agentleFS
HHFinAiHHFinAi/Institutional-Growth-Equity-Skills/skills/semi-dcf-modeler/SKILL.md

semi-dcf-modeler

Skill3 starsChanged 7 months agoMITView on GitHub

Expert semiconductor industry equity research and DCF valuation. Use when the user asks to: (1) Value a semiconductor company (equipment, memory, foundry, or fabless), (2) Build a DCF, SOTP, P/BV, or residual income model for semi stocks, (3) Analyze semiconductor industry cycles or cycle positioning, (4) Assess competitive moats and chokepoints (EUV, foundry, HBM), (5) Model geopolitical scenarios (China export controls, Taiwan risk), (6) Analyze 10-K filings for semi companies with industry-specific red flags, (7) Track memory pricing cycles or equipment demand, (8) Compare semi valuations using trading comps or precedent transactions, (9) Calculate fully diluted shares using Treasury Stock Method (TSM) and If-Converted method, (10) Build share dilution bridges and SBC-adjusted FCF models. Covers equipment (ASML, LAM, AMAT), memory (Micron, SK Hynix), foundry (TSMC), and fabless (Nvidia, AMD) using cycle-peak P/E, SOTP, P/BV, residual income, and forward P/E methodologies.

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Important: Remove in-the-money convertibles from debt in EV calculation.

    line 77
  2. Distinctive:

    Always show valuation at both basic and fully diluted shares:

    line 96

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Semiconductor DCF Modeler

This is an institutional-quality research and investment skill.

Expert-level semiconductor industry valuation for institutional investment research.

Workflow Overview

  1. Identify company type - Equipment, memory, foundry, or fabless
  2. Select methodology - Based on company type decision tree
  3. Gather data - Yahoo Finance, SEC filings, SemiAnalysis (high-confidence), other industry sources
  4. Build projections - Segment-specific revenue/margin models
  5. Calculate diluted shares - TSM for options, If-Converted for converts
  6. Apply valuation - Appropriate method with cycle/risk adjustments
  7. Adjust FCF for SBC - Subtract stock-based compensation
  8. Scenario analysis - Bull/base/bear with geopolitical scenarios
  9. Output Excel workbook - Multi-tab model per company type template

Valuation Method Selection

Company TypePrimary MethodKey MetricExamples
EquipmentCycle-Peak P/E25-35x peak EPSASML, LAM, AMAT, KLA
MemorySOTP + P/BV3x core + 6x HBM, 2-3.5x P/BVMicron, SK Hynix
FoundryResidual IncomeCoE 9-11%, g 4-5%TSMC, UMC
FablessForward P/E20-50x based on growthNvidia, AMD, Qualcomm
IDMSOTP or Norm. P/ESegment-specificIntel, TI, Analog Devices

Share Dilution Analysis (Required for All Models)

Dilution Bridge

Basic Shares Outstanding              (10-K/10-Q cover page)
+ RSUs                                (1-for-1 addition)
+ Net New Shares from Options (TSM)   (Treasury Stock Method)
+ Net New Shares from Warrants (TSM)  (Treasury Stock Method)
+ Shares from Convertibles            (If-Converted Method)
+ PSUs                                (Probability-weighted)
= FULLY DILUTED SHARES OUTSTANDING

Treasury Stock Method (TSM) - Options & Warrants

Only include IN-THE-MONEY options (Strike Price < Current Price)

Net New Shares = Options × (1 - Strike Price / Current Price)

Or equivalently:

Option Proceeds = Options × Strike Price
Shares Repurchased = Option Proceeds ÷ Current Price
Net New Shares = Options - Shares Repurchased

If-Converted Method - Convertible Securities

Only include IN-THE-MONEY convertibles (Current Price > Conversion Price)

Dilutive Shares = Principal Amount ÷ Conversion Price

Important: Remove in-the-money convertibles from debt in EV calculation.

SBC-Adjusted Free Cash Flow

Adjusted UFCF = Unlevered FCF - Stock-Based Compensation Expense

Semi SBC Benchmarks:

Company TypeSBC % of Revenue
Equipment3-5%
Memory2-4%
Foundry2-3%
Fabless (mature)8-12%
Fabless (growth)15-25%

Per-Share Value Impact

Always show valuation at both basic and fully diluted shares:

Value per Basic Share = Equity Value ÷ Basic Shares
Value per Diluted Share = Equity Value ÷ Fully Diluted Shares
Dilution Impact = (Basic Value - Diluted Value) / Basic Value

Industry Expert Sources

High-Confidence Sources

SourceWebsiteExpertiseUse Cases
SemiAnalysishttps://semianalysis.comDeep semiconductor supply chain, technology roadmaps, competitive analysis, AI/HPC trendsEquipment demand forecasts, foundry capacity analysis, memory market dynamics, geopolitical supply chain impacts, technology inflection points

Other Industry Experts

ExpertContribution
Dr. Walden RhinesIndustry constants and historical benchmarks

Industry Constants (Dr. Walden Rhines)

ConstantValueApplication
Learning curve decline~30%/yearASP trajectory anchor
Revenue per silicon area~$70/sq.inRevenue cross-check
R&D % of semi revenue~14%Expense validation
EDA % of semi revenue~2%Industry benchmark
Revenue vs. profit correlation0.07Specialization > scale

Reference Documents

DocumentLoad When
references/equipment_valuation.mdValuing ASML, LAM, AMAT, KLA, TEL
references/memory_valuation.mdValuing Micron, SK Hynix, Samsung Memory
references/foundry_fabless_valuation.mdValuing TSMC, Nvidia, AMD, Qualcomm
references/geopolitical_risk.mdChina scenarios, Taiwan risk, 10-K analysis
references/share_dilution.mdTSM, If-Converted, SBC adjustment, dilution bridge
references/quick_reference.mdQuick lookup of multiples, constants, red flags

Excel Model Structure by Company Type

Equipment Companies (12 Tabs)

  1. CoverSheet - Summary, price target, cycle position
  2. Tool_Demand - Customer capex → tool units
  3. Revenue_Model - Units × ASP × Mix
  4. Income_Statement - Full P&L with margins
  5. Cash_Flow - FCF calculation
  6. Balance_Sheet - Key items
  7. Diluted_Shares - TSM/If-Converted dilution bridge, SBC-adjusted FCF
  8. DCF_Valuation - Cycle-peak EPS × Multiple
  9. Cycle_Analysis - Historical P/E, indicators
  10. China_Scenarios - Revenue impact by scenario
  11. Sensitivity - EPS vs Multiple tables
  12. Assumptions - All inputs with sources

Memory Companies (13 Tabs)

  1. CoverSheet - SOTP summary
  2. DRAM_Model - Commodity DDR
  3. HBM_Model - High Bandwidth Memory (AI)
  4. NAND_Model - Flash storage
  5. Revenue_Rollup - Consolidated
  6. Income_Statement - With cost/bit analysis
  7. Cash_Flow - FCF with heavy capex
  8. Diluted_Shares - TSM/If-Converted dilution bridge, SBC-adjusted FCF
  9. SOTP_Valuation - Segment values + net debt
  10. PBV_CrossCheck - LT ROE / CoE
  11. Cycle_Analysis - Spot prices, inventory
  12. Sensitivity - ASP vs Volume tables
  13. Assumptions - Learning curve position

Foundry Companies (12 Tabs)

  1. CoverSheet - RIM-based target
  2. Capacity_Model - By node
  3. Revenue_Model - Capacity × Utilization × ASP
  4. Margin_Model - GM by node
  5. Income_Statement - Customer concentration
  6. Cash_Flow - Massive capex intensity
  7. Diluted_Shares - TSM/If-Converted dilution bridge, SBC-adjusted FCF
  8. RIM_Valuation - Residual Income Model
  9. Geopolitical_Risk - Taiwan scenarios
  10. Customer_Analysis - Apple, Nvidia concentration
  11. Sensitivity - CoE vs Growth tables
  12. Assumptions - Geopolitical probabilities

Fabless Companies (11 Tabs)

  1. CoverSheet - TAM-based target
  2. TAM_Analysis - By end market
  3. Revenue_Model - By segment
  4. Income_Statement - SBC adjustment
  5. Cash_Flow - Asset-light FCF
  6. Diluted_Shares - TSM/If-Converted dilution bridge, SBC-adjusted FCF
  7. DCF_Valuation - Three-stage for high growth
  8. Moat_Assessment - Architecture, ecosystem
  9. Scenarios - Bull/Base/Bear
  10. Sensitivity - Growth vs Multiple
  11. Assumptions - TAM sources

Diluted_Shares Tab Structure

Section A: Basic Shares

Basic Shares Outstanding    | XX.X M | Source: 10-Q Cover [Date]

Section B: RSUs (1-for-1)

Unvested RSUs Outstanding   | X.X M  | Source: 10-K Note X

Section C: Options (TSM)

In-the-Money Options        | X.X M  |
Weighted Avg Strike Price   | $XX.XX |
Current Share Price         | $XX.XX |
Option Proceeds             | $XX.X M|
Shares Repurchased          | X.X M  |
Net New Shares from Options | X.X M  |

Section D: Warrants (TSM)

In-the-Money Warrants       | X.X M  |
Net New Shares from Warrants| X.X M  |

Section E: Convertibles (If-Converted)

Convertible Principal       | $XX.X M|
Conversion Price            | $XX.XX |
In-the-Money? (Y/N)         | Yes    |
Dilutive Shares             | X.X M  |
Treatment in EV             | Equity |

Section F: PSUs

PSUs Outstanding            | X.X M  |
Probability of Achievement  | XX%    |
Expected Shares from PSUs   | X.X M  |

Section G: Dilution Summary

Basic Shares                | XX.X M |
+ RSUs                      | X.X M  |
+ Net Options (TSM)         | X.X M  |
+ Net Warrants (TSM)        | X.X M  |
+ Convertibles              | X.X M  |
+ PSUs                      | X.X M  |
= FULLY DILUTED SHARES      | XX.X M |
Dilution %                  | XX.X%  |

Section H: SBC-Adjusted FCF

($ millions)    | FY26E | FY27E | FY28E | FY29E | FY30E
------------------------------------------------------
Unlevered FCF   | $XXX  | $XXX  | $XXX  | $XXX  | $XXX
SBC Expense     | ($XX) | ($XX) | ($XX) | ($XX) | ($XX)
Adjusted UFCF   | $XXX  | $XXX  | $XXX  | $XXX  | $XXX

Section I: Per-Share Impact

                        | Basic  | Diluted | Impact
--------------------------------------------------
Equity Value ($M)       | $X,XXX | $X,XXX  |
÷ Shares               | XX.X M | XX.X M  | +XX.X%
= Value per Share      | $XX.XX | $XX.XX  | -XX.X%
Current Price          | $XX.XX |
Upside (Diluted)       | XX.X%  |

Key Parameters Quick Reference

Equipment Multiples (at Cycle Peak)

CompanyP/EChina Risk
ASML30-35xMedium
KLA30-32xMedium
LAM28-30xHigh
AMAT26-28xHigh

Memory Parameters

SegmentP/S MultipleP/BV Range
Commodity DDR2-3x1.5-2.5x
HBM5-7x3.0-3.5x
NAND1-2x1.0-2.0x

Foundry RIM Inputs

ParameterTSMCGFUMC
CoE9.2%11.5%10.5%
Terminal g4.5%2.5%2.5%
Terminal ROE20%13%11%

Fabless Forward P/E

CompanyP/E RangeSBC % Rev
Nvidia35-50x12-15%
AMD25-35x8-12%
Broadcom18-25x6-8%
Qualcomm12-18x5-7%

Typical Dilution by Company Type

Company TypeDilution %SBC % Rev
Equipment3-6%3-5%
Memory2-5%2-4%
Foundry2-4%2-3%
Fabless (mature)5-10%8-12%
Fabless (growth)10-20%15-25%

Geopolitical Risk Framework

China Scenario Probabilities

ScenarioProbabilityEquipmentMemory
Status Quo55-60%-5% rev-2% rev
Moderate25-30%-15-20% rev-5-8% rev
Severe10-12%-25-35% rev-10-15% rev

Taiwan Risk Discount

ScenarioProbabilityTSMC Discount
Status Quo70%0%
Elevated20%10-15%
Blockade7%30-40%
Conflict3%50%+

10-K Red Flags

Red FlagThreshold
Inventory days +50% YoYDemand weakness
AR growth > RevenueRevenue quality
Capex < DepreciationUnder-investing
R&D % decliningInnovation slowdown
SBC > 20% OpIncomeExcessive dilution
Dilution % > 15%High option overhang

Output Standards

Excel Formatting

ElementStyleUsage
Blue textInputUser-modifiable
Black textFormulaCalculated
Green textLinkCross-sheet
Yellow fillOutputKey results

Model Checklist

Before delivery:

  • Cycle position documented with indicators
  • Appropriate methodology for company type
  • Diluted shares calculated using TSM/If-Converted
  • SBC-adjusted FCF included
  • Per-share value at basic AND diluted
  • China/Taiwan risk explicitly modeled
  • Learning curve position for ASP assumptions
  • Sensitivity tables for key drivers
  • Bull/Base/Bear with explicit drivers
  • All assumptions sourced

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.