Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Some links on this page are affiliate links: if you buy through them we may earn a commission, at no extra cost to you.
On February 27, 2025, Microsoft President Brad Smith urged the Trump administration to revise the Biden-era AI Diffusion Rule. Microsoft argued that limits on advanced AI chips for U.S. allies could push them toward Chinese technology. It was not asking for unrestricted chip sales to China: Smith backed controls on adversaries, secure data centers and safeguards against diversion. The policy later changed, but not through a simple, completed repeal.
What Microsoft wanted changed
Smith’s proposal focused on the rule’s restrictions for countries in its middle tier. Microsoft wanted the administration to remove or substantially reduce quantitative limits on advanced chips for trusted allies and partners, and to stop treating important partners as a category requiring those caps. It also called for a simpler framework and faster Commerce Department licensing and enforcement.
Microsoft said security measures should remain: requirements for certified data centers, controls intended to keep advanced technology from adversaries, and stronger protections against chips or computing capacity being diverted to prohibited users. The distinction matters. The request was to loosen limits affecting many allies, not to allow advanced chips to flow freely to China.
Smith described the existing framework as a 41-page rule. His argument was that broad country quotas could constrain legitimate U.S. cloud infrastructure projects even when the operator and intended customers were not security threats. Microsoft’s statement named Switzerland, Poland, Greece, Singapore, India, Indonesia, Israel, the United Arab Emirates and Saudi Arabia as partners affected by the middle-tier approach.
#1 Best Overall
- ✅Powered by 26 Tera-Operations Per Second (TOPS) Hailo-8 AI Processor. 2.5W typical power consumption
- ✅Scalable, enabling simultaneous processing of multi-streams & multi-models
- ✅Enabling real-time, low latency and high-efficiency AI inferencing on the edge devices
- ✅Supports TensorFlow, TensorFlow Lite, ONNX, Keras, Pytorch frameworks
- ✅Supports Linux and Windows. Supports the temperature range of -40°C to 85°C
What the AI Diffusion Rule did
The Commerce Department’s Bureau of Industry and Security published the Framework for Artificial Intelligence Diffusion on January 15, 2025. It revised controls on advanced-computing chips, covered certain advanced closed-weight AI-model weights, and set up licensing and authorization mechanisms, including an updated pathway for validated data-center end users.
Its broad design divided countries into three groups:
- Tier 1: A limited set of highly trusted countries received the most favorable treatment and generally avoided the principal quantitative limits.
- Tier 2: Many other countries, including U.S. allies and partners, faced quantitative limits and additional conditions. This was not a blanket ban on buying chips; licensing, ceilings, exceptions and authorization arrangements affected what was permitted.
- Tier 3: Countries of concern, including China, Russia and Iran, faced the strictest restrictions on covered technologies.
The rule took effect on January 13, 2025, before its publication, and set May 15, 2025, as the general compliance date. Certain provisions had a delayed compliance date of January 15, 2026. Its stated rationale was that advanced AI could support military applications, cyber operations, surveillance and chemical or biological weapons, while secure access could also deliver substantial economic and social benefits.
The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Rank #2
- High-Performance AI Processing: The MX3 is designed to handle the most demanding AI computer vision workloads, delivering exceptional performance and efficiency.
- Flexible Integration: The MX3 can be easily integrated into your existing systems via its M.2 M-key form factor and support for Linux operating systems.
- Energy Efficient: The MX3 is designed to provide high performance while minimizing power consumption.
- Comprehensive Software Development Kit (SDK): The MX3 is supported by a comprehensive SDK that simplifies development and deployment.
- Hardware compatability: The MX3 is compatible with the PCI-SIG M.2 M-key 2280 Specification. It can be used with the Raspberry Pi 5 with a M-key 2280 HAT.
Why Microsoft said the limits could benefit China
Microsoft’s case was about where countries would build and buy AI infrastructure. If a U.S. provider could not be confident it could install enough computing capacity in a country, Microsoft argued, local customers might hesitate to rely on its services. Some could instead choose Chinese infrastructure or AI services. If that happened at scale, China could gain customers, commercial footholds and long-term influence in markets where American companies might otherwise set the standard.
Smith compared the risk with China’s expansion of 5G technology: he warned that a gap in U.S. availability could create room for Chinese providers. That was Microsoft’s strategic forecast, not proof that affected countries would switch suppliers or that the rule would automatically transfer business to China.
The mechanics are more complicated than a country label. A data center’s risk depends on who owns and operates it, who can access its systems, and which customers use its capacity. Cloud access also differs from direct chip ownership: a customer may use computing remotely without importing a GPU itself. Microsoft’s argument was that broad national quotas could constrain American providers before those distinctions were fully reflected in practical access.
Microsoft had a business stake
Microsoft’s position also aligned with its commercial plans. The company was building out Azure and AI infrastructure internationally, where local data centers can help serve customers with lower latency and meet local operational needs. Smith said Microsoft planned to spend $80 billion on AI infrastructure in 2025, with more than half in the United States. He also cited a planned $700 million expansion of Microsoft’s data-center infrastructure in Poland.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Clear out junk files and repair common Windows errors3Fix the driver behind crashes, sound loss and screen glitchesThat interest does not disprove the national-security argument, but it belongs in the analysis: broader access to overseas markets could help Microsoft sell cloud and AI services as well as support U.S. technology adoption. The policy question was therefore not simply whether export controls were good or bad. It was whether this particular mix of quotas and safeguards would protect U.S. security without undermining American providers’ ability to compete abroad.
The case for caution
Advanced chips can power large-scale AI training and inference. The resulting systems may have military, intelligence, cyber, surveillance or weapons-related uses. Even a data center located in an allied country can pose a diversion risk if its owners, operators, customers or supply chains are not adequately controlled. Model weights and remote computing access raise additional questions beyond the shipment of a chip across a border.
Rank #4
- ✅Powered by 26 Tera-Operations Per Second (TOPS) Hailo-8 AI Processor. 2.5W typical power consumption
- ✅Scalable, enabling simultaneous processing of multi-streams & multi-models
- ✅Enabling real-time, low latency and high-efficiency AI inferencing on the edge devices
- ✅Supports TensorFlow, TensorFlow Lite, ONNX, Keras, Pytorch frameworks
- ✅Supports Linux and Windows. Supports the temperature range of -40°C to 85°C
Supporters of a restrictive approach could argue that measurable ceilings and licensing rules are easier to apply than relying only on trust in individual companies or customers. Loosening them could expand the number of facilities requiring monitoring and create more routes for sensitive computing capacity to reach prohibited users. The disagreement was over the best way to secure U.S. leadership: broad limits to slow access, or wider availability of U.S. technology under targeted security conditions.
Microsoft was not alone, but industry opinion was not uniform
Nvidia criticized the rule as a threat to U.S. technological competitiveness, arguing that the restrictions could weaken the innovation ecosystem that helped the United States lead. Amazon CEO Andy Jassy separately raised concern that restrictions on friendly countries could cost U.S. providers business and relationships. These documented objections show a broader technology-industry pushback against the rule’s reach; they do not establish that every company opposed export controls or shared Microsoft’s precise proposal.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
What happened after the appeal
- January 15, 2025: BIS published the AI Diffusion Rule.
- February 27, 2025: Brad Smith made Microsoft’s case for revising it.
- May 13, 2025: Commerce announced that it would not enforce the rule and said it planned to rescind and replace it. The department also announced additional semiconductor-related controls and anti-diversion guidance.
- May 12, 2026: The Government Accountability Office concluded that Commerce’s non-enforcement announcement qualified as a rule for Congressional Review Act purposes. GAO noted that formal rescission had not been completed: the original rule remained technically in the Code of Federal Regulations, although Commerce was not enforcing it.
So it is imprecise to say that Trump simply repealed the rule, or that the United States abandoned AI-chip export controls. Commerce stopped enforcing this framework and announced plans to replace it, while taking other steps on semiconductor controls. The distinction between policy in practice and a rule’s formal legal status is important for understanding what changed.
Best Value
- DEEPX DX-M1M NPU: Powered by the DEEPX DX-M1M neural processing unit, purpose-built for efficient on-device AI inference workloads.
- COMPACT M.2 2242 FORM FACTOR: Fits the standard M.2 2242 slot, making it easy to integrate into embedded systems, edge devices, and compact computing platforms.
- EDGE AI ACCELERATION: Designed to accelerate deep learning inference at the edge, enabling real-time AI applications without relying on cloud connectivity.
- RADXA AICORE MODULE: The Radxa AICore DX-M1M delivers a plug-and-play AI compute solution ideal for robotics, smart cameras, and industrial automation.
- WARRANTY AND ORIGIN: Backed by a 1-year manufacturer warranty and crafted with quality components for reliable long-term performance in demanding environments.
In retrospect, Microsoft’s warning captures the underlying trade-off: restrictions meant to limit China’s access can also affect how readily allies can obtain American AI infrastructure. Whether broader access would have produced the shift toward Chinese providers Microsoft feared remains a forecast, not a settled outcome. The lasting policy question is how to protect sensitive technology without making U.S. alternatives harder for partners to use.
Sources: Microsoft, February 27, 2025; Federal Register rule; Commerce Department, May 13, 2025; GAO decision, May 12, 2026.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
Recommended Free Tools
