From Silicon to Intelligence — Defining the New Productive Forces of the Smart Era
联广科技 · Business Plan | June 2026 | Contact: Huang Libo, Director
Chapter 1 · Project Overview
At the Inflection Point of the AI Era
We stand at a historic inflection point — one driven by large AI models and a fundamental restructuring of compute infrastructure. Demand for computing power is growing exponentially, while the supply of advanced-node chips scales only linearly. This supply-demand tension is the most certain investment theme of the next decade.
While the AI race is crowded with algorithm and application players, 联广科技 asks a more foundational question: Who provides the relentless compute "brain" for AI's digital workers and robotic bodies — and who delivers that into real, commercializable value?
The answer is a company that has vertically integrated the entire stack: from compute chips (Brain) → system software (Soul) → physical terminals (Body) → scene applications (Landing). We are not a software company, nor a chip designer alone. We are a full-stack AI new infrastructure and productivity enablement platform.
The Strategic Integration
贵真科技 brings mature AI Digital Employee systems, the VV-BOT robot rental-and-education ecosystem, and Quantum Walkie-Talkie AI social products — proven with thousands of enterprise clients.
联广科技 is born from one of the world's top GPGPU design teams, holding the core technology behind China's only mass-produced full-precision GPGPU chip.
Chapter 2 · Strategic Positioning
Brain, Soul, Body — A Complete Intelligence Architecture
Our core conviction: the true commanding heights of the AI value chain lie in compute chips (Brain) and digital employee operating systems (Soul). Embodied intelligence — humanoid robots, robotic dogs — is the natural execution terminal of that brain and soul in the physical world.
1
🧠 Brain
Full-precision GPGPU — 2nm to 7nm mass production. The compute substrate for everything.
2
👻 Soul
AI Digital Employee Platform — the operating system that determines how AI is invoked and monetized.
3
🤖 Body
VV-BOT embodied intelligence terminals — the physical manifestation, operated via asset-light model.
4
🌐 Landing
Scene applications — enterprise clients, education, social, and HPC verticals generating recurring revenue.
Strategic Priority Matrix
Where We Place Our Bets
Strategic resource allocation is deliberate and disciplined. Maximum investment flows to the two highest-barrier layers — chips and software — while embodied intelligence is pursued in asset-light mode as a showcase and distribution window.
Layer | Our Position | Priority | Rationale |
AI Compute Chip (Brain) & Supply Chain | Proprietary full-precision GPGPU, 2nm–7nm mass production | ★★★★★ | Compute is the "oil" of the AI era. Domestic substitution imperative. Highest technical barrier. |
Digital Employee Platform (Soul) | AI Digital Employee system; Quantum Walkie-Talkie AI social | ★★★★★ | Determines how AI is called and how commercial value is generated — the SaaS revenue engine. |
Embodied Intelligence Terminal (Body) | VV-BOT robot rental, sales & STEM education ecosystem | ★★★☆☆ | Asset-light showcase window for Brain + Soul. No heavy R&D allocation. Natural market extension. |
Chapter 3 · Industry Pain Points
Three Critical Contradictions — and Our Answers
Drawing on 贵真科技's first-hand experience serving thousands of enterprise clients, combined with 联广科技's deep understanding of compute infrastructure, we have identified three structural market failures driving our entire strategy.
Pain Point 1: The Compute Bottleneck — China Needs Its Own NVIDIA
NVIDIA H100-class GPUs are expensive, scarce, and subject to U.S. export controls. Domestic GPU startups have emerged, but none can match full-precision scientific computing (FP64) and AI training (FP32) requirements — except us. HPC applications in meteorology, geological survey, drug discovery, and defense simulation are non-negotiable. Domestic substitution is urgent.
Pain Point 2: The "Last Mile" Problem of AI Application Deployment
贵真科技's Digital Employee system has validated massive market demand across sales, legal, and finance verticals. But achieving 24/7 seamless interaction, processing complex video streams, and handling omnichannel data requires powerful cloud and edge compute. General CPUs and low-precision AI chips cannot balance cost, performance, and versatility simultaneously.
Pain Point 3: Embodied Intelligence — High Valuations, No Commercial Loop
Humanoid robot valuations are sky-high, yet most lack a real compute brain and software ecosystem. The true commercial loop must be: chip provides compute → digital employee provides intelligence → robot body executes tasks → customer pays. Virtually no company on earth can close this entire loop. We can — and do.
Chapter 4 · Solution · Core 1
The Full-Precision GPGPU — China's Only Mass-Produced Compute Brain
Born from 红山微电子, our technical team carries over 20 years of GPGPU design expertise. Core members have led multiple world-class supercomputing chips — including M. Saito, the founding architect of the global MIMD architecture and a multi-time Green500 #1 winner.
Our IA Series chips are production-ready — not a tape-out on paper. They support the full x86 ecosystem, have passed Linpack benchmarking, and are the only full-precision general-purpose GPGPU from China in mass production today. Global peers: just three companies. Chinese peers: only us.
Metric | Our IA Series | NVIDIA A100 | Significance |
Process Node | 2nm–7nm | 2nm–7nm | Parity with global SOTA |
FP64 Performance | 19.2 Tflops | 9.7 Tflops | 2× scientific compute |
FP32 Performance | 38.4 Tflops | 19.5 Tflops | Leading AI training |
Full Precision | ✓ FP64/32/16 | ✓ Supported | HPC + AI coverage |
Mass Production | ✓ Already shipping | ✓ Shipping | China's only option |
Chapter 4 · Solution · Core 2
Digital Employee Platform — The Labor OS of the AI Era
贵真科技's Digital Employee system is not RPA. It is a full-capability AI agent that operates autonomously across entire business workflows — understanding natural language, taking initiative, and collaborating with humans across 50+ job functions including sales, customer service, legal, finance, and HR.
Physical World Presence
Runs autonomously on real devices (phones, computers), interacts with the real world — replacing human roles and delivering measurable cost reduction.
End-to-End Workflow
Handles complete business processes — from content creation to multi-platform publishing — with zero human intervention in the loop.
Omnichannel Capability
Seamlessly bridges phone calls, WeChat, APP messaging, and email — multiplying customer acquisition efficiency and elevating experience.
Full-Scene Coverage
Internal: meeting minutes, AI interviews, employee coaching. External: client acquisition, customer service, sales. Covers the entire enterprise lifecycle.
Traditional RPA executes only pre-programmed single tasks and requires heavy developer involvement. Our Digital Employees independently manage complete business processes, understand intent, and act proactively. Deployment time: 2 hours to 2 days vs. months and millions in traditional development costs.
Chapter 4 · Solution · Core 3
VV-BOT Embodied Intelligence — Asset-Light, Maximum Impact
VV-BOT is the physical manifestation of our Brain + Soul stack — an ecosystem of robot rental, sales, and STEM education that brings our technology into tangible, visible spaces. Critically, we do not invest heavily in developing robot bodies ourselves.
Instead, we integrate leading hardware from Unitree, UBTECH, and other top manufacturers, then empower them with our proprietary chip and digital employee software. This asset-light approach lets us enter the embodied intelligence market with minimal capital risk while concentrating strategic resources on the chip and platform layers.
Store Network Expansion
- Flagship Stores (300㎡, 15+ locations): Shanghai, Chengdu, Wuhan, Hangzhou, Guangzhou, Shenzhen
- Standard Stores (150㎡, 35+ locations): Tier-1 and Tier-2 city rollout
- Pop-Up Stores (50㎡, 100+ locations): High-traffic commercial zones
Revenue streams: robot leasing (with operator), robot sales, STEM education (Scratch + robotics, VEX competition), and in-store experience including coffee robot demonstrations.
Chapter 5 · Team
Academy-Led, Battle-Tested — A Rare Convergence of Talent
Our leadership team combines world-class chip architecture expertise, decade-long AI commercialization track records, and deep academic credentials — a convergence that is exceptionally difficult to replicate.
Zhou Jianwu · Executive Director
Foreign Academician, Russian Academy of Natural Sciences. Professor-level researcher in quantum technology and high-performance computing. 100+ patents, papers, and industry awards. Long-term pioneer in HPC industrialization.
Wang Yi · President
PhD in Computer Science, University of Tokyo. 15+ years in computing industry. One of China's earliest advocates for GPU-accelerated server computing since 2010. Expert in global resource integration and AI/HPC convergence strategy.
M. Saito · Chief Designer
Former Chief Scientist, PEZY. PhD, University of Tokyo. 20+ years chip design, 10+ years as GPGPU Chief Designer. Global MIMD architecture founder. Created world's highest-performance GPGPU chips. Multiple Green500 #1 wins. Founded Envoy AI and PEZY supercomputer.
Huang Libo · Co-Founder & Director
CS graduate, Huazhong University of Science & Technology. Founded 贵真科技 in 2014. Delivered 100+ software products and platforms. Current Xiaohongshu and cross-border ranking systems lead their respective categories. Deep expertise in AI product design, marketing, and channel development.
Chapter 6 · Business Model
Three Revenue Engines — Progressive, Compounding Growth
Engine 1: Chip Sales & IP Licensing (Bedrock, High Margin)
Full-precision GPGPU chip sales targeting NVIDIA A100 displacement. IP licensing across SMIC, UMC, TSMC process nodes from 55nm to 7nm. Quantum simulation cloud platform services leveraging Monte Carlo algorithms and the MIMD architecture. This is the highest-margin, most defensible revenue layer.
Engine 2: Digital Employee SaaS Subscriptions (High Growth, Recurring)
Monthly and annual subscriptions for 50+ digital employee roles — sales, customer service, finance, legal, HR — delivered globally. Fast onboarding: 2 hours to 2 days vs. months and millions via traditional development. Expected renewal rate exceeds 80%, driving high customer lifetime value and predictable ARR growth.
Engine 3: VV-BOT Embodied Intelligence & STEM Education (Stable Cash Flow)
Asset-light store network generating predictable monthly cash flows. Each standard store targets ¥640K annual revenue from robot leasing, toy sales, in-store experiences, and STEM education programs. National expansion to 150+ stores creates scale revenue and brand equity simultaneously.
VV-BOT Unit Economics
Standard Store Annual Revenue Model
600
264
96
48
Single Standard Store (150㎡) Annual Summary
¥768万
Total Annual Revenue
Per standard store, across all four streams
150+
National Target
Flagship, standard, and pop-up stores combined
With 35 standard stores at full operation, the VV-BOT network alone targets over ¥2.6 billion in annual revenue — before flagship and pop-up contributions. This scale also creates a nationwide physical presence that amplifies the chip and software brands.
Chapter 7 · Competitive Moats
Five Layers of Unassailable Competitive Advantage
Technical Barrier
7nm full-precision GPGPU is a national "choke point" breakthrough. We are one of only three companies globally — and the only one in China — with mass-produced full-precision general-purpose GPGPU.
Production Head Start
IA Series chips are already shipping — not a tape-out slide deck. Tape-out risk is resolved. We are in active market deployment, not development stage.
Closed-Loop Ecosystem
From chip ISA optimization → digital employee OS → embodied intelligence terminals: a complete "compute-algorithm-data-scene" loop that no competitor can replicate end-to-end.
Team Scarcity
Global top-tier chip design team (MIMD architecture founder) + proven AI product commercialization team. This combination exists nowhere else in China.
Market Position Lock-In
Simultaneously positioned at: domestic compute sovereignty (national strategy) + AI application deployment (trillion-RMB market) + embodied intelligence ecosystem (future direction).
Chapter 8 · Financials & Fundraising
Financial Projections & Investment Thesis
5-Year Revenue & Profit Forecast (RMB Millions, Base Case)
Total Revenue (¥M)
Net Profit (¥M)
Consolidated model based on chip shipments, Digital Employee SaaS subscriptions, and VV-BOT store expansion. Revenue grows from ¥1.6B in 2026 to ¥2.8B by 2030, with net profit margin expanding as SaaS and chip licensing revenues scale.
This Round: A+ / Pre-B
¥5亿
Raise Target
RMB 500 million
10-15%
Equity Offered
Subject to valuation negotiation
Use of Proceeds
- 40% — Chip mass production & next-gen 3nm/5nm R&D
- 25% — Digital Employee platform, LLM training, ecosystem
- 15% — VV-BOT national store network rollout
- 10% — Quantum Walkie-Talkie & consumer product iteration
- 10% — Global expansion, talent acquisition, working capital
Chapter 9 · Vision
Invest in Us. Invest in the Three Certainties of the AI Era.
Compute is the new oil. Digital employees are the new labor force. Embodied intelligence is simply the natural extension of both.
1
Short Term (1–2 Years)
Consolidate leadership in domestic full-precision GPGPU. Become "China's NVIDIA" — the indispensable compute backbone of AI infrastructure.
2
Mid Term (3–5 Years)
Digital Employee platform serves millions of enterprise clients globally. Become the de facto "Labor Operating System" of the AI era — the platform through which AI is summoned, deployed, and monetized.
3
Long Term (5–10 Years)
Build the complete intelligence ecosystem from chip to digital employee to embodied intelligence. Emerge as a core global supplier of AI foundational infrastructure.
✅ National Strategy
China's compute sovereignty — a state-level imperative with guaranteed demand and policy tailwinds.
✅ Trillion-RMB Market
AI empowerment across every industry — the largest application deployment wave in commercial history.
✅ Future Imagination
Embodied intelligence as the physical world's interface — a frontier market we enter with zero capital risk.
We do not aspire to be the loudest AI company. We are determined to become the most indispensable one — from silicon, to mind, to world.
Modesty will bring Abundance to your life. · Contact: Huang Libo, Director | 联广科技 × 贵真科技 · June 2026
