Geopolitics & Sovereign Positioning
Top Line
China's top economic planner declared the domestic chip supply chain 'significantly safer,' signalling that US export controls are accelerating indigenous semiconductor industrial capacity rather than containing it — a direct challenge to the core logic of Western technology denial strategy.
The Pentagon has formally approved an OpenAI product for unclassified military work, marking the first institutionalised integration of a frontier commercial LLM into US armed forces operations and setting a precedent for adversary militaries tracking American AI-defence integration.
Chinese AI chipmakers MetaX and Iluvatar have swung to profitability, demonstrating that Beijing's captive domestic market strategy for AI hardware is beginning to yield commercial viability — reducing the leverage of US export controls over China's AI development trajectory.
Japan's record 2027 defence budget request explicitly centres AI, unmanned systems, and standoff capabilities, reflecting a broader Indo-Pacific pattern of US allies accelerating AI-enabled military modernisation under the umbrella of US deterrence commitments.
Foreign institutional investors increased Chinese mainland stock holdings by 87 percent in value in Q2 2026, with AI hardware and green energy as primary drivers — indicating that Western capital markets are not aligned with Washington's technology containment posture toward Beijing.
Key Developments
China's Chip Self-Sufficiency Transition: From Isolated Breakthroughs to Full-Chain Industrial Synergy
Li Chao, a senior official at China's National Development and Reform Commission, stated publicly that the domestic semiconductor supply chain had achieved 'full-chain industrial synergy' — a qualitative shift from prior years when China could demonstrate point-specific technological wins but lacked integrated production capability. Domestic chipmaking equipment and materials are now described as 'growing steadily,' per South China Morning Post. Separately, first-half earnings from MetaX Integrated Circuits and Iluvatar CoreX show both companies swinging to profitability, while Biren Technology narrowed losses — the first cohort of domestic AI chipmakers to demonstrate commercial viability at scale, per South China Morning Post.
This cluster of developments directly challenges the strategic premise of US export controls: that denying access to leading-edge chips and chipmaking equipment would constrain China's AI capacity long enough for the US to consolidate a decisive lead. The NDRC framing of 'full-chain synergy' is partially a political signal intended to demonstrate resilience, but the commercial profitability of domestic AI chipmakers lends it empirical weight. The more consequential question is whether Chinese chips are now competitive enough to support frontier model training, or whether they remain adequate only for inference and mid-tier applications — a distinction US policymakers need to track with technical precision rather than treating China's progress as monolithic.
Pentagon's OpenAI Integration: Institutionalising AI in US Military Operations
The Department of Defense has formally approved an OpenAI product for unclassified military use following months of security testing, per Defense One. The restriction to unclassified work is significant: it reflects genuine information security concerns about LLM data handling, but it also means the tool is immediately applicable to a vast range of military administrative, logistics, planning, and open-source intelligence workflows that constitute the bulk of daily defence operations. The approval marks an institutionalisation point — not a pilot or experiment — with OpenAI now formally embedded in the US defence industrial base.
The geopolitical significance extends beyond US capability. Beijing and Moscow both track US military AI adoption as an indicator of the integration timeline they must match or counter. An approved, fielded ChatGPT-class tool for Pentagon unclassified work will accelerate pressure on Chinese military planners to demonstrate equivalent or superior capability, feeding the action-reaction dynamic in AI-enabled military competition. It also further entrenches OpenAI's role as a strategic national asset — with implications for how the US government treats the company's international partnerships and export of API access.
China's AI Video Dominance: Structural Advantages Beyond Compute
Chinese AI firms are establishing a lead in video generation that derives from structural ecosystem advantages rather than raw compute: access to the world's largest short-video training datasets via ByteDance and Kuaishou, looser copyright enforcement reducing data acquisition costs, and aggressive pricing strategies that are undercutting US competitors in third-party markets, per South China Morning Post. Meanwhile, MiniMax's expansion of its Alibaba Cloud compute agreement by 220 percent to a $1.2 billion three-year ceiling signals that Chinese AI developers are scaling infrastructure investment at a rate that will compound capability advantages, per South China Morning Post.
This matters geopolitically because AI video generation is not merely a consumer product — it is infrastructure for influence operations, synthetic media, and information environment shaping. A Chinese lead in video generation capability translates directly into asymmetric capacity for state and non-state actors to generate credible synthetic content at scale and low cost. The pricing dynamic is also a market-penetration tool: Chinese AI video tools priced to dominate emerging market adoption create dependency relationships that parallel China's infrastructure diplomacy playbook. Countries that adopt Chinese AI video infrastructure as a default become embedded in Chinese AI ecosystems in ways that are difficult to reverse.
Japan's AI-Centred Military Modernisation and Indo-Pacific Alliance Architecture
Japan's record 2027 defence budget request — described by The Diplomat as explicitly structured around AI, unmanned systems, and a high-low weapons mix — represents more than a national spending decision. It reflects the operationalisation of Japan's 2022 National Security Strategy commitments under real budgetary pressure and signals that Tokyo is building AI-enabled military capacity as a core element of its strategic posture rather than as an addendum to conventional procurement. The framing of the budget request as 'only the opening bid' suggests political consensus for sustained escalation.
Japan's trajectory is significant for the broader US-led alliance architecture in the Indo-Pacific. As the US military integrates commercial AI tools and Japan builds AI-enabled unmanned systems, the two allies are converging on compatible AI-enabled military architectures — a convergence that will deepen interoperability but also create shared dependencies on US commercial AI providers. This raises questions about whether Japan is developing genuinely sovereign AI military capability or building a more capable but still US-dependent force — a distinction that matters if Washington's security commitments are ever tested.
Signals & Trends
Western Capital Markets Are Decoupled from Washington's Technology Containment Posture
Foreign institutional investors increased their holdings of Chinese mainland stocks by 87 percent in value during Q2 2026, with AI hardware explicitly cited as a driver, per South China Morning Post. Simultaneously, China Life Insurance — the country's largest insurer with $64.6 billion in first-half revenue — is explicitly scaling investments in AI, chips, and biotech as a long-term capital allocator. The implication is a structural divergence: US government policy is attempting to restrict capital and technology flows to China's AI sector, but Western fund managers are simultaneously increasing exposure to Chinese AI supply chain equities. This creates a political economy pressure on export control regimes — as Western financial institutions profit from Chinese AI hardware companies, the domestic political coalition supporting stringent controls becomes harder to sustain. Policymakers should track whether outbound investment screening mechanisms are being extended to cover portfolio equity exposure, not just direct investment and technology transfer.
The China-US AI Cooperation Signal: Tactical Dialogue Alongside Strategic Competition
A Foreign Policy analysis argues that practical US-China AI safety cooperation is achievable despite mutual strategic distrust, pointing to shared interests in preventing AI-enabled catastrophic risks, per Foreign Policy. Finland's Foreign Minister is scheduled to address European AI competitiveness at Chatham House in September, per Chatham House. These two signals together point to a pattern: even as hardware and model capability competition intensifies, a parallel diplomatic track around AI safety norms is developing — one that China, the US, and Europe all have incentives to engage with for different reasons. For China, safety diplomacy offers legitimacy and intelligence on Western red lines. For the US, it offers a channel that doesn't require technology sharing. For Europe, it offers relevance in a competition where it lacks frontier model or chip manufacturing capacity. The risk is that safety diplomacy becomes a substitute for substantive engagement rather than a complement to it — creating the appearance of governance without constraining competitive behaviour.
Space-Computing Integration as the Next AI Infrastructure Battleground
Shanghai's launch of a dedicated space-computing hub — bringing together aerospace developers, semiconductor firms, AI specialists, and cloud vendors — signals that China is treating orbital data infrastructure as an integrated component of its AI ecosystem rather than a separate domain, per South China Morning Post. The convergence of space-based data collection, edge AI processing, and ground-based cloud infrastructure represents an emerging capability layer that sits outside current export control frameworks focused on chips and model weights. Countries that establish orbital AI infrastructure first will have persistent, hard-to-contest advantages in real-time intelligence, communications resilience, and data sovereignty. This is an early-stage signal, but the institutional architecture being built in Shanghai — co-locating the relevant industrial actors — mirrors the cluster-building playbook China used successfully in electric vehicles and solar manufacturing.
Explore Other Categories
Read detailed analysis in other strategic domains