Deep analysis of Guangdong Jinpeng Technology (金朋科技): a PLC IoT company with 100+ patents, co-author of China's national PLC standard, currently raising funds, and the ideal hardware backbone for ARLI's AI agents.
Jinpeng Technology is a deep-tech hardware company that transforms ordinary 220V electrical wiring into an internet backbone, no new cables, no Wi-Fi congestion, no signal loss through concrete walls. They are the only Chinese PLC company co-authoring the national standard, and their technology is deployed in national museums, Olympic centers, and hundreds of commercial buildings. But their software — 云影智能 (Cloud Shadow Intelligence), is a simple IoT remote control with zero AI, zero agents, and zero user engagement. This is the gap ARLI fills.
Jinpeng has the hardware moat (100+ patents, national standard co-authorship, 200+ landmark deployments). ARLI has the AI moat (58 models, multi-agent rooms, event log, RAG memory, trust core). Together, they are the only company that can deliver AI-governed buildings on PLC infrastructure. Neither can do it alone. Jinpeng's current app is 0 ratings on Xiaomi Store, they know their software is weak. This is not a cold call. This is a warm introduction to a company that needs exactly what ARLI has.
Jinpeng Technology was founded in November 2006 in Guangzhou by 张国梁 (Zhang Guoliang). The company's trajectory reveals a rare capability: complete business model reinvention.
Founded in Guangzhou. Initial business: internet services.
Began R&D on Power Line Communication (PLC) technology, a 8-year bet before commercialization.
Pivot to internet payments & digital commerce. ¥400 billion total flow over 4 years. Named top-100 taxpayer in Panyu District, Guangzhou.
Complete pivot, exit payments, focus exclusively on PLC IoT products. R&D, design, production, sales, full stack.
Launch of "云影智能 Smart PLC" brand. Batch delivery to industrial, commercial, and residential projects.
Recognized as Guangdong "专精特新" (Specialized, Refined, Special, New) SME. Co-author of national PLC home standard.
Won "2025 Real Estate AI Application Case Study" award. Listed on 36kr pitchhub, actively raising funds.
App v1.10.0 released (July 2026). 200+ projects deployed nationwide.
Jinpeng didn't stumble into PLC, they funded 8 years of hardware R&D with 4 years of fintech profits. The founder took ¥400B of payment flow and converted it into 100+ hardware patents. This is not a startup burning VC money on a thesis. This is a profitable company that chose to bet its own capital on a deep-tech hardware bet. The financial discipline this implies is the opposite of the typical AI startup, and exactly the kind of partner that values ARLI's revenue-generating AI capabilities.
Power Line Communication (PLC) is not a new concept, but Jinpeng's implementation solves the problems that made previous PLC attempts fail: signal noise on 220V lines, interference between phases, and reliability at scale.
| Technology | Range | Penetrates Walls | No New Wiring | Scalable (1000+ nodes) | Latency |
|---|---|---|---|---|---|
| Wi-Fi 6/7 | 30-50m per AP | Poor (concrete kills signal) | ✓ | Limited (AP congestion) | 5-20ms |
| Zigbee | 10-100m mesh | Mesh helps but slow | ✓ | ~200 nodes practical | 30-100ms |
| KNX / RS-485 (wired) | Unlimited | Irrelevant | No — requires dedicated wiring | Unlimited | <1ms |
| PLC (Jinpeng) | Building-wide | Natively — via existing power lines | ✓ — uses 220V wiring | Thousands of nodes | <5ms |
"We turn every power outlet into a network port. Every lamp, every switch, every sensor, already connected. No new cables. No signal loss through concrete."
: Jinpeng Technology core value propositionThe critical insight for commercial deployment: PLC is the only technology that makes retrofit (存量智改) economically viable. In existing buildings, running Ethernet cables means tearing down walls. Wi-Fi means dead zones and AP sprawl. Zigbee means mesh complexity and battery maintenance for thousands of sensors. PLC: the electrician swaps the switch or socket with a Jinpeng-compatible one, 5 minutes, zero dust, zero new wiring.
Jinpeng's product system follows a clean architecture: one PLC brain, three capabilities, N application scenarios.
| Component | Products | Role |
|---|---|---|
| Networking | PLC Smart Host, PLC Router/Server, Isolator | Injects data signal into 220V wiring; routes between phases |
| Interaction | 4-inch Smart Screen Host, Multi-touch Panels, Knob Remote | Physical control interfaces for users |
| Control | PLC Smart Breaker, Smart Socket, 0-10V Controller, Curtain Motor | Executes commands: power on/off, dim, open/close |
| Sensing | Multi-sensor, Temperature/Humidity Sensor | Environmental data collection — feeds AI |
This is the single most important section for ARLI's positioning in the meeting. Jinpeng's app — 云影智能, has 0 ratings on the Xiaomi Store. Version 1.10.0, 212MB. Its description reveals the gap:
| 云影智能 Can Do | 云影智能 CANNOT Do |
|---|---|
| Add and control devices | Natural language commands ("set the house to night mode") |
| Create scenes (scenarios) | AI-driven predictive scenarios based on behavior |
| Remote control of lights | Multi-agent coordination across buildings |
| Device status & alerts | Energy optimization recommendations |
| Share device access with family | Voice-first interaction with fallback reasoning |
| Resource management by role | Continuous learning from user behavior patterns |
So what: Jinpeng's hardware is world-class. Their software is a glorified remote control. This is not a criticism, it's the precise opening for ARLI. Every feature in the "CANNOT Do" column is an ARLI agent capability. Voice control → ARLI natural language agent. Predictive scenarios → ARLI behavior-learning agent. Multi-building coordination → ARLI multi-agent rooms. Energy optimization → ARLI analytics agent.
"Your PLC turns every outlet into a network port. My AI turns every outlet into an intelligent decision-maker. Together, we create the only building that thinks, while your competitors are still selling remote controls."
| Competitor | Technology | Standard Co-Author | AI Layer | Retrofit Compatible |
|---|---|---|---|---|
| Jinpeng 金朋 | PLC (power line) | ✓ National standard | ✗ — THIS IS ARLI'S SLOT | ✓ — native |
| Tuya Smart 涂鸦 | Wi-Fi + Zigbee + Bluetooth | ✗ — ecosystem, not standard | Basic LLM integration | Partial — needs gateways |
| Xiaomi 小米 | Wi-Fi + Zigbee + BLE Mesh | Ecosystem standard only | XiaoAI voice assistant | Partial — needs hubs |
| Huawei 华为 | Wi-Fi + PLC (limited) | HarmonyOS ecosystem | Celix/Pangu models | Limited — new builds only |
| UIOT 超级智慧家 | Zigbee 3.0 | ✗ | ✗ | Partial |
Jinpeng is not just another smart home company. They are a 参编单位, co-author of China's national standard for PLC-based smart home connectivity. This means: (1) competitors must comply with standards Jinpeng helped write, (2) government procurement favors standard co-authors for public projects, (3) the "专精特新" designation opens doors to government subsidies and preferential policies. No other Chinese PLC company has both the patents AND the standard authorship.
ARLI agent as app plugin. Jinpeng keeps 云影智能. ARLI provides an AI layer that users can enable: natural language control, predictive scenarios, energy analytics. Revenue: Jinpeng sells PLC hardware + ARLI subscription bundled. Split: Jinpeng 70% hardware revenue, ARLI 85% of AI subscription revenue. Effort: 4-6 weeks. Risk: low. Upside: limited.
New joint brand: "ARLI + Jinpeng AI Building." A co-branded AI building management platform that runs on Jinpeng's PLC hardware + ARLI's multi-agent system. Sold to property developers, hotel chains, and commercial building operators. Revenue: ¥500-2000/month per building. Split: Jinpeng 55%, ARLI 45%. Effort: 3-6 months. Risk: medium. Upside: significant, creates a new category.
ARLI takes 5-10% equity in Jinpeng in exchange for: (1) exclusive AI partnership rights for 3 years, (2) ARLI's AI infrastructure deployed across all Jinpeng projects, (3) joint go-to-market for "AI building" category. Jinpeng is currently actively fundraising on 36kr, they need both capital AND a strategic technology partner. ARLI's AI capabilities are worth more to them than cash. Effort: 1-3 months. Risk: highest. Upside: generational: ARLI becomes the AI brain for every Jinpeng building in China.
You are being introduced through a friend on the executive team. This is not a cold call. The goal of the first meeting is not to pitch, it's to understand their pain points and establish credibility. Let them talk 70% of the time. Your questions reveal expertise; your silence builds trust.
| # | Question | What You're Learning |
|---|---|---|
| 1 | "Of your 200+ deployed projects, which segment is growing fastest — commercial offices, public buildings, or residential?" | Their revenue concentration. Tells you where to focus the partnership. |
| 2 | "When you hand over a building to the client, who configures the scenes and automations — your team, a systems integrator, or the electrician?" | The "last mile" pain point. If it's a manual process, ARLI solves it directly. |
| 3 | "What features do building owners ask for that you currently can't deliver — voice control, predictive energy management, multi-building dashboards?" | Their product roadmap gap. Every answer is an ARLI feature. |
| 4 | "Your 云影智能 app has deep device control. Have you considered an AI layer — natural language, predictive automation, energy optimization?" | Their AI awareness. Opens the door to pitch ARLI without pitching. |
| 5 | "I see you're raising funds on 36kr. What kind of partner are you looking for — a pure financial investor, or a strategic partner with technology synergies?" | Their openness to Model C. Don't pitch equity — let them tell you what they need. |
| ☠ WHAT DIES | ♾ WHAT SURVIVES |
|---|---|
| Wi-Fi-only smart buildings (dead zones, congestion) | PLC as the backbone — every outlet is a network port |
| IoT apps that are "just remote controls" | AI agents that understand intent and execute autonomously |
| Hardware-only companies (commoditized by Tuya/Xiaomi) | Hardware + exclusive AI partnerships (defensible) |
| One-size-fits-all building automation | Building-specific AI that learns occupant behavior |
| Selling smart home components individually | Selling "AI-governed building" as a complete product |
| Risk | Probability | Impact | Severity | Mitigation |
|---|---|---|---|---|
| Jinpeng builds their own AI | Low | Critical | MEDIUM | They have 0 AI patents. Hardware DNA ≠ AI DNA. Move fast to become their default AI partner before they hire an ML team. |
| Huawei/Xiaomi PLC + AI bundle | Medium | High | HIGH | Jinpeng's standard co-authorship gives them a moat against giants. Plus, Huawei's AI doesn't do multi-agent rooms — ARLI's differentiator. |
| Jinpeng acquired by Tuya/Xiaomi | Low-Med | Critical | MEDIUM | Negotiate exclusivity clause in partnership agreement. If acquired, ARLI's AI rights survive the transaction. |
| Partnership stalls on pricing | Medium | Medium | MEDIUM | Start with Model A (zero integration cost). Prove value → upgrade to Model B/C. |
| Building owners don't want AI | Low | Medium | LOW | Energy savings alone (30%+) pays for AI subscription. ROI case is clear. |
Jinpeng's app has 0 ratings on Xiaomi Store. This is either: (a) nobody uses it (bad, means users don't engage with the software layer), or (b) it's pre-installed and users don't need to rate it (neutral). In either case, this is a signal that their end-user software experience is not their strength. ARLI's AI layer must demonstrably improve this, not just add features, but transform the user experience from "configure scenes manually" to "tell the building what you want."
| Scenario | Buildings | ARLI Monthly Fee/Building | Annual ARLI Revenue | Jinpeng Hardware Revenue (est.) |
|---|---|---|---|---|
| Conservative (Year 1) | 10 | ¥500 | ¥60,000 | ~¥2-5M |
| Base Case (Year 2) | 50 | ¥1,000 | ¥600,000 | ~¥20-40M |
| Optimistic (Year 3) | 200 | ¥1,500 | ¥3,600,000 | ~¥80-150M |
Meet Jinpeng through the warm introduction. Listen 70%, ask the 5 questions. Do NOT pitch yet, establish yourself as a domain expert who understands both hardware and AI.
Follow-up meeting: present a lightweight demo. ARLI agent controlling a simulated PLC device (or a real Jinpeng switch if they provide one). Show, don't tell.
Sign a Letter of Intent for Model A integration (light: ARLI as AI plugin). Define mutual KPIs: number of buildings, user engagement improvement, energy savings.
Pilot deployment: 3-5 buildings with ARLI AI layer. Measure: energy savings, user satisfaction, scene automation rate.
Based on pilot results, negotiate Model B (co-branded platform) or Model C (strategic equity partnership in their funding round).
Jinpeng has spent 15 years and ¥400B of payment profits building the best PLC hardware in China. They co-wrote the national standard. They're deployed in national museums and Olympic centers. But their software is a remote control with zero AI, and they know it. They're currently fundraising, which means they're open to strategic partnerships that increase their valuation. ARLI's AI layer is exactly the kind of technology synergy that justifies a higher valuation. This is not a vendor relationship. This is a partnership between two companies that each own one half of the future of smart buildings.
ARLI membership: the notes from the workshop
One site of yours, measured, with the price for the work in the same document
ARLI, Inc., a Delaware corporation. Questions: hq@arliwork.com.