export controls

A US official accused Chinese AI firm Moonshot AI of circumventing export controls to acquire advanced Nvidia chips for its Kimi K3 large language model. The Kimi K3, with 2.8 trillion parameters, shows reduced performance gaps compared to global leaders. The US is restricting high-performance AI chip exports to China to curb its AI ambitions. This incident highlights the intensifying AI competition and concerns over Chinese firms seeking computing power abroad.

ASML, a key supplier of chipmaking equipment, faces geopolitical pressure amid booming AI demand. While China remains a crucial market, the US is pushing for stricter export controls on less advanced DUV lithography machines. ASML, which doesn’t export its most advanced EUV technology to China, has raised its guidance due to strong customer demand for AI chips. Potential US legislation like the MATCH Act could impact its orderbook, forcing ASML to balance global demand with international relations.

A US official stated that very few Nvidia H200 AI chips have been shipped to China, suggesting a potential shift in export controls. This could boost Nvidia’s revenue, as they previously excluded Chinese sales from forecasts due to stringent restrictions. While H200s are powerful, Nvidia is transitioning to newer Blackwell architecture. The limited availability of advanced chips impacts China’s AI development, forcing reliance on less capable domestic alternatives.

The US Department of Commerce has lifted export controls on Anthropic’s Claude Fable 5 and Mythos 5 AI models, resolving a dispute with the Trump administration. This decision allows Anthropic to redeploy these advanced models after a temporary suspension due to national security concerns. The move comes amid rapid advancements in Chinese AI and aims to maintain US leadership in the field.

Nvidia CEO Jensen Huang affirmed that U.S. national security supersedes commercial interests, especially concerning advanced AI exports. He stressed that attempts to bypass export restrictions are futile, as Nvidia’s integrated systems require trusted support to function. Huang also highlighted AI’s proven return on investment through increased efficiency and demonstrated profitability, driving demand for computing power. Nvidia remains committed to shareholder value, planning to return 50% of free cash flow through buybacks and dividends.

Nvidia CEO Jensen Huang has been invited to testify before the Senate Banking Committee regarding the company’s significant business ties with China. Lawmakers are concerned that advanced U.S. AI chips could be used by China for military purposes, impacting national security. The hearing will also address U.S. export controls and Nvidia’s role in the global AI landscape, reflecting bipartisan and bicameral scrutiny of U.S. tech engagement with China.

The U.S.-China summit focuses on tech trade, with critical minerals and market access for American firms as key issues. U.S. tech leaders are present to influence AI supply chains and secure market penetration. While China signals openness, negotiations will likely involve conditional access to technology and address China’s dominance in critical minerals, a significant leverage point. The U.S. aims for tariff reductions and stable mineral supplies amidst ongoing geopolitical and economic tensions.

Super Micro Computer faces leadership turmoil after co-founder Yih-Shyan Liaw’s indictment for allegedly smuggling AI chips to China. Liaw resigned, while another executive is on leave. The company’s stock dropped 33% amid allegations of circumventing export controls. Super Micro appointed a new acting chief compliance officer and is enhancing its framework. The case highlights regulatory challenges in the AI hardware sector.

Meta’s acquisition of AI startup Manus is a case study in cross-border compliance. China is scrutinizing the deal, focusing on export controls and technology transfer rules, despite Manus’s relocation from Beijing to Singapore. This highlights that a vendor’s domicile doesn’t determine regulatory exposure; technology origin is key. Businesses must now conduct deeper due diligence on AI vendors, examining technology origin, transfer compliance, and operational continuity to navigate evolving geopolitical and regulatory landscapes.

In 2025, AI chip shortages, driven by geopolitical tensions and soaring demand, became the primary obstacle for enterprise AI deployment. This led to increased costs, with monthly AI spending projected to rise significantly, and longer deployment timelines. A critical memory chip crisis compounded these issues, driving up prices and creating extended lead times. Companies learned to diversify supply, budget for volatility, optimize efficiency, and consider hybrid infrastructure models to navigate these persistent constraints, acknowledging that hardware limitations now dictate AI strategy.