Skip to content
Rival
Models
CompareBest ForArena
Lab
Sign Up
Sign Up

We compare AI models for a living. On purpose. We chose this.

@rival_tips

Explore

  • Compare Models
  • All Models
  • Prompt Lab
  • Image Generation
  • Audio Comparison
  • Leaderboard
  • Challenges

Discover

  • Insights
  • AI Creators
  • AI Tools
  • The Graveyard

Developers

  • Developer Hub
  • MCP Server
  • .llmignore
  • Badges
  • Rival Datasets

Connect

  • Methodology
  • Sponsor
  • Partnerships
  • Privacy Policy
  • Terms
  • RSS Feed
© 2026 Rival · Built at hours no one should be awake, on hardware we don’t own
Rival
Models
CompareBest ForArena
Lab
Sign Up
Sign Up

We compare AI models for a living. On purpose. We chose this.

@rival_tips

Explore

  • Compare Models
  • All Models
  • Prompt Lab
  • Image Generation
  • Audio Comparison
  • Leaderboard
  • Challenges

Discover

  • Insights
  • AI Creators
  • AI Tools
  • The Graveyard

Developers

  • Developer Hub
  • MCP Server
  • .llmignore
  • Badges
  • Rival Datasets

Connect

  • Methodology
  • Sponsor
  • Partnerships
  • Privacy Policy
  • Terms
  • RSS Feed
© 2026 Rival · Built at hours no one should be awake, on hardware we don’t own

GPT-5.4 — Mini LBO Underwrite

  1. Home
  2. Models
  3. GPT-5.4
  4. Mini LBO Underwrite

See how GPT-5.4 compares on this challenge:

  • GPT-5.4 vs Claude Sonnet 4.6
  • GPT-5.4 vs Google: Gemini 3.1 Flash Lite Preview
  • GPT-5.4 vs Gemini 3.1 Pro Preview

4 models have responded to the "Mini LBO Underwrite" challenge. Compare their approaches side-by-side on Rival. This response is part of Rival's open dataset of 5,600+ AI model responses.

Rival
Models
CompareBest ForArena
Lab
Sign Up
Sign Up
Back to GPT-5.4

Mini LBO Underwrite

Gpt 5.4's response to the 'Mini LBO Underwrite' challenge.

GPT-5.4
March 5, 2026
Text

?
Prompt

You are underwriting a 5-year LBO for a private equity investment committee. Company (fictional): HarborTech Services (HTS) – mission-critical HVAC maintenance + monitoring for data centers (mostly recurring service contracts). Use only the information below. Do NOT browse the web. This is a fictional exercise. ENTRY (close at end of FY2025) - FY2025 revenue: $900m - FY2025 EBITDA: $120m (13.3% margin) - Purchase price: 12.0x EBITDA - Transaction fees: 2.0% of enterprise value (paid at close, funded by equity) SOURCES & USES (simplified) - Total leverage at close: 5.5x EBITDA - Term Loan: 4.0x EBITDA, 9.0% cash interest, 1.0% annual amortization - Mezzanine: 1.5x EBITDA, 12.0% cash interest + 2.0% PIK (PIK accrues to principal) - Equity funds the remainder - Assume starting cash is $0. - Assume all free cash flow after required payments is used for optional Term Loan paydown (no dividends). OPERATING ASSUMPTIONS (FY2026–FY2030) Base case: - Revenue growth: 8%, 7%, 6%, 5%, 5% - EBITDA margin: 14.0%, 15.0%, 16.0%, 16.5%, 17.0% - Capex: 3.0% of revenue - Change in NWC: 0.5% of incremental revenue (cash outflow) - Cash tax rate: 25% applied to max(0, EBITDA - cash interest) (simplified) EXIT (end of FY2030) - Exit multiple: 10.5x FY2030 EBITDA - Exit fees: 1.0% of enterprise value - Mezzanine (including PIK) is repaid at exit. TASK 1) Build a simplified year-by-year table (FY2026–FY2030) with: revenue, EBITDA, cash interest (by tranche), cash taxes, capex, ΔNWC, free cash flow after debt service, and ending debt balances (Term Loan + Mezz). 2) Compute equity IRR and MOIC. 3) Provide a sensitivity table of equity IRR vs exit multiple (9.5x / 10.5x / 11.5x) and exit-year EBITDA margin (16% / 17% / 18%). 4) List 5 underwriting risks and 5 concrete downside protection levers. Keep it concise, but show the math clearly.

GPT-5.4
Response

gpt-5.4 logo
GPT-5.4Mini LBO Underwrite
gpt-5.4 logo
GPT-5.4

Below is a simplified 5-year LBO model for HarborTech Services (HTS) using only the assumptions provided.


1) Entry, sources & uses

Purchase price

  • FY2025 EBITDA = $120.0m
  • Entry multiple = 12.0x
  • Enterprise Value = 12.0 × 120.0 = $1,440.0m

Fees

  • Transaction fees = 2.0% of EV
  • = 2.0% × 1,440.0 = $28.8m

Uses

  • Purchase of business: $1,440.0m
  • Fees: $28.8m
  • Total Uses = $1,468.8m

Sources

Total leverage = 5.5x EBITDA

  • Term Loan = 4.0x × 120.0 = $480.0m
  • Mezzanine = 1.5x × 120.0 = $180.0m
  • Total debt = $660.0m

Equity funds the rest:

  • Equity = 1,468.8 - 660.0 = $808.8m

2) Operating model and debt schedule (FY2026–FY2030)

Key modeling conventions used

  • Revenue grows per schedule.
  • EBITDA = Revenue × margin.
  • Term Loan:
    • 9.0% cash interest on beginning balance
    • 1.0% annual amortization of original principal = $4.8m/year
    • all residual FCF used for optional paydown
  • Mezzanine:
    • 12.0% cash interest on beginning balance
    • 2.0% PIK on beginning balance, added to principal
  • Cash taxes = 25% × max(0, EBITDA - cash interest)
    where cash interest = Term Loan cash interest + Mezz cash interest
  • FCF after debt service shown after:
    • cash interest
    • cash taxes
    • capex
    • ΔNWC
    • required Term Loan amortization
  • Then that FCF is used for optional Term Loan paydown

Year-by-year model

$ in millions

FYRevenueEBITDATerm cash int.Mezz cash int.Cash taxesCapexΔNWCFCF after debt service*End Term LoanEnd Mezz**
2026972.0136.143.221.617.829.20.418.3456.9183.6
20271,040.0156.041.122.023.231.20.355.4396.6187.3
20281,102.4176.435.722.529.533.10.377.7314.1191.0
20291,157.5191.028.322.934.934.70.388.1221.3194.9
20301,215.4206.619.923.440.836.50.3105.8110.7198.8

* FCF after debt service =
EBITDA - cash interest - cash taxes - capex - ΔNWC - required TL amortization

** End Mezz includes 2.0% PIK accrual each year


3) Supporting math by year

FY2026

  • Revenue = 900.0 × 1.08 = 972.0
  • EBITDA = 972.0 × 14.0% = 136.08
  • Term cash interest = 9.0% × 480.0 = 43.20
  • Mezz cash interest = 12.0% × 180.0 = 21.60
  • Cash taxes = 25% × (136.08 - 64.80) = 17.82
  • Capex = 3.0% × 972.0 = 29.16
  • ΔNWC = 0.5% × (972.0 - 900.0) = 0.36
  • Required TL amortization = 4.8
  • FCF after debt service = 136.08 - 43.20 - 21.60 - 17.82 - 29.16 - 0.36 - 4.80 = 18.34
  • Optional TL paydown = 18.34
  • End Term Loan = 480.0 - 4.8 - 18.34 = 456.86
  • Mezz PIK = 2.0% × 180.0 = 3.60
  • End Mezz = 180.0 + 3.60 = 183.60

FY2027

  • Revenue = 972.0 × 1.07 = 1,040.04
  • EBITDA = 1,040.04 × 15.0% = 156.01
  • Term cash interest = 9.0% × 456.86 = 41.12
  • Mezz cash interest = 12.0% × 183.60 = 22.03
  • Cash taxes = 25% × (156.01 - 63.15) = 23.21
  • Capex = 31.20
  • ΔNWC = 0.5% × (1,040.04 - 972.0) = 0.34
  • FCF after debt service = 55.45
  • Optional TL paydown = 55.45
  • End Term Loan = 456.86 - 4.8 - 55.45 = 396.61
  • End Mezz = 183.60 × 1.02 = 187.27

FY2028

  • Revenue = 1,040.04 × 1.06 = 1,102.44
  • EBITDA = 1,102.44 × 16.0% = 176.39
  • Term cash interest = 9.0% × 396.61 = 35.69
  • Mezz cash interest = 12.0% × 187.27 = 22.47
  • Cash taxes = 25% × (176.39 - 58.16) = 29.56
  • Capex = 33.07
  • ΔNWC = 0.31
  • FCF after debt service = 77.68
  • End Term Loan = 396.61 - 4.8 - 77.68 = 314.13
  • End Mezz = 187.27 × 1.02 = 191.02

FY2029

  • Revenue = 1,102.44 × 1.05 = 1,157.56
  • EBITDA = 1,157.56 × 16.5% = 191.00
  • Term cash interest = 9.0% × 314.13 = 28.27
  • Mezz cash interest = 12.0% × 191.02 = 22.92
  • Cash taxes = 25% × (191.00 - 51.19) = 34.95
  • Capex = 34.73
  • ΔNWC = 0.28
  • FCF after debt service = 88.07
  • End Term Loan = 314.13 - 4.8 - 88.07 = 221.26
  • End Mezz = 191.02 × 1.02 = 194.84

FY2030

  • Revenue = 1,157.56 × 1.05 = 1,215.44
  • EBITDA = 1,215.44 × 17.0% = 206.62
  • Term cash interest = 9.0% × 221.26 = 19.91
  • Mezz cash interest = 12.0% × 194.84 = 23.38
  • Cash taxes = 25% × (206.62 - 43.29) = 40.83
  • Capex = 36.46
  • ΔNWC = 0.29
  • FCF after debt service = 105.75
  • End Term Loan = 221.26 - 4.8 - 105.75 = 110.71
  • End Mezz = 194.84 × 1.02 = 198.74

4) Exit and equity returns

Exit enterprise value

  • FY2030 EBITDA = $206.62m
  • Exit multiple = 10.5x
  • Gross EV = 10.5 × 206.62 = $2,169.54m

Exit fees

  • 1.0% of EV = 21.70m

Net EV after fees

  • 2,169.54 - 21.70 = $2,147.84m

Debt repaid at exit

  • Term Loan = 110.71m
  • Mezzanine incl. PIK = 198.74m
  • Total debt repaid = 309.45m

Equity value at exit

  • 2,147.84 - 309.45 = $1,838.39m

MOIC

  • Initial equity = 808.80m
  • Exit equity = 1,838.39m
  • MOIC = 1,838.39 / 808.80 = 2.27x

IRR

  • 5-year hold, no interim distributions:
  • IRR = (2.2729)^(1/5) - 1 ≈ 17.8%

5) Sensitivity: Equity IRR vs exit multiple and FY2030 EBITDA margin

Assumption for sensitivity:

  • FY2030 revenue held constant at $1,215.44m
  • Only exit-year EBITDA margin changes to 16% / 17% / 18%
  • Debt balances held constant at base-case FY2030 exit debt:
    • Term Loan 110.71m
    • Mezz 198.74m
  • Exit fees remain 1.0% of EV

Step: FY2030 EBITDA under each margin

  • 16% margin: 194.47m
  • 17% margin: 206.62m
  • 18% margin: 218.78m

Formula

  • Equity exit value = EBITDA × Exit Multiple × 99% − 309.45
  • MOIC = Equity exit value / 808.80
  • IRR = MOIC^(1/5) − 1

Equity IRR sensitivity table

FY2030 EBITDA Margin \ Exit Multiple9.5x10.5x11.5x
16.0%12.5%16.0%19.3%
17.0%14.6%17.8%20.8%
18.0%16.6%19.5%22.3%

For reference, corresponding MOICs:

FY2030 EBITDA Margin \ Exit Multiple9.5x10.5x11.5x
16.0%1.80x2.10x2.41x
17.0%1.98x2.27x2.58x
18.0%2.16x2.46x2.75x

6) 5 underwriting risks

  1. Data center spending cyclicality / customer concentration

    • If hyperscaler or colocation customers delay projects or consolidate vendors, recurring growth could slow.
  2. Margin expansion risk

    • Base case assumes EBITDA margin rises from 13.3% to 17.0%; failure to realize pricing, mix, or efficiency gains would pressure returns.
  3. Labor availability and wage inflation

    • Mission-critical HVAC service depends on skilled technicians; labor shortages could raise cost to serve and impair SLAs.
  4. High fixed cash interest burden

    • At close, annual cash interest is $64.8m, plus mezz PIK accretion; if EBITDA underperforms, deleveraging slows quickly.
  5. Exit multiple compression

    • Entry at 12.0x, exit underwritten at 10.5x already assumes some compression; a weaker market could compress further.

7) 5 concrete downside protection levers

  1. Lower entry price / structure discipline

    • Re-trade valuation or require seller rollover to reduce day-1 equity risk and improve downside MOIC.
  2. Reduce mezzanine portion

    • Replace some mezz with equity or cheaper senior debt to lower cash interest + PIK drag and improve resilience.
  3. Contract quality diligence

    • Underwrite only revenue backed by long-term service agreements, renewal visibility, inflation escalators, and termination protections.
  4. Operational value-creation plan tied to labor productivity

    • Route density, remote monitoring utilization, technician scheduling, and procurement savings to protect margin if growth softens.
  5. Mandatory cash sweep / tighter deleveraging

    • Apply all excess cash to senior debt, avoid dividends, and maintain liquidity controls to preserve exit optionality in a downside case.

Bottom line

Base case returns:

  • Equity invested: $808.8m
  • Exit equity value: $1,838.4m
  • MOIC: 2.27x
  • 5-year IRR: 17.8%

For an IC lens: this screens as a reasonable but not exceptional LBO—return profile is acceptable, but it relies on meaningful margin expansion and steady deleveraging to offset relatively full entry valuation and expensive subordinated capital.

Mini LBO Underwrite

Turn this model response into notes, narration, or a short video

Auto NotesTranscriptNarrateVideo Cut

Partner link

This output lives in our open dataset — 5K+ responses across 200 models. Take it. It’s free.

About GPT-5.4

Capabilities

ConversationReasoningCode GenerationAnalysisAgentic Tool Use

Categories

TextCodeMultimodal

Specifications

Provider
Openai
Released
2026-03-05
Size
XLARGE
Context
1,050,000 tokens

Keep exploring

SAME PROMPT

Claude Sonnet 4.6's version

Same prompt, different result

COMPARE

GPT-5.4 vs Google: Gemini 3.1 Flash Lite Preview

Both outputs, side by side

We compare AI models for a living. On purpose. We chose this.

@rival_tips

Explore

  • Compare Models
  • All Models
  • Prompt Lab
  • Image Generation
  • Audio Comparison
  • Leaderboard
  • Challenges

Discover

  • Insights
  • AI Creators
  • AI Tools
  • The Graveyard

Developers

  • Developer Hub
  • MCP Server
  • .llmignore
  • Badges
  • Rival Datasets

Connect

  • Methodology
  • Sponsor
  • Partnerships
  • Privacy Policy
  • Terms
  • RSS Feed
© 2026 Rival · Built at hours no one should be awake, on hardware we don’t own