Alexandr Wang, head of Meta AI, urged American startups to treat serving the US government as a "bedrock principle" rather than a commercial convenience. His remarks followed a Forbes report revealing that Chinese companies are sourcing training data from Silicon Valley firms. Aakash Sabharwal, VP of engineering at Scale AI, warned that American companies should not sell data to Chinese AI labs, stating that his firm has "turned down revenue" for this reason and that companies doing such work are "undermining American AI leadership and risking national security." The investigation found that American vendors serving OpenAI, Anthropic, and in some cases the U.S. government, including Surge AI, Mercor, AfterQuery and Turing, have also supplied Chinese buyers such as Tencent, Ant Group, Alibaba and ByteDance. Forbes estimated that the top six Chinese AI labs collectively spend around $500 million a year with American data-labeling companies. The report noted that while the U.S. has restricted China's access to advanced AI chips, it has not placed similar restrictions on training data. Separately, AI Czar David Sacks called the performance of China's new Kimi K3 model "concerning" for American technological dominance, noting it took #1 on the Frontend Code Arena. Beijing-backed startup Moonshot AI open-sourced a massive 2.8 trillion-parameter system.