The SemiAnalysis Foundry Metrics Model helps you understand the fundamental drivers of the foundry market, offering a holistic view by tracking the most critical metrics. Our model details the dynamic foundry segment, providing key metrics including: wafer shipments, revenue (wafer and non-wafer), wafer ASPs, capacity, utilization rates, process node mix (revenue, capacity, ASP, and utilization), wafer starts per month (by node and foundry), margins (gross and operating), capex, revenue by application/region, wafer starts by customer for major foundries, wafer starts by end market, and forecasts through 2030.

Our Foundry Metrics Model cuts through the noise, delivering a decision-ready data book that lets you benchmark every major foundry. The model includes a dedicated focus on industry leader TSMC, providing an in-depth view of its top customers, wafer starts by customer, and its relationship with Apple.

We are also developing derivative models that will complement the Foundry Metrics Model and be available as add-ons. These include a foundry capacity demand model by end market, a detailed industry forecast by vendor/node, top customer data for leading foundries, process node roadmaps, and profit and loss modeling for key leading-edge nodes and fabs. We plan to continually enhance the model’s breadth and depth.

Why This Model?

• Quantify capacity expansions, wafer prices, leading-edge customer base, and mix shifts before they impact earnings.
• Track utilization, node mix, and wafer pricing trends that lead revenue and margin inflections.
• Leverage rich historical data on leading foundries to understand competitive shifts.
• Monitor which companies are driving leading-edge demand at TSMC.
• Get continuous updates on the foundry landscape with insights and expert commentary.
• Reconcile foundry commentary with customers, suppliers, and peers, backed by expert interpretation.

Comprehensive Metric Coverage

Our dataset includes the following metrics for all leading semiconductor foundries.

• Foundry revenue
• Wafer revenue
• Wafer ASP
• Wafer shipments (WSPM) by node
• Wafer revenue/ASP by node
• Non-wafer revenue streams
• Installed capacity by node
• Utilization by node
• Capex
• 20+ years of historical data for the top five companies
• 10+ years of historical data for other foundries
• Quarterly forecasts through 2030
• TSMC-Apple chip: detailed bottom-up and top-down model
• Fabless vs. foundry relationship matrix
• Wafer starts by customer for TSMC and other major foundries
• Capacity by fab location
• Foundry process roadmaps
• China fab coverage (SMIC, Hua Hong, Nexchip, etc.)
• Wafer starts by end market (smartphone, PC, datacenter, auto, IoT, etc.)
• Wafer demand and capacity forecasts by node and by end market
• TSMC top-customer analysis
• TSMC segment analysis (HPC, etc.)
• TSMC node cost modeling
• Samsung internal vs. external wafer starts
• Financial metrics (margins, R&D, depreciation, COGS, inventory, DOI, capex, etc.)
• COGS per wafer (fixed and variable)
• Profit margins (gross and operating)
• Revenue by application
• Revenue by region
• Inventory levels and days
• Methodology & definitions

Company Universe

• TSMC
• Samsung Foundry
• Intel Foundry
• GlobalFoundries
• SMIC
• UMC
• Hua Hong Semiconductor
• Tower Semiconductor
• HLMC
• Powerchip (PSMC)
• Nexchip
• DB HiTek
• VIS
• X-Fab
• SkyWater
• Rapidus
• Tata Electronics
• SK Hynix System IC
• XMC
• Others

We are planning to extend the list based on client feedback and our ability to consistently get the accurate information.

Deliverables & Update Cadence

• Quarterly Excel workbook: Raw tables updated through the latest quarter, plus near-term forecasts.
• Quarterly PDF snapshot: Charts, summaries, and executive insights.
• Interactive dashboard: On the roadmap.
• Regular updates: Market-event addenda and new features (e.g., wafer starts by customer).

Our Differentiators

• Industry expertise: Deep coverage across IP, equipment, foundry, fabless, OSAT, and OEMs enables realistic supply/demand views.
• Triangulated approach: We reconcile foundry commentary with customers, suppliers, and peers via expert interpretation.
• Rich historical data: Quarterly-level datasets for TSMC (from the mid-1990s), Samsung, UMC, SMIC, and GlobalFoundries.

Methodology Snapshot

• Blend proprietary bottom-up wafers-per-tool calculations, conference materials, public filings, customs data, and primary research.
• Analyze fabless revenue, COGS, inventory, and end markets to reconcile utilization and wafer demand by node.
• Track fab build plans and ramp timings to refine capacity forecasts.
• Use SemiAnalysis’ Accelerator and WFE models to inform TSMC leading-edge wafer starts.
• Build bottom-up cost models to estimate TSMC leading-edge gross margin by node.
• Back-test each quarter against reported results and adjust utilization curves.
• Standardize definitions to SEMI norms.