The Pentagon’s Consumer-Tech Blacklist Has a Point, and a Problem

The most important thing about the Pentagon’s new China list is not that it names defense companies. It is that it names companies most Americans would not think of as defense companies at all.

On June 8, 2026, the Pentagon added Alibaba, BYD and Baidu, along with other Chinese firms, to its list of “Chinese military companies” operating directly or indirectly in the United States, according to the Associated Press¹. Alibaba is the e-commerce and cloud giant. BYD is China’s electric-vehicle and battery champion. Baidu is the search, artificial-intelligence and autonomous-driving company behind the Ernie chatbot. None looks like a missile maker. That is exactly why the move matters.

I think Washington is right to treat these firms as defense-supply-chain risks. I also think it should stop pretending the public evidence is equally strong for all of them. The right conclusion is not “blacklist every big Chinese technology firm.” It is narrower and sharper: the Pentagon has a sound reason to keep these platforms out of U.S. defense procurement, but Congress and the administration need a more transparent evidentiary process before the label becomes a broader decoupling weapon.

Start with the legal machinery. The Pentagon 1260H list takes its name from Section 1260H of the 2021 National Defense Authorization Act, which requires the Defense Department to identify Chinese military companies operating directly or indirectly in the United States. The statutory definition is broad by design. It covers not only firms owned or controlled by the People’s Liberation Army, or PLA, but also commercial firms affiliated with China’s Ministry of Industry and Information Technology, known as MIIT, China’s state asset regulator, the Ministry of State Security, SASTIND, which oversees defense science and industry, or firms identified as “military-civil fusion” contributors to China’s defense industrial base, according to 10 U.S.C. § 113 notes setting out Section 1260H². Military-civil fusion means Beijing’s strategy of blurring civilian and military innovation so that commercial technologies, factories, labs and data systems can help modernize the PLA.

That breadth is not a drafting accident. It is the point. A modern military does not just buy tanks. It needs batteries, cloud computing, mapping, logistics software, artificial intelligence, robotics, sensors and data centers. Defense procurement, in plain English, is government buying for military use. In that setting, a civilian-looking platform can become part of military capacity long before it appears on a weapons invoice.

The strongest evidence for this logic comes from China’s AI procurement ecosystem. Georgetown’s Center for Security and Emerging Technology reviewed 2,857 PLA AI-related defense contract award notices from January 2023 through December 2024 and found that, while legacy defense firms still mattered, nontraditional vendors and research institutions played a consequential role; nontraditional vendors won 764 contracts, according to CSET’s report³. That finding does not prove Alibaba, BYD or Baidu did anything wrong. It does prove the old mental model, civilian company over here and military supplier over there, is too simple for China’s technology economy.

Alibaba is the easiest case for Washington. The Pentagon reportedly cited Alibaba’s affiliation with MIIT, and AP reported that Alibaba denied being a Chinese military company or part of any military-civil fusion strategy. But the public record contains more than MIIT affiliation. A 2023 Senate letter alleged that Alibaba Cloud signed a 2017 military-civilian fusion cloud cooperation agreement with Norinco, a major Chinese state-owned defense company, and a cooperation agreement with China’s National University of Defense Technology; the same letter said an Alibaba Cloud executive participated in a PLA Academy of Military Sciences Military Big Data Forum about moving e-commerce technologies into national defense, according to Senator Bill Hagerty’s office⁴.

Those are allegations from lawmakers, not a court-tested record. Still, cloud infrastructure is exactly the kind of dual-use layer that belongs under defense scrutiny. A cloud provider can supply storage, compute power, data analytics and AI tooling. If a U.S. defense contractor were quietly dependent on a firm with credible links to Chinese defense research or industrial planning, the Pentagon would be negligent not to care.

Baidu is more ambiguous. The company has denied that it is a military company, and the best-known public incident cuts both ways. In January 2024, Baidu denied a report tying its Ernie chatbot to Chinese military research after an academic paper from PLA Information Engineering University described using Ernie to generate simulated military response plans; Baidu said it had no business collaboration with the authors and that Ernie was available to the general public, according to AP’s account at the time⁵. That denial matters. A public chatbot being tested by military researchers is not the same thing as Baidu building a PLA system.

But it also points to why Washington is nervous. Frontier AI models are not like ordinary consumer apps. They can be adapted for decision support, simulation, targeting workflows, cyber operations, logistics and intelligence analysis. If the PLA is experimenting with commercial AI tools, and if Chinese law and industrial policy make clean separation between civilian champions and state security needs hard to trust, Baidu belongs in a defense-procurement risk file even if the public evidence does not yet justify treating it like a sanctioned arms dealer.

BYD is the thinnest public case, and that should bother people. The Pentagon cited BYD’s affiliation with MIIT, according to AP, which also described BYD as dominant in the global electric vehicle market. EVs and batteries are plainly dual-use technologies: militaries need mobile power, unmanned vehicles, resilient logistics, fleet electrification and battery supply chains. A company that can make cars at vast scale can also matter to industrial mobilization. But public reporting so far has not shown the kind of specific BYD-to-PLA evidence that exists, at least in allegation form, for Alibaba Cloud.

That distinction should not be waved away. National-security policy gets worse when every case is argued at the level of vibes. If MIIT affiliation plus a dual-use sector is enough, then nearly every major Chinese company in autos, drones, cloud, chips, robotics, batteries or industrial software can be called military-linked. Some of those labels will be right. Some will be lazy. The difference matters for investors, suppliers and the credibility of U.S. policy.

The list has real bite, even if it is not a full sanctions regime. A 1260H designation by itself is not the same as an Office of Foreign Assets Control blocking sanction. It does not automatically freeze assets, ban all U.S. dealings or impose a general export-control prohibition. Legal advisers have long described the list as reputationally powerful but historically limited in immediate legal effect, as Akin Gump noted⁶.

That is changing. Section 805 of the fiscal 2024 defense law bars the Defense Department, effective June 30, 2026, from entering into, renewing or extending contracts with companies on the 1260H list or entities they control; a broader goods-and-services restriction tied to listed entities phases in on June 30, 2027, according to Crowell & Moring’s analysis⁷. For defense contractors, that means the list becomes a supply-chain screening problem. For investors, it becomes a warning label. For U.S. companies using Chinese cloud, battery, robotics or vehicle technology near sensitive government work, it becomes a reason to unwind exposure before a lawyer or contracting officer forces the issue.

This is where I part ways with the most aggressive version of Washington’s argument. Keeping Alibaba, BYD and Baidu away from U.S. defense procurement is prudent. Using the same designation as a public proof of military culpability, a delisting argument and a commercial no-contact order is a different move. AP reported that the House Select Committee on the Chinese Communist Party called the list a warning and said publicly traded listed companies should be delisted and that no American company should do business with them. That leap goes beyond the evidence the public has seen.

The counterargument is serious: China’s military-civil fusion system is opaque, and waiting for courtroom-grade proof may mean acting only after U.S. technology, capital and supply chains have already strengthened the PLA. I accept that risk. A defense supply chain has to be more paranoid than a consumer marketplace. The Pentagon should not need to prove a direct weapons contract before deciding that a Chinese AI or battery platform is too risky for military procurement.

But opacity cuts both ways. Beijing can hide military ties, and Washington can hide weak evidence behind classified process. The answer is not to make the list toothless. The answer is to create tiers. A procurement-exclusion tier can rest on risk indicators such as MIIT affiliation, sensitive technology, defense-industrial partnerships and ownership links. A broader investment or commercial restriction should require more specific public findings: contracts, controlled subsidiaries, military research programs, procurement-platform activity, data-sharing obligations or defense-conglomerate partnerships.

There is also a retaliation channel. China has already used tools such as export controls, rare-earth restrictions and unreliable-entity measures in disputes with the United States, and legal analysts have described Beijing’s 2025 rare-earth and lithium-battery-related controls as part of the same leverage toolkit, according to Jones Day⁸ and Holland & Knight⁹. That does not mean Washington should back down. It means the United States should reserve its most escalatory labels for cases it can defend with evidence.

My forecast is simple: the 1260H list will become a de facto map for defense-sector decoupling, not a standalone sanctions list. Watch three indicators next: whether DoD publishes more detailed unclassified justifications for Alibaba, BYD and Baidu; whether major defense contractors disclose supply-chain changes before the June 30, 2027 indirect-procurement deadline; and whether Congress tries to convert 1260H status into mandatory delisting or investment bans. If the third happens before the first, Washington will have turned a defensible defense screen into an opaque economic weapon.