explain the specific legal charges, so this describes the general principle rather than the court’s exact reasoning.

AI models are software systems that learn patterns from examples and use those patterns to produce results, such as predictions, generated content, or decisions. Training systems provide the data handling, computing processes, and software needed to build those models. Infrastructure supports the models as they are developed and used.

The article describes Zhisuan as focused on specialized software infrastructure designed to accelerate the training of large-scale AI models. Faster or more efficient systems can help developers reduce delays and handle demanding model-development work. That gives the underlying technology strategic and commercial value.

The dispute shows why this area attracts intense competition. Google treated related AI technology as confidential, while Ding allegedly used inside knowledge when seeking funding for Zhisuan. The article does not describe a specific model or product. It does show that infrastructure and training capabilities were central to the startup’s business proposition and the alleged trade-secret case.