\n\n
Key developments: Washington and Beijing are pursuing sharply divergent diplomatic playbooks. The United States bundles advanced chips, cloud services, models, compute power and financing into a “full‑stack” export package for favored allies, while China offers locally owned AI models, training, capacity‑building and new multilateral bodies, even pledging 5,000 AI seminars for developing nations and setting up AI application centers with groups such as BRICS and ASEAN. Beijing’s collaborative approach is already bearing fruit, exemplified by the ASEAN‑China AI Industry Innovation Center opened in May and the rapid expansion of Chinese compute capacity—24 GW today, with another 50 GW under construction, narrowing the gap with the U.S.’s 56 GW. Regional actors are lining up: South Korea is earmarking $3.5 B for sovereign frontier AI, and Finland, Norway and 19 other nations have called for global cooperation on frontier AI risks.
Why it matters: The competing models shape global influence. The U.S.’s “full‑stack” strategy secures market dominance but risks alienating nations that prefer ownership and local control. China’s model, by providing affordable, locally operable AI and extensive training, wins goodwill in the Global South and creates dependency on Chinese stacks—potentially shifting standards and data governance away from Washington. The scale‑up of compute in China, together with deals like Tencent’s access to Oracle’s Southeast Asian chips and the Huawei‑Qualcomm licensing pact, illustrates how export controls can be sidestepped, further eroding U.S. leverage.
What to watch next: • Expansion of Pax Silica – how many allies adopt the U.S. supply‑chain initiative.
• Progress of China’s AI application centers with BRICS, ASEAN and the African Union.
• The impact of South Korea’s $3.5 B sovereign AI investment on regional tech balance.
• Responses from Europe and the EU’s AI Act as they try to carve an independent path.
• Market moves such as Nvidia’s talks with Reflection AI and further chip‑licensing deals that could alter the hardware landscape.
These trends indicate that the AI rivalry is moving beyond model contests to an industrial‑scale competition over compute, chips, and governance frameworks, with geopolitical ramifications for every major economy. [[1, 2]]