Runjian’s Low-Latency Bet on Southeast Asian Manufacturing
17 Nov 2026 · Stories from AIMX by Industrial Gemba’s Editorial Team Runjian’s Weijian Ou, Marketing Director, and Jake Yin, Vice President of its AI Business Group, on a Guangxi compute...
17 Nov 2026 · Stories from AIMX by Industrial Gemba’s Editorial Team
Quick Skim
Runjian’s Weijian Ou, Marketing Director, and Jake Yin, Vice President of its AI Business Group, on a Guangxi compute centre, 20-millisecond latency, and an ontology-native platform already running at Malaysian manufacturers.
Runjian’s first sales pitch to regional plant operators is not an algorithm. It is geography.
Headquartered in Guangxi, China, the telecommunications and enterprise technology provider operates a proprietary computing hub strategically positioned on Southeast Asia’s doorstep. According to Weijian Ou, Marketing Director, and Jake Yin, Vice President of Runjian’s AI Business Group, this geographic footprint enables cross-border token delivery with latencies as low as 20 milliseconds into manufacturing nodes across ASEAN, including Singapore.
Apollo 11: The Full Stack at the Edge
Layered directly on top of this compute layer is Apollo 11, an enterprise software suite Runjian characterizes as an “ontology-native platform” purpose-built for heavy manufacturing.
Rather than acting strictly as an infrastructure vendor, the company positions itself as an end-to-end partner. The architecture spans the entire operational stack, from the underlying computing hardware and fine-tuned domain-specific large language models down to agentic orchestration engines running automated shop-floor workflows.
High-Volume Latex, Certified Aerospace
While compute metrics draw interest, Runjian’s strongest validation lies in early regional deployments.
The company points to live industrial implementations in Malaysia, specifically across high-volume latex manufacturing and precision aircraft components. Given the stark operational difference between continuous-flow chemical batching and tightly audited aerospace tolerances, these reference sites represent critical proving grounds for Apollo 11’s agentic capabilities.
The Latency vs. Residency Trade-off
Sub-20-millisecond response times make cross-border cloud AI feasible for edge machinery, but distributed architecture creates regulatory friction. Aerospace manufacturers and tier-one industrial suppliers operate under strict non-disclosure regimes and contractual data governance frameworks.
Because the interview left cross-border data routing, model weight storage, and jurisdictional compliance unaddressed, manufacturers must treat legal data residency as a core technical specification rather than an afterthought.
The Gemba Takeaway
Before routing production-line telemetry outside your borders:
- Validate Live Reference Architectures: Request an operational briefing or reference contact from the Malaysian aircraft-parts or latex deployments to see whether Apollo 11 handles mission-critical closed-loop control or advisory analytics.
- Audit Data Residency in Writing: Review the physical route your shop-floor data travels. Cross-reference the compute cluster’s location and edge caching policies against your end-customers’ strict security and export-control requirements before initiating any pilot.
This interview is part of Industrial Gemba’s coverage from AIMX – the region’s AI event featuring real-world use cases, innovation showcases, and strategic insights. Responses have been edited for length and clarity. Watch the full interview on YouTube.


