Data-driven business analysis on AI infrastructure companies, using public financial data, company filings, and market indicators.
A personal analytics portfolio for business analytics preparation: framing questions, structuring public inputs, and planning visuals for AI infrastructure companies. The first company profile collects NVIDIA filing figures on a dedicated page; cross-company comparison remains in development.
Overview
Three lanes—same site, different deliverables.
Long-form notes and frameworks.Systems context and falsifiable written tests of ideas.
Interactive valuation and finance tooling.Hands-on reverse framing and consistency checks.
Structured business analytics projects.Datasets, methodology notes, dashboards, and case studies.
Core project
Compares selected names across growth, profitability, R&D intensity, CapEx intensity, and supply-chain positioning—once annual metrics are collected and reconciled from public sources.
Dashboard preview
Wireframe only. No charts or numbers.
Comparison set and qualitative labels.
Cross-company summary once inputs are aligned.
Time-series view after collected pulls.
Margin panels with shared definitions.
Roles vs qualitative AI exposure mapping.
Industry framework
Industry-level framework for comparing profit capture, capital burden, pricing power, and cycle risk across the AI infrastructure value chain.
Collected company-level evidence nodes that support the broader AI Infrastructure Economics framework and future cross-company comparison views.
Collected Form 10-K data for FY2021–FY2025, including revenue, margins, R&D intensity, CapEx intensity, and initial company-level interpretation.
Status: collected
Collected Form 10-K data for FY2021–FY2024, including revenue, margins, R&D intensity, CapEx intensity, and initial company-level interpretation.
Status: collected
Company universe
Summary columns below; full qualitative fields remain in project JSON.
| Ticker | Company | Category | AI Exposure | Supply Chain Position |
|---|---|---|---|---|
| NVDA | NVIDIA Corporation | Accelerated computing & AI chips | Core AI training and inference hardware and software stack | Fabless chip design; GPU and AI accelerator supply to cloud and enterprise |
| AMD | Advanced Micro Devices, Inc. | CPUs, GPUs, and adaptive computing | Data center GPUs and CPUs for AI workloads; embedded and edge compute | Fabless design; relies on external foundry capacity for advanced nodes |
| TSM | Taiwan Semiconductor Manufacturing Company Limited | Semiconductor foundry | Enables AI silicon production for fabless and IDM customers at scale | Leading-edge and mature wafer fabrication for global chip designers |
| ASML | ASML Holding N.V. | Lithography systems | EUV and DUV systems that define feasible geometry for AI-related leading-edge chips | Critical equipment supplier for leading-edge chip manufacturing |
| AVGO | Broadcom Inc. | Semiconductors and infrastructure software | Custom AI accelerators for hyperscale customers; networking for AI clusters | Diversified semis plus networking and enterprise software attach |
| MU | Micron Technology, Inc. | Memory and storage semiconductors | High-bandwidth memory and dense server memory content for AI training and inference | DRAM and NAND manufacturing; supplies memory into servers and devices |
| ANET | Arista Networks, Inc. | Cloud networking equipment | Network fabrics for AI training clusters and scalable cloud backbones | High-speed data center switching and routing for cloud and enterprise |
| VRT | Vertiv Holdings Co | Critical digital infrastructure and thermal management | Cooling and power systems supporting higher-density AI racks and facilities | Power, thermal, and services for data centers and adjacent infrastructure |
Metric framework
Definitions for the broader comparison framework. Collected company profiles will align metrics under this rule set before cross-company views.
Reported top-line sales for the fiscal period.
Year-over-year change in revenue with aligned fiscal periods.
Gross profit divided by revenue from the same statements.
Operating income divided by revenue before non-operating noise.
R&D expense divided by revenue using reported line items.
Capital expenditure divided by revenue under a single documented rule set.
Methodology note
Names span silicon, equipment, memory, networking, and facility-level roles for comparison design—not for ranking or recommendations.
Focus on growth, profitability, and reinvestment from commonly disclosed statement lines, with one documented rule set before any numbers ship.
Categories and chain roles live in the universe table and follow how each issuer describes its business in public materials.
Work will draw on public financial data, company filings, and disclosures. Fiscal calendars, restatements, and segment definitions differ by issuer, so comparisons need explicit reconciliation and may stay approximate. AI exposure classification is qualitative from public business descriptions and must not be read as a precise quantitative score.
Reference
Compact definitions for planned tables and files.
Planned case studies
Business question
How can selected AI infrastructure companies be compared across growth, profitability, reinvestment intensity, and supply-chain positioning?
Expected output
Interactive dashboard, methodology note, data dictionary, and case study.
Business question
How does hyperscaler capital expenditure flow through the AI infrastructure value chain?
Expected output
Transmission map, exposure table, and selected financial charts.
Business question
Which companies show stronger alignment between AI-related narratives and reported business fundamentals?
Expected output
Metric comparison, classification matrix, and written analysis.
Current status: framework, data structure, and first NVIDIA company profile are in place.
Next step: expand collected annual metrics to the next company in the universe, then build cross-company comparison views.
This project is for educational and portfolio purposes only and does not constitute investment advice.