In a regional capital some 2,000 km from China's coastal tech hubs, officials are telling a different AI story — one aimed south, toward the border, not east toward the sea. The pitch is simple: let the rich coastal cities invent the models, let Guangxi bolt them onto factories and farms, and let Southeast Asia be the customer.
Key takeaways
- Guangxi's 2026 Government Work Report (opened 2 February 2026) commits RMB 45 billion (≈ US$6.3bn / HK$49.5bn) over three years to AI-led "new quality productive forces."
- The province reports its AI core-industry output exceeded RMB 89 billion (≈ US$12.5bn) in 2025, with a China–ASEAN AI centre built in 90 days and 208 signed AI projects.
- The model is "R&D in Beijing/Shanghai/Guangzhou, integration in Guangxi, application in ASEAN" (北上广研发+广西集成+东盟应用).
- 2026 targets: 30 AI flagship products, 80 new smart factories and digital workshops, and humanoid-robot output above 5,000 units.
- Stage note: this is an adopted provincial work-plan commitment, not a national law; the three-year outlay is announced, not yet fully disbursed.
The ASEAN angle is the whole point
Most Chinese provinces race to build the smartest model or the biggest data centre. Guangxi cannot win that race against Beijing, Shanghai or Guangdong, so it is competing on geography. It borders Vietnam and sits closest to the ASEAN bloc, and it has spent the past year building institutions for cross-border AI: the China–ASEAN AI Application Cooperation Centre (中国—东盟国家人工智能应用合作中心), built and opened in 90 days, and cooperation on large-model research with eight ASEAN countries.
The province's own results sheet is the evidence it is leaning on. In 2025, Guangxi's AI core-industry output passed RMB 89 billion, it signed 208 new AI industrial projects, and it framed artificial intelligence as a "new business card" for the region. The 2026 plan doubles down.
What RMB 45 billion is supposed to buy
The three-year allocation is not a single fund but a planning commitment spread across the provincial budget and supporting vehicles. The work report pairs the money with concrete, measurable targets for 2026: 30 AI flagship products, 80 new autonomous-region-level smart factories and digital workshops, and humanoid-robot output above 5,000 units. It also promises a Guangxi AI Academy (广西人工智能学院), a Guangxi–ASEAN embodied-intelligence pilot training ground (广西—东盟具身智能中试训练场), and an ASEAN corpus (东盟语料库) of training data.
Crucially, the plan names where AI should land: vertical models and agents for sugarcane, new-energy vehicles and non-ferrous metals, plus applications in agriculture, tourism, education, healthcare and anti-fraud work. These are Guangxi's actual industries, not abstract benchmarks.
Why this model can work
The "北上广研发+广西集成+东盟应用" formula is a recognition of comparative advantage. Coastal hubs have the talent density to train frontier models; Guangxi has the manufacturing base, the border logistics, and the political mandate to be China's ASEAN-facing node. If a model trained in Shanghai can be integrated into a广西 factory and then sold into Vietnam or Indonesia, everyone in the chain captures value.
It also fits Beijing's broader "AI Plus" push, which rewards deployment and usage over leaderboard scores. A border province that turns AI into exported, paid services is exactly the kind of adoption the national plan wants to see.
The caveats
Announcements in government work reports are commitments, not cash in hand. The RMB 45 billion is a three-year planned figure; how much is genuinely new budget versus relabelled existing spending is not spelled out in the report. And Guangxi's industrial base is smaller than its coastal rivals, so execution risk is real.
The plan also leans on a familiar provincial playbook: attracting coastal-headquartered firms to set up integration and application units in Guangxi through the South-AI Centre, rather than trying to grow frontier research locally. That lowers the bar to early results, but it also means the province is betting that being the "integration layer" between Beijing's models and ASEAN's markets is a defensible, durable position — a claim that has yet to be tested at scale. If the bet pays off, Guangxi becomes a conduit; if it does not, the RMB 45 billion buys headline projects without a self-sustaining industry.
Honest limitations
The RMB 45 billion is a three-year planned allocation stated in the work report; I have not traced how much has been budgeted versus actually spent in 2026. The Guangxi government work report is the primary source, corroborated by China News Service, but I treated it as a policy commitment rather than audited expenditure. The RMB 89 billion 2025 AI-core output is a provincial government figure; definitions of "AI core industry" vary and are not independently audited. I did not verify the identities of the eight ASEAN countries cooperating on large-model R&D beyond the government's statement, and the "90 days" for the South-AI Centre is the government's own claim.
What readers can do now
- If you build AI products for Southeast Asian languages, watch Guangxi's ASEAN corpus and the China–ASEAN AI application cooperation centre for data and deployment partnerships.
- If you are an investor, track whether the RMB 45 billion materialises as procurement, subsidies or funds, and which verticals get first calls.
- If you follow China's regional AI map, treat Guangxi as the ASEAN-facing node and compare its approach with Shanghai's finance tilt and Guangdong's manufacturing base.