AI & Algorithm Governance

China's AI policy is still built for speed, but the control track is no longer frozen. At a State Council Information Office press conference on August 26, MIIT Vice Minister Xin Guobin (辛国斌) said China has formulated nearly 200 key AI standards and is organizing multiple cities to carry out AI ethics review and service practices ¹. That is the clearest governance move in this batch, yet it is framed as steering "sound and orderly development" of an AI already "largely integrated across the entire chain"—governance as enabling infrastructure, not a brake. The promotion-vs-control balance still tilts decisively toward deployment, but outside readers should now track standards and ethics pilots, not just compute lines.

Guizhou remains the sharpest baseline of the promotion machine. Provincial compute capacity has reached 176 EFLOPS, with Guiyang-Gui'an at 174 EFLOPS and smart-computing share above 98% in that core area ^2,3. The province counts 50 key data centers in operation or under construction (40 already running), up from seven a decade ago, and 26 large data centers in Guiyang-Gui'an alone, including Huawei, Tencent, and the three telecom operators ^4,3. Guizhou's 2024 digital-economy value added exceeded 250 billion yuan, 53.3% of regional GDP, and the province targets 190 EFLOPS and a digital-intelligence industry above 300 billion yuan by end-2026 ^3,4. The sales pitch is unchanged: deployment in Guiyang is about 30% cheaper than the coast, latency to Chengdu-Chongqing is 6 ms and to the Greater Bay Area 10 ms, more than 90% of customers come from outside Guizhou, and utilization exceeds 96% ².

The MIIT announcement complicates the old "control track is flat" reading. Nearly 200 key standards, plus AI ethics-review pilots in unspecified cities, mean governance is being operationalized on the ground. But the same official put standards in the service of industrial integration across R&D, design, production, manufacturing, quality inspection, and operations and maintenance ¹—i.e., coherence for a national supply chain, not a clampdown. No new algorithm or large-model rule appears in this batch; the tightest sectoral mechanism remains the staged pathway for medical AI, from concept to insurance listing ⁵. The Shanghai formula makes the same point: at a municipal Party committee study session, Party Secretary Chen Jining and Shanghai AI Laboratory director Zhou Bowen called for seizing the "new foundation, new paradigm, new industry, new governance" window and building an "innovation hub for AI development and governance," with development kept "safe and controllable" ⁶. Governance is being absorbed into the promotion narrative.

The international layer is also shifting from principles to process. An official English-language media account frames Xi Jinping's 2026 World AI Conference keynote as the basis for turning global AI governance principles into "deeper, more concrete action" ⁷. That is consistent with China selling itself as both an AI adopter and a governance standard-setter. The domestic standards push is the concrete counterpart: nearly 200 standards give China a homegrown baseline it can carry into global forums. The test is whether these standards and ethics-review pilots harden into market-access requirements, especially for foreign AI services and chips.

The real economy continues to transmit the promotion signal. PCB leader Hudian (沪电股份) reported H1 revenue of RMB 13.689 billion, up 61.17%, and net profit of RMB 2.923 billion, up 73.72%, with 32-layer-plus PCB revenue surging 190.83% ⁸. Industry data show PCB exports up 27.63% in H1 2026, with four-layer-plus boards exporting $10.863 billion, up 47.74% ⁸. The human-capital layer is thickening on the mainland too: Jiangsu has built 264 craft academies and is running "AI+技能夜校" (AI+skills night school) at a Shanghai Zhenhua Heavy Industry subsidiary in Nantong, alongside the 1,400 Hong Kong school leaders already trained in AI leadership ^9,10. Fujian is adding factory-floor robotics for precision grinding of power-generation components ¹¹. The build-out is reaching skills and factory floors, not just data centers.

The key uncertainty is whether governance catches up with deployment. MIIT's ethics-review pilots and the standards count are the first signs of a governance layer being assembled; they could become compliance obligations as AI products scale. The concrete tests for outside readers are Guizhou's 190 EFLOPS and 300 billion yuan digital-intelligence target for end-2026 ⁴, and whether the staged-review model now used for medical AI is extended to autonomous driving or embodied intelligence. If it is, the promotion-control balance will have meaningfully moved.