Short answer: Nvidia CEO Jensen Huang’s public praise of President Donald Trump was followed by a controlled U.S. licensing pathway for Nvidia’s H200 AI accelerator, but there is no public evidence that praise caused the policy change. The January 2026 rules did not restore unrestricted access to advanced chips. They allowed case-by-case review with customer-screening, testing, end-use, supply and processing-power limits. The strategy was partly intended to keep Chinese companies tied to U.S. technology instead of Huawei. Yet earlier restrictions had already encouraged Chinese buyers and Beijing to develop domestic alternatives, so regulatory permission has not equaled a recovered Nvidia market.
What Nvidia actually praised
In 2025, Jensen Huang argued that earlier U.S. AI-chip export controls had failed commercially and strategically. In Nvidia’s account, the restrictions cut off sales while encouraging Chinese companies to buy from Huawei and invest in domestic substitutes. Huang also argued that U.S. companies were not the only source of advanced technology.
Reuters reported Huang’s criticism and his support for Trump’s approach to China-chip policy: Reuters reporting on Nvidia’s export-control criticism. The timing creates an obvious political narrative, but it does not establish a quid pro quo. Public praise, private lobbying and a later policy decision can be related without proving that one caused the other.
The defensible sequence is narrower: Huang criticized the previous policy; the Trump administration signaled that some sales might resume; and the Commerce Department later created a conditional licensing route.
#1 Best Overall
- AI Performance: 767 AI TOPS
- OC mode: 2632 MHz (OC mode)/ 2602 MHz (Default mode)
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- Axial-tech fan design features a smaller fan hub that facilitates longer blades and a barrier ring that increases downward air pressure
- A 2.5-slot design maximizes compatibility and cooling efficiency for superior performance in small chassis
What changed on January 13, 2026
The Bureau of Industry and Security (BIS) moved applications for Nvidia H200, AMD Instinct MI325X and similar products to case-by-case review when specified conditions are met. That is a significant relaxation from a broadly restrictive posture, but it is not a general lifting of controls or a presumption that every application will be approved.
| Requirement | Practical meaning |
|---|---|
| Eligible products | H200, AMD MI325X and similar products named by BIS; the policy does not automatically cover Blackwell, Rubin or later top-end platforms. |
| Customer controls | Chinese purchasers must maintain export-compliance and customer-screening procedures. |
| Independent testing | U.S.-based third-party testing must verify performance and security. |
| Supply protection | Exports cannot reduce semiconductor production capacity available to U.S. customers. |
| End use | Products cannot be destined for prohibited military, military-intelligence, nuclear, missile or chemical/biological-weapons uses. |
| Remote access | Rules restrict access by prohibited end users through cloud or other remote arrangements. |
| Processing-power ceiling | Aggregate processing power shipped to China or Macau must not exceed 50% of comparable processing power shipped to U.S. customers for U.S. end use, under the applicable BIS definitions. |
See the BIS policy announcement and EAR §748 regulatory text. The 50% ceiling is not a 50% unit quota. BIS calculations use technical measures including total processing performance, memory and interconnect bandwidth, memory capacity and power consumption.
Why the H200 matters
The H200 is a high-performance data-center accelerator for AI training and inference. It was more capable than Nvidia’s China-specific H20, which had been designed to comply with earlier restrictions, but it was not Nvidia’s absolute newest platform.
Nvidia’s filings show the financial cost of policy volatility. The company recorded a $4.5 billion charge tied to H20 excess inventory and purchase obligations after an April 2025 licensing requirement. Nvidia also reported approximately $50 million in H20 revenue under licenses granted later, measured as of the filing cited. Those figures are company disclosures, not a complete measure of Chinese demand. Nvidia fiscal 2026 filing.
Free tools Windows power users keep installed
One-click scans. No signup required.
H200 access therefore represents a possible product-level reopening, not permission to sell every advanced Nvidia system. Nvidia says product classifications and export requirements can change without notice: Nvidia export-regulations policy.
Rank #2
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- Powered by GeForce RTX 5060
- Integrated with 8GB GDDR7 128bit memory interface
- PCIe 5.0
- WINDFORCE cooling system
Why Washington thought this could constrain Huawei
The administration’s theory has three linked parts:
Keep Chinese customers in the U.S. ecosystem
If Chinese developers can lawfully obtain capable Nvidia hardware and continue using its software ecosystem, they may have less reason to standardize on Huawei accelerators. An installed Nvidia base can preserve influence over tools, models, support and future infrastructure choices.
Preserve U.S. commercial and technological strength
Sales provide revenue for research, manufacturing commitments and supplier investment. BIS framed controlled H200 access as a way to strengthen the American technology ecosystem while managing national-security risks: BIS statement.
Use licenses as a security instrument
Customer screening, end-use certifications, testing and shipment ceilings are intended to limit diversion while retaining some commercial access. This is a compromise between a near-total blockade and unrestricted sales.
Why the policy may not restore Nvidia’s Chinese business
China had already adapted
Earlier restrictions reduced Nvidia’s ability to serve China and increased incentives for Huawei and other domestic suppliers. The same controls intended to protect U.S. technological leadership also gave Beijing a reason to accelerate local substitution.
Rank #3
- Powered by the NVIDIA Blackwell architecture and DLSS 4. System Requirements: Minimum 850W PSU with 16-pin 12V-2x6 (12VHPWR) connector required. Verify before purchasing.
- Military-grade components deliver rock-solid power and longer lifespan for ultimate durability. Compatibility: 348mm (13.7") length, 3.6 slots, 4.3 lbs. Confirm case clearance and slot spacing. GPU bracket included.
- Protective PCB coating helps protect against short circuits caused by moisture, dust, or debris
- 3.6-slot design with massive fin array optimized for airflow from three Axial-tech fans
- Phase-change GPU thermal pad helps ensure optimal thermal performance and longevity, outlasting traditional thermal paste for graphics cards under heavy loads
Permission is not delivery
A company can move through several separate stages:
- U.S. policy makes a product potentially licensable.
- A specific exporter and customer receive a license.
- The customer places an order.
- The product is physically shipped.
- The system is installed and used.
- The customer returns for repeat purchases.
Reuters later reported that deliveries had not yet occurred to roughly 10 cleared Chinese firms as of May 2026, while AP reported that Nvidia’s China sales were stalling as Huawei and domestic chips gained ground. Those reports show why an opening in the rules cannot be treated as a market recovery. Reuters-linked report on delivery uncertainty and AP reporting on China sales.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Repair Windows errors before they cause bigger problems3Fix the driver behind crashes, sound loss and screen glitchesBeijing controls the other side of the transaction
U.S. authorization does not require Chinese authorities or customers to accept Nvidia products. Beijing can encourage domestic procurement, impose security reviews, limit deployment or treat U.S. accelerators as temporary tools while subsidizing Huawei.
Customers dislike policy uncertainty
Licenses can change, shipment caps can disrupt supply, and a future administration can impose new restrictions. Chinese buyers building multi-year data centers may prefer a technically weaker domestic product if its legal and political availability is more predictable.
Huawei’s restrictions and the enforcement problem
Huawei has faced extensive U.S. export controls and Entity List restrictions for years. The January H200 policy was not itself a newly identified Huawei ban. Its strategic effect was indirect: Washington paired limited Nvidia access with continued pressure on Chinese procurement networks and Huawei’s competitive position.
Rank #4
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- Powered by GeForce RTX 5070 Ti
- Integrated with 16GB GDDR7 256bit memory interface
- PCIe 5.0
- WINDFORCE cooling system
That enforcement challenge extends beyond mainland factories. In May 2026, Reuters reported that Commerce moved against a potential route in which advanced Nvidia chips could reach subsidiaries of Chinese companies outside China, including through Malaysia: Reuters reporting on foreign-subsidiary shipments.
- A foreign subsidiary may still be connected to a restricted Chinese parent.
- Cloud access can allow a prohibited user to use a chip without directly owning it.
- Third-party testing can verify a system’s performance and security without proving that the customer will prevent later diversion.
- Know-your-customer controls are difficult once hardware is installed, resold or accessed remotely.
Reuters reported that Nvidia had not agreed to proposed customer-screening conditions at an earlier stage, illustrating the difference between government requirements and a company’s position during negotiations: Reuters on proposed guardrails.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.The strongest objections
It may transfer strategic capability
Lawmakers and former officials argued that allowing H200 sales could improve Chinese military, intelligence and surveillance capabilities and weaken the U.S. lead. Reuters on congressional and former-official criticism.
It may reward the problem the controls created
If restrictions pushed customers toward Huawei, reopening access after that shift may leave Nvidia with fewer buyers while giving Huawei a stronger domestic position.
Guardrails may be hard to verify
End-use promises, remote access and overseas subsidiaries create monitoring problems that do not disappear because a license includes certifications.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallCrashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteBest Value
- AI Performance: 1005 AI TOPS
- OC mode boosts clock 2587 MHz (OC mode) / 2557 MHz (Default mode)
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- SFF-Ready enthusiast GeForce card compatible with small-form-factor builds
- Axial-tech fans feature a smaller fan hub that facilitates longer blades and a barrier ring that increases downward air pressure
The commercial upside may be too small
China may block or discourage purchases, licenses may arrive too slowly, and capped supply may not justify the cost of designing infrastructure around a politically exposed product.
The broader trade-off
| Policy objective | Potential benefit of controlled H200 sales | Risk |
|---|---|---|
| U.S. commercial competitiveness | Nvidia and AMD retain some revenue and customer relationships. | China may use the hardware while replacing it domestically. |
| Containing Huawei | Chinese firms remain connected to U.S. hardware and software. | Restrictions may already have strengthened Huawei’s incentive and market. |
| National security | Licensing and testing create formal controls. | Diversion and remote use can be difficult to detect. |
| Supply-chain resilience | Chip revenue supports U.S. research and production. | Policy changes can create inventory losses, as Nvidia’s H20 charge demonstrates. |
| Diplomatic leverage | Access becomes a bargaining tool. | Unpredictability reduces customer confidence. |
What this means for AI-infrastructure buyers
Companies evaluating Nvidia, AMD or Chinese accelerators should assess more than benchmark performance:
- Export eligibility by country, ownership, end user and intended use.
- Whether affiliates or cloud users in restricted jurisdictions can access the system.
- The exact product generation; H200 is not interchangeable with Blackwell, Rubin or successor products.
- Repeat-order reliability under licensing and manufacturing constraints.
- Software dependence, including CUDA, Nvidia AI Enterprise and migration costs.
- Total cost of ownership: networking, power, cooling, support and compliance.
- The possibility that a currently permitted product becomes restricted later.
Cloud procurement does not automatically bypass chip controls. Hosted access can create additional jurisdiction and remote-access questions. Nvidia documentation for enterprise infrastructure is available at Nvidia AI Enterprise reference architecture.
What the evidence supports
Trump’s administration softened restrictions for specified AI accelerators by creating a conditional, case-by-case licensing path. The policy was partly justified as a way to preserve U.S. influence and prevent Huawei from becoming China’s default supplier.
It is not established that Huang’s praise caused the change. Nor is there evidence that the policy restored Nvidia’s Chinese market. The available reporting instead points to a gap between approval and shipment, continuing enforcement risks and a Chinese market that had already moved toward domestic substitutes.
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.
