In 2025, Armenia and Azerbaijan announced a preliminary agreement to end nearly four decades of conflict around Nagorno-Karabakh, a disputed region in the South Caucasus. This followed Azerbaijan's reoccupation of the enclave in September 2023. While Armenian and Azerbaijani leaders credited U.S. President Donald Trump with playing a role in the peace process, the resolution also involved the use of advanced American technologies, particularly AI chips, as a negotiation tool. According to the Wall Street Journal, American negotiators reportedly used the promise of expanded access to Nvidia chips to encourage the conclusion of the agreement, including authorizing more purchases of chips for a data center project in Armenia. Armenia has since begun turning this promise into tangible infrastructure. The Firebird AI plant, located in Hrazdan, was officially inaugurated on August 8, 2026. According to the Armenian Ministry of High Technology Industry, the plant's capacity is expected to exceed 100,000 Nvidia Blackwell and Vera Rubin GPUs and reach more than 400 megawatts once the third phase is completed by the end of 2027. Nvidia has also announced its intention to invest in the project. In April 2026, the ministry and Firebird signed an agreement under which the Armenian state will purchase $25 million worth of high-performance computing resources from the company over five years. In Azerbaijan, Nvidia is already involved in several projects aimed at advancing the country's artificial intelligence capabilities. At the end of 2025, the state-owned AzInTelecom launched the country's first supercomputing center, an infrastructure capable of processing large amounts of data using 32 Nvidia H200 graphics processors, essential for running AI models. In April 2026, Nvidia also signed a partnership with the Azerconnect Group operator to deploy its GeForce NOW cloud gaming service. This agreement also paves the way for the creation of future "AI factories," data centers specifically designed to run and train artificial intelligence systems. This "deployment" strategy is not limited to the Caucasus. Similar negotiation frameworks, based on the promise of state-of-the-art Nvidia computing infrastructure, are also offered by the White House to strategic powers such as the United Arab Emirates and Saudi Arabia. Nvidia has become much more than a semiconductor giant in recent years. Thanks to its technologies, its economic weight, and its central role in the development of artificial intelligence, the company has become an indispensable actor in global geopolitics, to the point that its chips can now serve as an important tool for international relations. Beyond the "chip deals," Nvidia has also become one of the key players in recent years in Sino-American relations. While the semiconductor war pits Washington against the People's Republic of China, with sanctions, export restrictions, and bans targeting the most advanced models, Nvidia is trying to preserve its access to the Chinese market. Through the voice of its CEO, Jensen Huang, Nvidia defends the idea that a complete cutoff of exports could push Chinese companies to accelerate the development of their own chips. Continuing to sell some chips to China allows the company to maintain a technological advantage over actors such as Huawei or Cambricon and, in the long term, to preserve its dominance on the global market. An issue far from trivial for Nvidia, which estimates the potential of the Chinese market to be nearly $50 billion. According to Leevi Saari, a researcher at the University of Amsterdam, the idea behind this strategy is to maintain a sufficient capacity gap to prevent Chinese companies from turning to local alternatives. A positioning that illustrates the paradox of the American giant: its chips have become a tool of power for Washington, but their sale remains also a major commercial issue for Nvidia, whose interests do not always coincide with those of American foreign policy.