The Simplest Explanation of Valuation You'll Ever Read
(By Someone Who's Still Figuring It Out)
What The Hell Is Valuation Anyway?
Valuation is answering one question: What is something worth?
That's it. Everything else – the DCF models, the EBITDA multiples, the Monte Carlo simulations that nobody actually uses – is just different ways of answering that question.
Valuation isn't science. It's educated guessing with formulas that make it look scientific.
But here's the kicker: nothing has inherent value. Value only exists when someone wants to buy and someone wants to sell. Everything else is educated guessing with Excel.
The difference between a $50M and $60M valuation often comes down to whether someone used 8% or 9% as their discount rate. That's it. That's the whole game.
But here's the thing: even though it's mostly made up, there's a right way and a wrong way to make things up. Today, I'll show you the right way. Or at least, the way that gets you invited back to meetings.
The Two Types That Matter
Project Valuation: What's this specific solar farm/wind project/battery worth?
Think of it as valuing a machine that prints money for 25 years then dies. You know (roughly) how much money it'll print each year. You know (roughly) when it'll break. You know (roughly) how much it costs to keep running. The "roughly" is doing a lot of work here. Project Valuation is Relatively predictable (lies, but we'll get there)
Company Valuation: What's this entire business worth?
This is valuing the machine that builds money-printing machines. Infinitely more complex because now you're betting on:
Machines they've already built (easy part)
Machines they're building (medium part)
Machines they might build (fantasy part)
The people building the machines (chaos part)
The dirty secret: We use the same basic math for both. We just add more rows to the Excel and pretend it's sophisticated.
Why Valuation Matters (And Why It Doesn't)
Valuation matters because it's how deals get done. Someone needs to put a number on something so money can change hands. It's the language of transactions.
Valuation doesn't matter because it's all made up anyway. I've seen the same project valued at $50M by the seller, $30M by the buyer, and they closed at $38M. Neither model was "right" – they just found a number both could live with.
The skill isn't getting the "right" answer. It's building a defensible argument for your answer.
The Time Value of Money (The Only Concept That Actually Matters)
Everything in finance comes down to this: $100 today is worth more than $100 tomorrow.
Why? Three reasons that took me embarrassingly long to internalize:
1. Inflation (The Obvious One) Your $100 today buys 100 $1 tacos. Next year, tacos cost $1.03. Your $100 only buys 97 tacos. You've lost 3 tacos to the universe. This is tragedy.
2. Risk (The Sneaky One) That $100 promised next year might not show up. Company goes bankrupt. Contract gets disputed. Revolution happens. Aliens invade. The future is uncertainty incarnate.
3. Opportunity Cost (The One Everyone Forgets) Your $100 today could be making money. Even in a savings account at 5%, it becomes $105 next year. So accepting $100 next year means losing that $5 you could have made.
This concept breaks most people's brains because we think linearly. Our brains see $100 and think "$100 is $100." Finance says "depends when."
Present Value: The Magic Trick
Present Value (PV) converts future money to today money. Here's the formula I had tattooed on my arm (kidding, it's on a sticky note):
PV = FV / (1 + r)^n
Where:
FV = Future Value (money later)
r = Discount rate (how impatient you are)
n = Time periods (how long you wait)
But something to note specifically is: that little "r" is doing ALL the work. Change it from 8% to 10% and your valuation drops 20%. It's the most important made-up number in finance.
Example:
Someone offers you $1,000 in one year. What's it worth today?
At 5% discount rate: $1,000 / 1.05 = $952
At 10% discount rate: $1,000 / 1.10 = $909
At 15% discount rate: $1,000 / 1.15 = $870
Same $1,000. Three different values.
The discount rate IS the valuation.
Compounding: The Eighth Wonder
Einstein supposedly called compound interest the eighth wonder of the world. He didn't, but the myth persists because it should be true.
Here's the formula going the OTHER direction:
FV = PV × (1 + r)^n
$100 at 10% for 30 years becomes $1,745. Not $400 (which would be simple interest). The extra $1,345 is your interest earning interest earning interest. It's money creating money creating money.
This is why VCs want 25% IRR. At 25% compounded:
Year 5: 3x your money
Year 10: 9x your money
Year 15: 28x your money
It's also why they're usually disappointed.
Discount Rates (Where Everyone Pretends They're Not Guessing)
The discount rate is how impatient you are for money. Higher rate = more impatient = future money worth less.
But really, it's the return you could get doing something else with similar risk. This is where the guessing starts.
The Discount Rate Hierarchy (From Safe to YOLO)
Risk-Free Rate (3-4%) What you'd get from US Treasury bonds. Called "risk-free" because if the US government defaults, we have bigger problems than our Excel models.
I use the 10-year Treasury yield. Currently ~4.5%. Some use 20-year to match project life. Doesn't matter – it's all made up anyway.
Bank Lending Rates (5-7%) What banks charge for senior secured loans. This is their "safe" return because they get paid first if things go wrong.
Add 200-300 basis points (2-3%) to risk-free rate. Banks are cowards, so this is very safe money.
WACC - Weighted Average Cost of Capital (6-10%) The blended cost of debt and equity. Used for established projects. Here's the formula I pretend to understand:
WACC = (E/V × Re) + (D/V × Rd × (1-Tax))
Where:
E = Equity value
D = Debt value
V = E + D (Total value)
Re = Cost of equity
Rd = Cost of debt
Tax = Tax rate (haha good luck)
Reality: I just use 7-8% for operating projects and adjust if someone argues. Nobody has ever asked me to show the WACC calculation. We're all faking it.
Cost of Equity (12-15%) What equity investors demand for established projects. Higher than debt because equity eats first losses.
Theoretically calculated with CAPM:
Re = Rf + β(Rm - Rf)
Where β (beta) measures correlation to market returns, Rm is market return, and nobody actually uses this formula for renewable energy.
Infrastructure Fund Returns (8-12%) Big boring funds buying operating assets. They want predictable, low-risk returns. The Brookfields and Blackstones of the world.
Fun fact: They use 8% discount rates then layer on 70% leverage to juice returns to 15%. Financial engineering at its finest.
Developer/Sponsor Returns (15-20%) What developers want for taking development risk. "Development risk" = everything that can go wrong before a project operates.
This includes:
Permits denied (happens constantly)
Interconnection costs exploding (happens more)
Offtaker backing out (happens most)
Panels arriving broken (personal experience)
VC/Growth Equity Returns (20-30%) What VCs demand because 7 out of 10 investments will fail. They need the 3 winners to cover the 7 losers plus make profit.
Anyone asking for 30%+ returns is either:
Lying
Confused about the difference between levered and unlevered
Actually finding the needle in the haystack deals
Using aggressive leverage assumptions
How to Actually Pick a Discount Rate
After all that theory, here's what I actually do:
Start with asset type:
Operating solar with PPA: 6-8%
Operating merchant: 9-12%
Construction ready: 10-15%
Development stage: 15-20%
Adjust for specifics:
Strong offtaker (Google): -1%
Weak offtaker (startup): +2%
Proven technology: -1%
New technology: +3%
Good location (Texas): -1%
Bad location (Hawaii): +2%
Sanity check:
Would I personally invest at this return?
What are similar deals pricing at?
What's the buyer's cost of capital?
Round to nearest percent
Because false precision is worse than acknowledged uncertainty
Cash Flows (The Numbers That Actually Matter)
Cash flow is money in minus money out. Revolutionary, I know.
But which cash flow you use determines everything. I screwed this up for two years before someone explained it properly.
The Four Types You Need to Know (Plus Two Nobody Talks About)
1. Unlevered Cash Flow (EBITDA)
Revenue
- Operating Expenses
= EBITDA
This is the project's raw earning power. No debt, no capex, no complications. It's what the project earns before financiers take their cut.
Use for: Enterprise value calculations
2. Levered Cash Flow
EBITDA
- Interest
- Principal
= Levered Cash Flow
What's left after paying the bank. This is what equity holders actually get. Usually 30-50% of EBITDA because debt is expensive.
Use for: Equity value calculations
3. Unlevered Free Cash Flow
EBITDA
- Taxes (on unlevered basis)
- Capex
- Change in Working Capital
= Unlevered Free Cash Flow
The theoretical cash available if you had no debt. This is what academics love because it's "pure."
Use for: Academic papers and showing off
4. Levered Free Cash Flow
EBITDA
- Interest
- Principal
- Taxes (actual)
- Capex
- Change in Working Capital
+ Debt Proceeds
= Levered Free Cash Flow
The actual cash you can take out of the business. The only number that really matters.
Use for: Actual investment decisions
5. The "Distributions" Reality: What actually gets paid to equity after:
Required reserve accounts
Debt covenants
Cash sweep mechanisms
Senior/subordinated payment waterfalls
This is usually 50-70% of theoretical free cash flow because banks are paranoid.
6. The "Tax Equity" Nightmare: For US solar projects with tax equity:
Years 1-6: Tax equity gets 99% of cash
Year 6: Flip to 5% tax equity, 95% sponsor
Unless you hit yield targets, then it's different
Unless there's a cash sweep, then it's really different
I've built models with 47 rows just for the tax equity waterfall. It's insanity codified in Excel.
A Real Project Example (With Actual Numbers)
Let me show you with a theoretical 100 MW solar project:
Year 1 Operations:
Revenue: $10,000,000
Operating Expenses: $2,000,000
EBITDA: $8,000,000
Debt Service: $5,000,000
Levered Cash Flow: $3,000,000
Maintenance Capex: $500,000
Free Cash Flow: $2,500,000
Reserve Requirements: $500,000
Distributable Cash: $2,000,000
Tax Equity Takes 99%: $1,980,000
Sponsor Gets: $20,000
Same project, six different cash flow numbers. Pick wrong = wrong valuation.
The worst part? All six are "correct" for different purposes.
The DCF Model (Where Dreams Go to Die)
DCF (Discounted Cash Flow) is just adding up all future cash flows in today's dollars. Simple concept, nightmare execution.
The Basic Formula That Hides All The Complexity
Value = CF₁/(1+r)¹ + CF₂/(1+r)² + CF₃/(1+r)³ + ... + CFₙ/(1+r)ⁿ
Or in Excel (because who does math by hand):
=NPV(discount_rate, cash_flow_range)
But that assumes annual periods. Real projects are monthly. So you need:
=XNPV(discount_rate, cash_flow_values, date_range)
Building Your First DCF (That Won't Completely Embarrass You)
Step 1: Build Your Timeline
Monthly. Always monthly. Annual models are for people who hate accuracy.
Start Date: 1/1/2024
Month 1: =DATE(2024,1,1)
Month 2: =EDATE(Month1,1)
[Copy across for project life]
Step 2: Project Revenue
For solar with a PPA:
Monthly Revenue =
System Size (MW)
× Capacity Factor (%)
× Hours in Month
× PPA Rate ($/MWh)
× Degradation Factor
× Availability Factor
But wait, there's more:
Time-of-delivery factors (power at 2pm ≠ power at 2am)
Seasonal adjustments (winter sun angle sucks)
Curtailment estimates (grid says "no thanks")
Contract price escalators (usually 2% annually)
My first model had one row for revenue. My current template has 47 rows building to revenue. Every row is a place something can go wrong.
Step 3: Project Operating Expenses
The basics everyone includes:
O&M (operations & maintenance): $7-12/kW/year
Insurance: 0.3-0.5% of project value
Land lease: $100-500/acre/year
Asset management: 1-2% of revenue
The stuff people forget:
Property tax (can be 30% of opex)
Inverter replacement reserves
Module cleaning (dusty panels = less power)
Security (copper theft is real)
Community payments (keeping neighbors happy)
Transmission charges
Scheduling coordinator fees
Performance guarantee shortfalls
Step 4: Calculate Tax (Haha Good Luck)
Solar tax is insane. Let’s just take a saner US version (haha. jokes on sane)
30-40% Investment Tax Credit
5-year MACRS depreciation
Bonus depreciation (sometimes)
State credits (sometimes)
Local property tax exemptions (sometimes)
Tax equity partnership structures (always complicated)
I've seen tax sections with 200+ rows. Just for tax.
Step 5: Debt Modeling
If there's project debt:
Monthly Payment = PMT(rate/12, months, -principal)
Interest = Beginning Balance × Rate/12
Principal = Payment - Interest
Ending Balance = Beginning Balance - Principal
But actually:
Construction loan converting to term loan
Different rates for construction vs operation
Cash sweep mechanisms
DSCR maintenance requirements
Balloon payments
Mini-perm structures
Debt models are where junior analysts go to cry.
Step 6: The Actual DCF Calculation
After 500 rows of build-up, the actual NPV is anticlimactic:
=XNPV(DiscountRate, CashFlowRange, DateRange)
That's it. One formula. Everything else is foreplay.
Terminal Value (The Number Everyone Makes Up)
Projects last 25-35 years but PPAs often end at year 20. What's years 21-35 worth?
Option 1: Assume Zero Conservative. Project has no value after PPA.
Option 2: Perpetuity Value
Terminal Value = Final Year Cash Flow / (Discount Rate - Growth Rate)
Aggressive. Assumes project runs forever.
Option 3: Multiple of Final EBITDA
Terminal Value = Year 20 EBITDA × 5-8x
Common. Equally made up.
Option 4: Replacement PPA Model a new PPA at lower rates. Most realistic, most work.
I've seen terminal value be 0% to 60% of project value. It's the wild west of assumptions.
Multiple Valuation (The Lazy Person's DCF That Everyone Uses)
Multiples are shortcuts. Instead of projecting 300 months of cash flows, you multiply something by something.
The Multiples That Matter
$/W (Dollar per Watt)
The industry standard that means nothing and everything.
Current ranges (late 2024):
Operating solar with good PPA: $1.00-1.50/W
Operating solar merchant: $0.60-0.90/W
Construction ready: $0.15-0.30/W
Mid-stage development: $0.05-0.15/W
Early development: $0.01-0.05/W
But $/W ignores everything important:
PPA price (huge impact)
Location (Nigeria ≠ Dubai)
Remaining PPA term
Technology (tracking vs fixed)
Offtaker credit
Interconnection costs
I've seen identical $/W projects have 50% different values when you actually model them.
EV/EBITDA Multiple
What sophisticated people use to feel sophisticated.
Current multiples:
Operating solar/wind: 10-14x
Operating + development platform: 12-18x
Pure development platform: 8-12x
Distressed assets: 4-8x
The dirty secret: EBITDA is manipulated more than a yoga instructor's spine. I've seen:
"Adjusted EBITDA" (adjusted for what? happiness?)
"Run-rate EBITDA" (if everything goes perfectly)
"Pro-forma EBITDA" (pure fiction)
"Normalized EBITDA" (abnormalized from reality)
Price per MWh
Theoretically clean: Value per unit of energy produced.
Value = Annual MWh Production × $/MWh × PPA Years Remaining
Nobody uses this because it requires thinking.
Why Multiples Suck (But We Use Them Anyway)
Multiples assume comparability. But every project is a snowflake:
Location Matters:
Texas project: Good interconnection, bad PPA prices
California project: Bad interconnection, good PPA prices
Hawaii project: Great PPA prices, impossible development
Contract Matters:
Google PPA: 20 years, investment grade, fixed price
Startup PPA: 10 years, questionable credit, complex escalators
Merchant: No PPA, pure price risk, cowboy territory
Technology Matters:
Fixed tilt: Cheap, simple, lower capacity factor
Single-axis tracking: 20% more generation, 30% more complexity
Bifacial panels: 5-10% bonus, 100% more arguments about modeling
Yet we still use multiples because:
They're fast
They're simple
Everyone else uses them
DCF models are too much work
Enterprise vs Equity Value (The Distinction That Still Confuses Me Sometimes)
This concept is simple in theory, nightmare in practice.
The House Analogy That Actually Works
You buy a house:
House price: $500,000 (Enterprise Value)
Your down payment: $100,000
Mortgage: $400,000
Your equity: $100,000 (Equity Value)
If the house appreciates to $600,000:
New enterprise value: $600,000
Mortgage still: $400,000
Your equity now: $200,000
You doubled your equity (100% return) while the house only went up 20%. This is leverage.
The Formula Bridge
Enterprise Value = Equity Value + Net Debt + Preferred Stock + Minority Interest - Cash
But for projects, it's usually simpler:
Enterprise Value = Equity Value + Project Debt
Why People Screw This Up
Scenario 1: The Acquisition "We bought that project for $100 million!"
But did you:
Pay $100M cash for equity? (Equity value = $100M)
Pay $100M and assume $50M debt? (Enterprise value = $150M)
Pay $30M and assume $70M debt? (Enterprise value = $100M)
All three could be described as "$100 million deal." Words matter.
Scenario 2: The Valuation Your model shows project value of $100M.
But is that:
Before debt? (Enterprise value)
After debt? (Equity value)
After debt and tax equity? (Sponsor equity value)
I once saw a buyer and seller agree on "$100 million value" then spend six weeks arguing because one meant enterprise, one meant equity.
The Leverage Effect
This is where it gets fun. Leverage amplifies returns (both ways).
Example Project:
Enterprise value: $100M
Debt: $70M (70% leverage)
Equity: $30M
If project value increases 10% to $110M:
Debt still: $70M
Equity now: $40M
Equity return: 33% (3.3x the 10% increase)
If project value decreases 10% to $90M:
Debt still: $70M
Equity now: $20M
Equity return: -33% (3.3x the 10% decrease)
Leverage is financial cocaine. Great high, terrible comedown.
The Three Metrics Everyone Uses (And Nobody Calculates The Same Way)
IRR (Internal Rate of Return) - The King of Metrics
IRR is the discount rate that makes NPV equal zero. If that made sense, you're lying.
Better explanation: IRR is your annual compound return. 15% IRR means your money grows 15% per year, like a very optimistic savings account.
The Excel Formula:
=IRR(annual_cash_flows) for annual
=XIRR(cash_flows, dates) for monthly (use this one)
What's Actually Good:
Unlevered IRR (before debt):
Operating solar/wind: 6-9%
Solar + storage: 7-10%
Standalone storage: 8-12%
Offshore wind: 9-14%
Levered IRR (after debt):
Conservative structure (50% debt): 12-15%
Normal structure (70% debt): 15-20%
Aggressive structure (80% debt): 20-30%
Insane structure (90% debt): -100% to +100%
The IRR Tricks Nobody Mentions:
Multiple IRRs: If cash flows go negative-positive-negative, Excel might find multiple IRRs or error out. This happens with projects that need major maintenance.
Reinvestment Assumption: IRR assumes you can reinvest distributions at the IRR rate. 20% IRR assumes you have infinite 20% opportunities. You don't.
Time Sensitivity: Quick payback inflates IRR. I've seen 50% IRRs on projects that only return 1.2x. Front-loaded returns look better than they are.
The Monthly vs Annual Thing:
Annual IRR: 15%
Monthly equivalent: 14.2%
Because compounding periods matter
MOIC (Multiple on Invested Capital) - The Sanity Check
How many times you get your money back. Simple. Beautiful. Harder to manipulate.
MOIC = Total Cash Received / Total Cash Invested
What's Good:
<1.0x = You lost money
1.0-1.2x = Why bother?
1.2-1.5x = Acceptable
1.5-2.0x = Good
2.0-3.0x = Great
>3.0x = Check your model
The MOIC/IRR Relationship:
High IRR + Low MOIC = Quick but small returns
Low IRR + High MOIC = Slow but large returns
I'll take 12% IRR with 2.5x MOIC over 20% IRR with 1.3x MOIC every time. But VCs disagree because they have 10-year fund lives.
Payback Period - The Metric for Humans
How long until you get your money back. The only metric normal people understand.
Find when: Cumulative Cash Flow > Initial Investment
Typical Paybacks:
Solar without tax credit: 10-14 years
Solar with 30% ITC: 7-10 years
Wind: 8-12 years
Battery storage: 10-15 years
If >15 years: probably shouldn't do it
The Payback Nobody Talks About: There's simple payback and discounted payback.
Simple: When cumulative cash = investment
Discounted: When cumulative NPV = investment
Discounted payback is 2-3 years longer. Guess which one sellers quote?
Valuing Companies vs Projects (Where It Gets Messy)
Projects are machines. Companies are organisms. One is predictable, one is chaos.
What Makes Companies Harder
Companies have all the project complexity plus:
Multiple Projects at Different Stages
5 operating (easy to value)
3 in construction (medium difficulty)
10 in development (mostly worthless)
50 in "pipeline" (pure fantasy)
Corporate Overhead
Salaries (expensive humans)
Office rent (WeWork recovering)
Legal fees (lawyers always win)
Development costs (90% wasted)
CEO's Tesla (necessary for success)
Growth Expectations
"Hockey stick" projections
"Platform synergies"
"Exponential scaling"
Other lies we tell investors
Human Capital
Key employees who might quit
Relationships that might die
Knowledge that walks out the door
Culture (whatever that means)
The Build-Up Method (How Bankers Do It)
Start with projects and add corporate stuff:
Operating Projects Value
= Operating MW × $/W multiple
= Or EBITDA × 10-12x
Construction Projects Value
= Invested Capital + 10-20% margin
Late Development Value
= MW × $100-200k/MW
Early Development Value
= MW × $10-50k/MW × 20% success rate
Corporate Overhead
= (Annual G&A × Multiple) - negative value
Platform Premium
= 0-50% of above
= For team, systems, relationships, magic
Total Company Value
Real Example:
Operating: 200 MW × $1.2M/MW = $240M
Construction: 100 MW × $1.0M/MW = $100M
Late development: 300 MW × $150k/MW = $45M
Early development: 1,000 MW × $30k/MW × 20% = $6M
Overhead: -$5M/year × 5 = -$25M
Subtotal: $366M
Platform premium (30%): $110M
Total: $476M
But the buyer will say $300M and you'll settle at $380M.
The Comparable Method (How Everyone Else Does It)
Find similar companies that sold recently:
Identify "Comparables"
Similar size (loosely)
Similar geography (vaguely)
Similar business model (arguably)
Recent transaction (within 3 years)
Calculate Their Multiples
EV/EBITDA
EV/MW (operating)
EV/MW (total pipeline)
Apply to Your Company
With "adjustments"
Many adjustments
So many adjustments it's meaningless
Present Range
Low: Worst comparable
High: Best comparable
Recommendation: Whatever supports your argument
Platform Value (The Hand-Waving Premium)
"Platform value" is what we call the extra value beyond projects. It's for:
Tangible Stuff:
Development team (costs money, might make money)
Proprietary technology (usually Excel macros)
Offtaker relationships (one email ≠ relationship)
Pipeline portfolio (90% won't happen)
Operating systems (spreadsheets and hope)
Intangible Stuff:
"Market knowledge" (we know stuff)
"Execution capability" (we've done stuff)
"Strategic positioning" (we exist in good places)
"Growth potential" (we might do more stuff)
"Synergies" (1+1=3, allegedly)
Platform premiums range from 0% (buyer's view) to 100% (seller's view). Usually settle around 20-30%.
The Dark Arts of Adjustment
Every model starts clean and ends up full of "adjustments." Here are the ones everyone makes but nobody admits.
EBITDA Adjustments (Making Numbers Look Better)
"Normalized" EBITDA:
Remove one-time events (there are always one-time events)
Add back excessive owner compensation
Remove non-recurring legal costs
Adjust for "market" management fees
Pro-forma recent acquisitions
I've seen EBITDA double through "adjustments."
Pipeline Adjustments (Accepting Reality)
The Pipeline Discount Ladder:
Identified sites: Value at 5% of developed
Site control: 10%
Interconnection study: 20%
Permits filed: 30%
Permits approved: 50%
PPA signed: 70%
Financing committed: 90%
NTP issued: 100%
Reality: Multiply all percentages by 0.5
Working Capital Adjustments (The Hidden Value Transfer)
Working capital is current assets minus current liabilities. It's the cash trapped in the business.
Sellers say: "Normalized working capital is zero"
Buyers say: "You need 3 months of opex as working capital"
Difference: 5-10% of purchase price
This negotiation happens after price agreement. It's where deals die.
Tax Adjustments (Good Luck)
Tax credits taken vs available
NOLs (tax losses) valuable or worthless?
State credits transferable or trapped?
Depreciation recapture on sale
Tax equity flip timing
I've seen tax adjustments swing value by 30%. Tax is where liberal arts majors go to feel stupid.
How Deals Actually Get Done
After all this theory, here's how valuations actually work in practice:
The Dance
Month 1: Initial Indication
Seller whispers number to banker
Banker adds 30%
Teaser shows "attractive opportunity"
Buyers guess wildly
Month 2: First Round
Buyers submit IOIs (Indications of Interest)
Range from 0.5x to 2x seller expectations
Seller shocked at "market disconnect"
Banker says "we need to educate buyers"
Month 3: Management Presentations
100-page deck explaining why company is amazing
Buyers ask same 47 questions
Nobody changes their price
Seller considers firing banker
Month 4: Second Round
Buyers who survive get data room access
10,000 documents, 7 are useful
Models built on assumptions and prayer
Binding bids submitted
Month 5: Negotiations
Top 2 bidders identified
Extensive negotiation on price
More extensive negotiation on terms
Deal almost dies 3 times
Month 6: Exclusivity
Winner selected
Due diligence begins
Everything wrong discovered
Price reduced 10-20%
Month 7: Documentation
Lawyers make $2M
Purchase agreement is 200 pages
Nobody reads it all
Close or die trying
What Actually Drives Price
From 50+ transactions, here's what matters in order:
Competitive Tension
Multiple bidders = higher price
One bidder = you're screwed
Fake bidders = sometimes works
Seller Desperation
Need money = lower price
Patient seller = higher price
Distressed seller = vultures circle
Buyer's Cost of Capital
Cheap capital = higher price
Infrastructure funds pay more than developers
Strategics pay more than financials
Market Timing
Hot market = multiples expand
Cold market = multiples contract
Timing matters more than fundamentals
Story Quality
Good story = platform premium
Bad story = asset discount
No story = commodity pricing
The actual valuation model? That's just backup for whatever price psychology determines.
The Mistakes Everyone Makes (Including Me, Constantly)
Modeling Mistakes
The Classics:
Wrong cash flow for wrong purpose - Using levered for enterprise value
Double-counting - Tax benefit in price AND returns
Circular references - Interest during construction loops
Sign errors - Negative investment = positive IRR
Date misalignment - Monthly model, annual discounting
The Subtle Ones:
Ignoring reserve requirements - Banks trap cash
Missing working capital - Need money to run business
Forgetting replacements - Inverters die
Assuming perpetual escalation - PPAs have caps
Straight-line construction - Nothing builds evenly
Assumption Mistakes
Capacity Factor Crimes:
Using P50 when you need P90
Ignoring degradation
Forgetting curtailment
Missing availability losses
Using Arizona factors in New Jersey
Cost Disasters:
Forgetting property tax (huge)
Missing interconnection upgrades
Ignoring snow removal (Northeast)
Skipping vegetation management
Assuming O&M never increases
Revenue Fantasy:
Merchant price projections (fiction)
Assuming PPA extension (unlikely)
Banking on capacity payments (volatile)
REC prices staying high (they won't)
Process Mistakes
The Negotiation Errors:
Revealing your model - Never show formulas
Anchoring too low - First number sticks
Negotiating against yourself - Shut up sometimes
Focusing on price only - Terms matter more
Believing projections - They're all wrong
The Diligence Misses:
Not visiting the site - Google Earth lies
Ignoring local opposition - NIMBYs kill projects
Missing permit conditions - Devils in details
Skipping reference checks - People lie
Assuming contract assignment - Often forbidden
The Reality Check
After 8,000 words, here's what actually matters:
The 80/20 of Valuation
For Projects: 80% of value comes from:
PPA rate and term
Operating history
Location quality
Offtaker credit
Everything else is details.
For Companies: 80% of value comes from:
Current EBITDA
Growth trajectory believability
Management team quality
Strategic buyer interest
Everything else is negotiation.
The Numbers That Actually Trade
Operating Solar (2024):
With good PPA: $1.00-1.30/W
Merchant: $0.60-0.80/W
EBITDA multiple: 10-12x
Unlevered IRR: 6-8%
Development Projects:
NTP ready: $150-250k/MW
Late stage: $50-100k/MW
Mid stage: $20-40k/MW
Early stage: Worth nothing
Platform Companies:
Growing fast: 14-18x EBITDA
Stable: 10-14x EBITDA
Struggling: 6-10x EBITDA
Distressed: 4-6x EBITDA
The Meta-Truth About Valuation
Valuation is a social construct. It's a negotiation wrapped in math. The model doesn't determine value – it justifies the value you already decided.
The skill isn't building the perfect model. It's building a model that tells a compelling story supported by defensible assumptions presented with appropriate confidence.
My models are wrong. Your models are wrong. Everyone's models are wrong. The winner is whoever's wrong in the most useful way.
The Bottom Line
After 200+ valuations, here's what I know:
All valuations are wrong - But some are useful
Discount rates are made up - But pick them carefully
Projections are fiction - But make them plausible
Models don't determine value - Markets do
Simple beats complex - If you can't explain it, it's wrong
The real skill isn't technical modeling. It's understanding what drives value in the real world and translating that into numbers that make sense to people with money.
Build your model. Make your assumptions clear. Check your math three times. Then remember it's all made up anyway and negotiate based on leverage, not logic.
That's valuation. It's not complicated – it's just uncertain dressed up as precision. And we're all pretending otherwise.
—S
P.S. - Found an error? There are definitely several. My math is more suggestion than fact. Email me. First person to find a real error gets credit in the correction.
P.P.S. - Yes, I wrote 8,000 words about finance on a Friday evening. Yes, my social life is just Excel and coffee. No, I don't see this changing. Yes, I need hobbies. No, I won't get them.
P.P.P.S. - Next week: "Why Your Financial Model Is Broken (And Mine Is Too)" or whatever spreadsheet disaster I'm debugging by then.
P.P.P.S. - My math might be wrong. Please check it.
The Impostor's Guide to Clean Energy: Where we turn confusion into slightly less confusion, one formula at a time. Come for the math errors, stay for the existential crisis about whether any of this matters.
