Saudi Arabia recently celebrated the launch of what it proudly calls the world’s most advanced Arabic-language artificial intelligence model, a milestone intended to showcase the kingdom’s growing technological ambitions. Yet, beneath the diplomatic ribbon-cutting lies a complex reality that defines the modern geopolitical landscape: the core technology powering this achievement came from China, even as Riyadh continues to spend billions of dollars building out its broader artificial intelligence infrastructure in close partnership with American tech giants.
The groundbreaking Arabic model, known as Humain-m3, is built upon an open-source architecture developed by MiniMax, an emerging Chinese artificial intelligence startup. Meanwhile, Humain, the state-backed enterprise that commissioned the model, has simultaneously forged major multi-billion-dollar agreements with Nvidia and several other prominent U.S. technology corporations to secure the heavy computing power necessary to sustain its national AI strategy.
This duality is far from an isolated incident. Across the globe, from the emerging markets of Latin America to the fast-growing economies of Southeast and Central Asia, nations are actively pursuing a dual-track strategy. They are purchasing advanced American silicon and hardware while simultaneously adopting Chinese open-source foundation models, effectively splitting their critical artificial intelligence projects between companies originating from the world’s two dominant technological superpowers.
This pragmatic approach presents an increasingly complicated challenge for Washington in its ongoing strategic competition with Beijing. Even as foreign governments pour vast financial resources into American-made artificial intelligence infrastructure, they are demonstrating that they are not necessarily inclined to choose American technology for every single layer of the complex artificial intelligence stack.
“Many countries around the world, particularly in the Global South, want the best of both worlds,” explained Kyle Chan, a research fellow at the John L. Thornton China Center at the Brookings Institution. Chan noted that many developing and middle-power nations often do not attach the exact same security stigma to Chinese technological products and services that Washington routinely emphasizes. Ultimately, these governments want to keep their geopolitical and technological options wide open, ensuring they never become excessively dependent on any single source of technology or vulnerable to external political pressure.
Staying in the Middle
The tension and strategy behind this balancing act became vividly apparent in South America. In August, the Brazilian government announced a sweeping investment package worth approximately $444 million dedicated entirely to artificial intelligence development. Rather than channeling the funds exclusively through traditional Western allies, Brasilia effectively split the capital between Chinese and American technology providers.
Under the plan, roughly $255 million is earmarked for advanced supercomputing infrastructure to be constructed in Rio de Janeiro, developed through close partnerships with Chinese telecommunications and technology giants Huawei and iFlytek. The remaining portion of the funding, amounting to around $189 million, is directed toward a separate national supercomputing project that Brazil has been tendering, with expectations that the hardware will be sourced directly from Nvidia.
The administration of President Luiz Inácio Lula da Silva articulated this deliberate philosophy in an official statement, emphasizing that the overarching national strategy is explicitly designed to avoid dependency on any single company, proprietary technology, or foreign country. Officials added that these strategic investments were fundamentally intended to strengthen Brazil’s national sovereignty over its own data and digital future.
For most sovereign states navigating the modern tech landscape, achieving technological sovereignty does not mean attempting the prohibitively expensive feat of building an entire artificial intelligence stack entirely at home, according to Jesse Marks, a China-Middle East research scholar at the Australian National University. As the costs associated with competing at the absolute technological frontier have skyrocketed, governments are forced to make pragmatic calculations about which specific layers they can develop independently and where they have no choice but to rely upon foreign suppliers.
For a nation like Saudi Arabia, Marks observed, choosing Chinese foundations to anchor sovereign artificial intelligence models represents a natural decision if those platforms demonstrate superior performance on regionally relevant benchmarks, regardless of the potential political friction such a choice might generate with Washington.
At the same time, the United States retains substantial leverage in specific layers of the technological stack where its proprietary systems remain exceptionally difficult to replace. This includes advanced semiconductors, specialized networking equipment, and sensitive cloud computing infrastructure. However, Washington faces a distinct disadvantage when dealing with nations that urgently desire advanced artificial intelligence capabilities but simply cannot afford premium American systems or software licenses.
Outside of the core network of trusted American allies who possess the financial resources to afford comprehensive U.S. stacks, the market inevitably turns toward Chinese alternatives and accessible open-source models, Marks explained. These technological decisions are increasingly becoming deeply intertwined with the broader diplomatic relationships and strategic maneuvering of middle powers. Artificial intelligence has firmly established itself as one of the primary portfolios that these governments must carefully manage when negotiating complex partnerships with major global players like the United States.
These commercial and strategic developments unfold against a backdrop of rising regulatory and political friction. Washington has repeatedly accused Chinese artificial intelligence laboratories of engaging in "free-riding," alleging that Beijing-backed entities have trained their domestic models using frontier American artificial intelligence systems at an unprecedented, industrial scale. China has consistently rejected these accusations and has issued stern warnings that it is fully prepared to retaliate if the United States implements severe new curbs or restrictions against Chinese artificial intelligence firms.
China Catching Up
Chinese technology companies are simultaneously ramping up their global efforts to aggressively compete with American firms across every layer of the artificial intelligence ecosystem. In North Africa, Huawei has reportedly submitted competitive bids to construct advanced artificial intelligence data centers for the Egyptian government, going head-to-head against a rival proposal centered around Nvidia technology. Securing such a contract would grant Chinese providers another vital foothold in a regional market where enterprise and government AI infrastructure has historically been dominated by American vendors.
Southeast Asia has emerged as another primary theater where the high-stakes competition for technological influence between Washington and Beijing is becoming increasingly visible. In July, the United States launched the U.S.-ASEAN AI Spark initiative, a strategic program designed to offer targeted financial and technical support to Southeast Asian governments interested in adopting American artificial intelligence technologies for public-sector modernization.
Despite these American overtures, individual nations in the region are actively evaluating alternative pathways. Malaysia has reportedly considered utilizing Huawei’s Ascend 910C accelerators to serve as the technological backbone of its ambitious national sovereign artificial intelligence initiative, highlighting the persistent pull of Chinese hardware even in regions heavily courted by Western diplomacy.
For now, the world’s two primary artificial intelligence superpowers appear to be deploying distinctly different strategic playbooks in their efforts to build and project influence abroad.
Arindrajit Basu, a doctoral researcher specializing in the global governance of emerging technology at Leiden University, noted that China, much like any leading global power, naturally seeks an international order that functions in a manner favorable to its core interests and reflects its governance values. In the context of the internet and its allied digital technologies, this vision prioritizes a framework where the state maintains a strong, central role in preserving public order and security online. However, Basu pointed out that China has generally avoided the overt, coercive ultimatums that have occasionally characterized recent Western diplomatic approaches.
Washington, by contrast, has weighed policies aimed at discouraging partner nations from attempting to straddle the divide between U.S.-led and China-led technology ecosystems. Such a hardline stance risks placing developing economies and traditional allies alike in extraordinarily difficult diplomatic positions.
Kazakhstan provides a clear illustration of this complex balancing act. The Central Asian nation recently joined the U.S.-led Pax Silica initiative, a coalition designed to secure resilient supply chains spanning semiconductors, compute hardware, and other critical technologies. Simultaneously, Kazakhstan has engaged with China’s World Artificial Intelligence Cooperation Organization, known as WAICO, which focuses heavily on collaborative artificial intelligence development and international governance standards. Uzbekistan signaled its own intent to engage with both rival frameworks.
Basu explained that these competing multilateral initiatives currently possess markedly different strategic objectives. Pax Silica operates as an exclusive, invitation-only grouping centered on the direct sharing of critical resources and hardware among trusted members. In contrast, WAICO presents itself as an inclusive governance body designed to complement United Nations frameworks, without currently tying membership directly to the allocation of scarce computing resources or joint hardware research.
While Beijing would theoretically prefer that countries make a definitive choice between technological blocs, Chinese leadership likely recognizes that issuing heavy-handed ultimatums is unrealistic at this stage of the technological cycle. Consequently, Beijing continues to rely on norms-oriented diplomacy and capacity-building efforts that steadily encourage foreign adoption of Chinese products and infrastructure.
Analysts warn that Washington’s efforts to compel third-party nations to take a definitive side in the global artificial intelligence race will likely encounter growing resistance. Even long-standing European allies are actively rethinking their deep dependencies on American technology platforms. Emerging and middle-power nations are determined to preserve their strategic flexibility by maintaining active working relationships with both the United States and China.
Singapore may soon serve as an early testing ground for this diplomatic tightrope. As a founding signatory of the U.S.-backed Pax Silica framework, the Southeast Asian nation has publicly confirmed that it has also received an official invitation to join the China-led artificial intelligence organization and is currently evaluating the proposal.
Addressing her nation’s Parliament, Singapore’s Minister for Digital Development and Information, Josephine Teo, outlined the pragmatic approach guiding her government’s strategy. Singapore regularly engages different global partners across a wide array of multilateral and bilateral platforms, she noted, explaining that these diplomatic and technological engagements arise at different times, serve distinct national purposes, and naturally evolve as Singapore’s strategic interests and regional circumstances continue to change.
