The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
The Trump administration pledged in April 2026 to work with U.S. AI companies to identify and defend against alleged campaigns to extract capabilities from American models, and to seek ways to punish offenders. On September 8, CISA, the NSA, and the FBI named six China-based firms in an advisory alleging large-scale activity. The agencies’ claims are contested; the public advisory is not an independent adjudication, and the sources available through October 4, 2026, do not document company-specific punishments arising from the pledge.
What the administration said it would do
On April 23, 2026, White House science and technology adviser Michael Kratsios said the administration would coordinate with U.S. AI companies to identify alleged industrial-scale distillation campaigns, strengthen defenses, and find ways to punish offenders. That was a pledge and policy direction, not evidence that a particular company had already been sanctioned or otherwise punished. The Associated Press reported the announcement.
There was also legislative activity, but it should not be confused with an enacted law. The House Foreign Affairs Committee gave unanimous bipartisan support to a proposed bill that would establish a process to identify foreign actors extracting key technical features from closed-source U.S.-owned models and potentially punish them, including through sanctions. Committee support does not itself create new sanctions authority.
Recommended Free Tools
Which companies U.S. agencies named
In a September 8, 2026 advisory issued with the NSA and FBI, CISA named DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI. The agencies alleged that these firms had extracted “billions of tokens across millions of exchanges/requests” from U.S. frontier models, including Claude, GPT, Gemini, and Grok variants, since at least late 2024. These scale figures and allegations are the agencies’ characterization, not independently verified totals. Read the CISA advisory.
#1 Best Overall
- EVOLUTION AMD RYZEN AI MAX+ 395 MINI PC - GMKtec EVO-X2 is the next evolution in AI mini PC Ryzen Strix Halo series. Thanks to AMD Simultaneous Multithreading (SMT) the core-count is effectively doubled, to 32 threads. Ryzen AI Max+ 395 has 64 MB of L3 cache and can boost up to 5.1 GHz, depending on the workload. The Ryzen AI Max+ 395 is currently rated as the "most powerful x86 APU" on the market for AI computing.
- AI NPU with XDNA 2 ARCHITECTURE - Powered by 16 “Zen 5” CPU cores, 50+ peak AI TOPS XDNA 2 NPU and a truly massive integrated GPU driven by 40 AMD RDNA 3.5 CUs, the Ryzen AI MAX+ 395 is a transformative upgrade and delivers a significant performance boost over the competition. The Ryzen AI Max+ 395 excels in consumer AI workloads like the llama.cpp-powered application: LM Studio. Shaping up to be the must-have app for client LLM workloads, LM Studio allows users to locally run the latest language model without any technical knowledge required and unleash their creativity and productivity.
- AMD RADEON 8090S iGPU GAMING PC - The AMD Radeon RX 8060S offers all 40 CUs with up to 2.9 GHz graphics clock and uses the new RDNA 3.5 architecture. The powerful iGPU is positioned between an RTX 4060 and 4070 laptop GPU and therefore enables gaming in FHD at maximum details in most demanding games. The 8060S can also utilize the full 128GB pool, which is perfect for running LLMs such as Deepseek 70B Q8, which runs comfortably on this machine.
- EIGHT CHANNEL LPDDR5X - LPDDR5X is a new ground breaking memory small form factor installed on-board. With blazing speeds up to to 8000MT/s, it runs 1.5x faster than the DDR5 SODIMMs; 90% better performance over DDR5 SODIMMs in video conferencing and photo editing; 30% better performance in productivity apps; 12% better performance in digital content workloads.
- QUAD SCREEN 8K DISPLAY SUPPORT - EVO-X2 AI Mini PC support 4-screen 4K/8K output via HDMI 2.1 (8K@60Hz), DisplayPort 1.4 (4K@60Hz), and dual USB 4 40Gbps Transfer speed (supporting PD3.0/DP1.4/DATA). Ideal for gaming, video editing, and multitasking, it provides expansive and crisp multi-display support.
The advisory alleges systematic efforts to obtain restricted proprietary functions and capabilities, including through multiple request pathways and activity that violated U.S. providers’ terms of use. The public materials cited here do not establish what role, if any, Chinese authorities played, nor do they independently establish each named firm’s conduct.
What AI model distillation is—and why the dispute is about conduct
Knowledge distillation is a standard machine-learning technique: a less capable model is trained using outputs from a more capable one. The technique can be used legitimately. CISA’s advisory distinguishes the method from what it describes as misuse: unauthorized, systematic extraction from proprietary systems.
Rank #2
- Built for Local AI Development: AMD Ryzen AI Halo is designed for local AI development and inference, featuring 128GB unified memory and support for up to 200B parameter models to build and run intensive AI workloads locally.
- 128GB Unified Memory: Features 128GB LPDDR5x unified memory at 8000 MT/s with 256 GB/s memory bandwidth, providing a shared memory pool across the CPU, GPU, and NPU to support larger AI models.
- AMD Ryzen AI Max+ 395 Processor: Features 16 cores, 32 threads, and Zen 5 architecture, paired with AMD Radeon 8060S integrated graphics featuring 40 RDNA 3.5 compute units and an AMD XDNA 2 NPU with up to 50 TOPS.
- Linux AI Developer Platform: Purpose-built for Linux-based AI development with full AMD ROCm software support and preloaded tools, models, and workflows optimized for local AI development.
- Compact, Connected Design: Includes a 2TB M.2 SSD, 10GbE LAN, Wi-Fi 7, Bluetooth 5.4, USB-C connectivity, and HDMI 2.1b.
In an April 16, 2026 House testimony, Yusuf Mahmood described alleged campaigns involving fraudulent accounts and large-scale querying to generate synthetic data for training other models. His testimony relayed company claims, rather than offering an independent forensic determination. It attributed more than 16 million exchanges through approximately 24,000 fraudulent accounts to an Anthropic report, and more than 13 million exchanges to MiniMax. Those figures are company-reported allegations relayed in testimony; they do not measure how much capability was transferred or establish the resulting competitive effect. Read Mahmood’s House testimony.
Free tools Windows power users keep installed
One-click scans. No signup required.
What China said in response
China’s Commerce Ministry called the U.S. claims groundless, said distillation is common practice among AI companies worldwide, and warned of countermeasures if U.S. restrictions use distillation as a pretext. China’s Foreign Ministry also urged the United States to stop what it called unfounded accusations and smears. These are the Chinese government’s positions, just as the advisory is the U.S. agencies’ account; the public material cited here does not independently resolve the allegations. The Associated Press reported the rebuttal.
Rank #3
- EVOLUTION RYZEN AI MAX+ 395 MINI PC - GMKtec EVO-X2 is the next evolution in AI mini PC Ryzen Strix Halo series. Thanks to AMD Simultaneous Multithreading (SMT) the core-count is effectively doubled, to 32 threads. Ryzen AI Max+ 395 has 64 MB of L3 cache and can boost up to 5.1 GHz, depending on the workload. The Ryzen AI Max+ 395 is currently rated as the "most powerful x86 APU" on the market for AI computing.
- AI NPU with XDNA 2 ARCHITECTURE - Powered by 16 “Zen 5” CPU cores, 50+ peak AI TOPS XDNA 2 NPU and a truly massive integrated GPU driven by 40 AMD RDNA 3.5 CUs, the Ryzen AI MAX+ 395 is a transformative upgrade and delivers a significant performance boost over the competition. The Ryzen AI Max+ 395 excels in consumer AI workloads like the llama.cpp-powered application: LM Studio. Shaping up to be the must-have app for client LLM workloads, LM Studio allows users to locally run the latest language model without any technical knowledge required and unleash their creativity and productivity.
- AMD RADEON 8090S iGPU GAMING PC - The AMD Radeon RX 8060S offers all 40 CUs with up to 2.9 GHz graphics clock and uses the new RDNA 3.5 architecture. The powerful iGPU is positioned between an RTX 4060 and 4070 laptop GPU and therefore enables gaming in FHD at maximum details in most demanding games. The 8060S can also utilize the full 128GB pool, which is perfect for running LLMs such as Deepseek 70B Q8, which runs comfortably on this machine.
- EIGHT CHANNEL LPDDR5X - LPDDR5X is a new ground breaking memory small form factor installed on-board. With blazing speeds up to to 8000MT/s, it runs 1.5x faster than the DDR5 SODIMMs; 90% better performance over DDR5 SODIMMs in video conferencing and photo editing; 30% better performance in productivity apps; 12% better performance in digital content workloads.
- QUAD SCREEN 8K DISPLAY SUPPORT - EVO-X2 AI Mini PC support 4-screen 4K/8K output via HDMI 2.1 (8K@60Hz), DisplayPort 1.4 (4K@60Hz), and dual USB 4 40Gbps Transfer speed (supporting PD3.0/DP1.4/DATA). Ideal for gaming, video editing, and multitasking, it provides expansive and crisp multi-display support.
What evidence is public—and what remains unresolved
The available accounts describe agency allegations, company claims relayed in congressional testimony, and the Chinese government’s denial. They do not provide an independent adjudication of each named firm or a verified estimate of the knowledge, model performance, or economic value transferred. Nor do they establish that every form of distillation or model querying is unlawful.
Distinguishing suspicious extraction from legitimate use is difficult when ordinary requests can be numerous and providers support some uses of model outputs. Brookings fellow Kyle Chan described the distinction as “looking for needles in an enormous haystack” in the AP’s April report. The central dispute is therefore narrower than whether distillation exists: it concerns whether particular activity was unauthorized, systematic, and in violation of provider terms or other applicable rules.
What the advisory recommends providers do
CISA’s recommendations are directed to frontier AI companies and infrastructure providers, not ordinary users. The advisory focuses on three defensive measures:
- Detect and mitigate suspicious activity. Monitor unusual prompts, accounts, networks, and behavior; compare subscription levels with usage; and flag accounts that reach maximum use immediately or generate enterprise-scale throughput.
- Make targeted response changes. Subtly alter responses to suspected distillation attempts to reduce their value to campaign operators.
- Share intelligence. Correlate suspicious activity across model providers, cloud platforms, and API aggregators so distributed campaigns are more visible.
The recommendations seek to identify patterns across services, but the advisory does not establish a universal technical method that can definitively classify every request as legitimate or abusive. It also does not report implementation results.
How this fits the administration’s broader AI policy
The crackdown pledge sits alongside a separate effort to promote U.S. AI exports. A July 23, 2025 White House fact sheet described an American AI Exports Program intended to support full-stack packages that may include hardware, data systems, models, cybersecurity measures, and sector-specific applications, subject to export controls and other requirements. That export policy provides context for the administration’s interest in protecting U.S. capabilities, but it does not itself establish enforcement authority against alleged distillation. Read the White House fact sheet.
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.
