In May 2025, as the global AI race reignites, the latest actions of the U.S. government can be described as a decisive move, with full force. The U.S. Department of Commerce's Bureau of Industry and Security (BIS) quietly released three 'guiding opinions,' which, although not officially enacted into law, send very clear political signals and industrial intentions: the 'technological Cold War' in the AI field is entering a more direct and aggressive phase.


The core goal of this initiative is undoubtedly to build a global technological blockade around 'Huawei Ascend chips,' forcing global companies to choose sides between the 'U.S. technology stack' and the 'Chinese technology route.'



One, 'guiding opinions' do not equal laws, but are sufficient to change the global game rules.


It should be made clear that the 'guiding opinions' issued by BIS are not legal texts within the U.S. administrative system but rather statements by law enforcement agencies regarding the interpretation and enforcement tendencies of existing regulations. They do not require public hearings and are not subject to the constraints of the Federal Administrative Procedure Act, but their 'soft deterrence' power should not be underestimated.


Particularly in these three opinions, the first one has received widespread attention — 'Using Huawei Ascend chips anywhere in the world is a violation of U.S. export controls.' Although this statement still falls under policy 'suggestions,' it basically establishes the U.S.'s regulatory boundaries for the global AI ecosystem: no matter where you are or whether you directly import U.S. equipment, as long as you use Huawei H series chips, you may be considered in violation of export controls.


In Mlion.ai's international policy monitoring system, this subtle change in rule hierarchy has been marked as a 'high-risk event,' indicating that the policy's potential impact on market dynamics has reached a red alert level.



Two, it appears to be a technological issue, but in essence, it is a struggle for sovereignty and rules.


Interestingly, this ban under the guise of 'chip usage' is not fundamentally about the product's inherent risks but about ensuring that the U.S. maintains its strategic dominance in the AI ecosystem.


The 'AI diffusion rules' established by the Biden administration were just about to be enacted when the Trump administration decisively withdrew them, citing simple reasons: too many restrictions that even U.S. companies couldn't tolerate. In other words, the U.S. strategy has shifted to: internal relaxation, external blockade, high pressure on China, and external inclusivity.


BIS Deputy Secretary Kessler's public statement further pointed out bluntly: 'We no longer enforce the unreasonable AI regulations of the Biden era; what we want is a bolder, more inclusive, but clearly confrontational AI strategy.'


Mlion.ai has already modeled warnings in this month's AI policy trend prediction model: the Trump administration is attempting to create a 'rules lock chain' in the AI field through a tripartite approach of export controls, global technology alliances, and 'informal guiding opinions.'



Three, direct strikes on China and indirect deterrence against third countries.


This statement of 'global ban on Huawei Ascend chips' superficially targets China, but the real strategic focus is on 'blocking third countries' cooperation with China.'


Currently, Huawei's Ascend H series chips are primarily used in the domestic market and have not yet formed large-scale exports. However, the U.S. has preemptively set a 'red line,' defining their use as 'violations,' clearly to establish rules and deterrence.


Its implicit threat is very clear:



'You may not use it now, but once you do, it's like crossing a line.'



This poses a significant pressure to choose sides for global AI companies, especially emerging market countries with tight resources but eager for technological breakthroughs.


In Mlion.ai's chip supply and demand dynamics panel, related countries such as Saudi Arabia, the UAE, and Indonesia are accelerating cooperation with U.S. companies like NVIDIA, Microsoft, and Amazon, and the evolution path of their AI infrastructure procurement decisions can be accurately tracked to provide investors with forward-looking intelligence support.



Four, the dual-wheel strategy of 'pushing U.S. technology + blocking Chinese technology' has taken shape.


This set of three guiding opinions is not an isolated policy, but a comprehensive launch of the 'AI Geostrategy.'


On one hand, the U.S. government is loudly promoting the 'U.S. technology stack': GPUs, IaaS cloud services, large model infrastructure, API services, etc., fully led by U.S. companies to ensure complete dependency from chips to algorithms; on the other hand, it is blocking Chinese technology penetration through rules, guiding opinions, export restrictions, and building a 'U.S.-style technology ally circle.'


Correspondingly, Mlion.ai's technology dependency analysis tool has begun to automatically identify the model training paths, cloud computing procurement sources, and underlying chip origins used by AI companies in various countries, helping users quickly assess the 'U.S.-China technology exposure' of a certain project or company to gauge its risk level of facing policy shocks.



Five, the law is not yet formed, but rules are already taking precedence — this is the most frightening aspect.


What is most alarming is that these three 'guiding opinions' are not based on the new bill or the revised Export Administration Regulations (EAR); they are merely BIS's interpretation and expanded application of existing rules. In other words, it is a strategy of 'execute first, legislate later.'


Especially the first clause regarding Ascend chips, which is likely to be expanded through FDPR (Foreign Direct Product Rule) and the 'knowing assistance ban' under EAR Section 736.2(b)(10), reinforcing the enforcement standard of 'knowing violation equals violation.'


This is especially dangerous for Chinese AI companies, as many projects may still contain unreleased U.S. technology residues when deploying domestic computing power. Mlion.ai's code and technology origin tracing tool is particularly crucial in such scenarios, assisting companies in identifying hidden U.S. technology components in their system architecture and avoiding compliance traps of 'knowingly violating.'



Six, cloud services have become the focus of control in the next phase.


In addition to chip use, the second 'guiding opinion' clearly states: U.S. AI chips used for training Chinese models will face risks. Although this currently lacks a clear legal basis, it is clearly laying the groundwork for the next phase of 'cloud computing control.'


This trend was evident during the Biden administration but failed to pass legislation at that time. Looking back, Trump's approach is more pragmatic — no legislation, directly regulating the market.


Therefore, in the short term, we are likely to see U.S. cloud service providers such as AWS, Azure, and GCP strengthen scrutiny of overseas clients' model training needs and set more 'red flag rules' regarding customer data flow and model weight delivery.


Mlion.ai's AI model training path supervision module can assist users in tracking the hardware sources, training locations, and computing power rental records during the model training process, identifying potential cross-border regulatory risks, and ensuring the compliance and transparency of technology use.



Seven, the true starting point of the AI Cold War era.


'Guiding opinions' are not the end, but the beginning.


These three seemingly inconspicuous policy adjustments mark the U.S. systematically positioning AI chips, model training, computing power scheduling, and cloud services as strategic resources. The new regulations represented by the 'global ban on Huawei Ascend chips' also mean that the global AI industry is being forcibly pushed into a more closed and oppositional framework.


This also brings a practical issue:



For global AI companies, how to accurately identify legal red lines, strategic risks, and market opportunities in an increasingly complex compliance environment?



The global policy event tracking, export control impact assessment, model origin auditing, and compliance research modules provided by Mlion.ai are digital solutions designed for this need. It is not merely a news platform, but a complete set of AI strategic intelligence and compliance tools that help users see the direction in chaos and seize opportunities amidst changing rules.



Conclusion:


The release of these three 'guiding opinions' by the U.S. not only marks the start of a new round of technology competition between China and the U.S., but is also likely to influence the future global landscape of AI infrastructure. Whoever can build a trustworthy, compliant, and independent technology chain first may gain the upper hand in the next industrial revolution.


And you, are you ready to find your place in this new era?



Disclaimer: The above content is for information sharing only and does not constitute any investment advice. Please make rational judgments based on your risk tolerance.