NVIDIA and Oracle have publicly opposed proposed US regulations on AI chip sales and export controls.

The US government announced plans on Monday for new regulations governing AI chips manufactured in America, establishing restrictions on the computational power of chips that US companies can sell to most countries. The framework distinguishes between allied nations and other international buyers, reflecting a strategic approach to semiconductor export control.

Eighteen key allies and partners will face no restrictions on their purchases of American-made AI chips. All other countries, however, will be limited to purchasing no more than 50,000 AI GPUs over the period between 2025 and 2027. This cap represents a significant constraint on global distribution of advanced semiconductor technology.

NVIDIA responded to the regulations with a statement emphasizing potential market consequences: "Data center related GPU products and technologies are widely integrated into the personal computers used by consumers worldwide. Imposing controls on them is pointless and could prompt the global market to seek alternatives, which would undermine US interests." The company's position reflects concerns that overly restrictive policies may drive international buyers toward non-American suppliers.

Oracle offered a more dire assessment of the regulations, characterizing them as potentially one of the most destructive policies in the history of the US tech industry. The company warned that the new restrictions would directly cut US companies' global chip market share by 80 percent, suggesting substantial competitive harm alongside the stated foreign policy objectives.