Moonshot AI Seeks Advanced Nvidia Chips for Kimi K4 Amid US Embargo

A report from X account @Polymarket on Tuesday, July 28, 2026, stated that Moonshot AI is seeking access to more advanced Nvidia chips to train its Kimi K4 model. This move comes just days after a White House official accused the startup of illegally accessing Nvidia GB300 chips through servers in Thailand to train its previous model, Kimi K3. According to a CNBC report on July 23, 2026, a White House official said that Moonshot AI, which released the powerful Kimi K3 model last week, had accessed Nvidia GB300 chips in Thailand. The allegation marks a new escalation in Washington's efforts to enforce export rules for advanced technology to China, which have been in place since October 2022 and gradually tightened.

Moonshot AI, based in Beijing, is known as one of China's leading developers of large language models (LLMs). The Kimi K3 model released last week is said to have capabilities comparable to leading models from OpenAI and Anthropic. Now, with plans to develop Kimi K4, the company appears undeterred by US regulatory pressure. Reports from CapWolf and CryptoBriefing, though with lower confidence levels, also echoed the allegation that Moonshot AI exploited loopholes in the global supply chain to obtain Nvidia chips. Both reports stated that access to the GB300 was arranged through a third party in Thailand, a country not subject to direct US export restrictions. US restrictions on advanced chip exports aim to slow China's military and technological progress, particularly in artificial intelligence.

The Nvidia GB300 chip falls into the category of chips banned for export to China without special permission. However, diversion through third countries has become a major concern for US regulators. As of this writing, there has been no official statement from Moonshot AI or Nvidia regarding the allegations. The Chinese government has also not responded. Meanwhile, Nvidia's stock (NVDA) closed at style="background-color: #ffffff;"94.46 in the last trading session, down 1.04%, though it is unclear whether the decline is directly related to the report. Moonshot AI's move to seek more advanced chips for Kimi K4 indicates that the company remains optimistic about continuing AI model development despite regulatory hurdles.

If successful, this could widen the gap between Chinese and US AI capabilities and trigger a stronger response from Washington. Analysts predict that the Moonshot AI case could set a precedent for other Chinese AI companies to seek alternative routes to obtain advanced chips. Thailand, Singapore, and Malaysia have become suspicious transit points in the global chip supply chain. The US government is likely to tighten oversight of these countries. On the other hand, Moonshot AI's success in accessing GB300 chips for Kimi K3 shows that loopholes in the embargo system still exist. This raises questions about the long term effectiveness of US export policy. If Chinese companies continue to access cutting edge technology, the US strategic goal of slowing China's AI progress may not be achieved.

This development also impacts the global semiconductor industry. Nvidia, the primary supplier of AI chips, is in a difficult position between complying with US regulations and maintaining global market share. The company has designed special chips for China that meet export limits, such as the H100 and H800 series, but demand for more advanced chips remains high. There is no official confirmation regarding the specifications of the chips Moonshot AI is seeking for Kimi K4. However, if the company targets chips equivalent to the B200 or newer Blackwell architecture, it would mark a significant escalation in efforts to circumvent the embargo. Currently, Nvidia has not released chips with performance equivalent to the B200 for the Chinese market.

The Moonshot AI case highlights the major challenges in enforcing technology export controls in the era of globalization. Although the US has strict rules, complex supply chains and high market demand create strong incentives for companies to find loopholes. Going forward, the effectiveness of this policy will heavily depend on international cooperation and the US ability to close existing gaps. As the situation unfolds, the implications for US China tech competition are profound. The ability of Chinese AI firms to access advanced hardware despite restrictions could accelerate their development and challenge US leadership in AI. Meanwhile, US policymakers face pressure to strengthen enforcement without disrupting global supply chains or alienating allies.

The coming months will be critical in determining whether the embargo can achieve its intended goals or if it will merely drive innovation in circumvention strategies.