Industrial research. Quantitative intelligence. Built for the AI era.
InAI Capital turns that depth into two products: early-stage investments in frontier AI, semiconductor, and robotics companies, and published industrial research & quantitative intelligence for the broader stock market, so you invest smarter and earn more.
Direct Investment
We invest along the full path intelligence travels: from the hardware it's computed on, through the software that serves it, to the machines it moves. When the current paradigm saturates, we back the science that builds the next one.
The physical layer of AI performance: accelerator architectures, next-gen HBM memory, and high-bandwidth interconnects that move data to compute faster.
The critical software layer powering AI workloads: LLM serving and inference optimization, data orchestration, and cloud-native developer tooling.
Where intelligence meets the physical world: industrial automation, embodied AI, and intelligent machines that convert bits into motion, force, and precision.
When scaling laws plateau, alpha returns to the lab. Next-generation architectures, novel reasoning paradigms, and fundamental breakthroughs that redefine what's possible.
We don't just study the frontier. We fund it. First-hand access to the companies building tomorrow's infrastructure is the signal source behind every research report and quantitative model we publish.
InAI Capital Advisor
Industrial research builds our proprietary Bit-Watt framework, which defines the stock universe for quantitative intelligence: AI-native data products and institutional technology licensing.
Deep-dive analysis on AI model architecture, GPU economics, US-China dynamics, robotics, and energy infrastructure. The theoretical foundation.
The Bit-Watt framework: mapping coupling points between compute density (Bit) and power density (Watt). The industry knowledge graph that defines our investment universe.
5,190 A-share stocks tracked daily across 150+ quantitative dimensions. Sector overview, company financials, signal optimizer, sentiment intelligence, delivered as interactive dashboards, JSON, and API. US stock coverage coming soon.
5+
Quantitative analytical frameworks
15+
Research reports
50+
Research advisory clients
5,190
A-share stocks tracked
150+
Dimensions per stock
Daily
Automated rebuild
Track Record
Over the last several years, Yumei Dou has served as an expert advisor to global investment banks, sovereign wealth investors, and leading venture and private equity funds.
The clearest framing I've seen of where the China AI stack actually has a moat versus where it's just price competition. We routed our team to her notes.
Yumei has the rare combination of having built AI companies and underwritten them. Her perspective on Ascend migration timelines reframed our hardware exposure.
Numbers-first, ego-second. Exactly the kind of independent analyst we wish more of our coverage looked like.
Selected advisory engagements
Enterprise
Our A-share quantitative system, from theoretical framework to production deployment, is available for technology licensing to private funds and brokerages seeking systematic quantitative capabilities.
Multi-period momentum, chip structure analysis, and institutional flow tracking, with clear mathematical and physical meaning behind every signal.
AI-powered news processing combined with supply-chain value-stock scoring. Sentiment signals that drive factor activation, not just headlines.
9×9 grid-search optimization, per-stock custom signals, W-phase detection, with overfitting markers and statistical confidence scoring.
Automated factor health monitoring, signal quality evaluation, and hypothesis testing framework. The research pipeline behind the signals.
Production-ready bot with sentiment reports, signal alerts, portfolio tracking, and interactive menu system, deployable on your infrastructure.
Complete factor documentation with mathematical derivations, physical analogies, and implementation specifications. Not a black box.
Get in touch
Questions, feedback, media, or partnership inquiries.