The X-Ray Machine: Separating Operations From Financing

بِسْمِ اللَّهِ الرَّحْمَٰنِ الرَّحِيمِ

In the name of Allah, the Most Gracious, the Most Merciful

Imagine two hospitals. Both have identical patient outcomes — same survival rates, same recovery times. From the outside, they look equally well-run.

But one hospital owns its building outright and has $50M in the bank. The other has a $200M mortgage and $2M in cash. Their "performance" looks the same on a basic report. But are they really in the same position?

Now imagine you had an X-ray machine that could see inside each hospital's finances. You could see: "This one is genuinely excellent at operations AND has financial safety. That one is genuinely excellent at operations BUT is highly leveraged."

That's what Modified DuPont does. It's the financial X-ray machine.

In Post 3, we learned that NVIDIA and Intel have almost identical leverage ratios (1.80x vs 1.75x). Traditional DuPont says they're similarly leveraged. But NVIDIA has MORE CASH than debt. Intel has MORE DEBT than cash. How can they have the "same" leverage? Because traditional DuPont doesn't separate operating decisions from financing decisions. The X-ray does.

Quick Summary

Traditional DuPont mixes operating and financing — you can't tell WHERE returns come from
Modified DuPont separates them: ROE = RNOA + Financial Leverage Gain
RNOA (Return on Net Operating Assets) = pure operating performance, ignoring all debt/cash decisions
Financial Leverage Gain = Spread x Net Financial Leverage — shows whether borrowing helps or hurts
The big reveal: NVIDIA's RNOA is 119.5% (higher than ROE of 91.5%) because excess cash drags ROE down
Free Cash Flow shows whether profits translate to actual cash — Intel's net income was $1.7B but FCF to equity was NEGATIVE $5B

This is for you if...

You finished Post 3 and wondered "but what if two companies have the same leverage but completely different debt situations?"
You want to evaluate a company's operations independent of how it's financed
You've heard "profitability doesn't equal cash flow" and want to understand why

Building on Post 3

In Is This Company Actually Making Money?, we decomposed ROE using traditional DuPont — margins x turnover x leverage. We noted a key limitation: it mixes operating and financing decisions. Now we fix that.

The X-Ray Analogy

When you go to the doctor with a complaint, they start with basic vitals — temperature, blood pressure, heart rate. That's traditional DuPont. It tells you the headline numbers.

But if the vitals don't explain your symptoms, the doctor orders an X-ray. They want to see INSIDE — separate the bones from the muscles, the organs from the fat.

Modified DuPont is the X-ray. It separates the operating skeleton (how well the business actually runs) from the financing tissue (how it's funded). Two companies can look identical on vitals but completely different on the X-ray.

Why Traditional DuPont Isn't Enough

Problem 1: Net income mixes operating results with interest expense. A company with identical operations but different debt levels will show different margins. If Company A has zero debt and Company B has $10B in debt, Company B pays hundreds of millions in interest — making its net profit margin lower. But the operations are the same.

Problem 2: The financial leverage ratio (Assets/Equity) doesn't distinguish between cash-rich and debt-laden companies. A company sitting on $50B in cash and one with $0 in cash could have the exact same leverage ratio. Both have lots of assets relative to equity — but one's assets are a safety cushion, and the other's are factories funded by debt.

The NVIDIA/Intel example from Post 3

Traditional DuPont Says

NVIDIA leverage: 1.80x

Intel leverage: 1.75x

Conclusion: "Similarly leveraged"

The X-Ray Reveals

NVIDIA: Negative net debt (more cash than debt)

Intel: Positive net debt (more debt than cash)

Conclusion: Completely opposite situations

Traditional DuPont treats them the same because it lumps all assets together. Cash, factories, inventory, patents — all in one bucket. Modified DuPont separates them: what's operating? What's financial? That separation changes everything.

The mental model that makes this click

Think of it this way:

NOA (Net Operating Assets) = the engine. The factories, inventory, receivables — minus operating liabilities like accounts payable.
NOPAT (Net Operating Profit After Tax) = the profit from running the engine. No interest, no financing — just what the operations produce.
RNOA = how efficient the engine is (NOPAT / NOA). Like miles per gallon.
Net Debt = borrowed fuel. How much debt minus how much cash.
Net Cost of Debt = the price of that fuel (the effective interest rate on net debt).
Spread = engine efficiency minus fuel price. If the engine earns 30% and fuel costs 5%, the spread is 25%. Every borrowed dollar creates 25 cents of value.
Financial Leverage Gain = spread x amount of fuel per dollar of owner money.

Think of RNOA as the return on equity for a company funded entirely by equity — no debt, no interest, no financing effects. It's the purest measure of how well the business itself performs.

Quick reference: all traditional DuPont ratios (from Post 3)

Before we go deeper into Modified DuPont, here's the full set of traditional ratios for reference. These are the ratios that Modified DuPont improves upon.

Margin Ratios (Profitability)

Gross Margin = Gross Profit / Sales
Operating Margin = Operating Income / Sales
EBITDA Margin = EBITDA / Sales
Net Profit Margin = Net Income / Sales

Efficiency Ratios (Asset Management)

Working Capital Turnover = Sales / Avg Working Capital
AR Turnover = Sales / Avg Accounts Receivable
Inventory Turnover = COGS / Avg Inventory
AP Turnover = Total Purchases / Avg Accounts Payable
Long-Term Assets Turnover = Sales / Avg Long-Term Assets
PP&E Turnover = Sales / Avg PP&E

Leverage Ratios (Capital Structure)

Leverage Ratio = Total Liabilities / Total Equity
Debt-to-Equity = Total Debt / Total Equity
Interest Coverage = Operating Income / Interest Expenses
Dividend Payout = Cash Dividends / Net Income

Liquidity Ratios (Short-Term Health)

Current Ratio = Current Assets / Current Liabilities
Quick Ratio = (Cash + ST Investments + AR) / Current Liabilities
Cash Ratio = Cash & ST Investments / Current Liabilities

Why Modified DuPont is better: These traditional ratios mix operating and financing decisions. The Modified DuPont framework separates them, so you can see how well the business performs independent of how it's funded.

Decision Card 1

Company A and Company B both have ROE of 25% and traditional leverage of 2.0x. Company A has $10B in cash and $8B in debt. Company B has $1B in cash and $15B in debt. Which traditional DuPont ratio fails to distinguish them?

Net profit margin

Financial leverage ratio

Asset turnover

Correct!

Assets/Equity lumps all assets together — it can't see that Company A's assets include a huge cash cushion ($10B) while Company B's include mostly debt-funded factories. Both could show 2.0x leverage with completely different risk profiles. Net profit margin and asset turnover would actually differ because of different interest expenses and asset compositions.

Not quite

Net profit margin would actually differ (Company B pays more interest, lowering its margin), and asset turnover would also differ (different asset compositions). The financial leverage ratio (Assets/Equity) is the one that fails — it treats all assets equally, whether they're cash cushions or debt-funded factories.

Reclassifying the Balance Sheet

Splitting the Balance Sheet in Two

Imagine you own a restaurant AND an investment portfolio. Your accountant combines everything into one balance sheet. But if you want to know how well the RESTAURANT is doing, you need to separate the restaurant's assets (ovens, inventory, receivables) from the investment assets (stocks, bonds, cash). That's exactly what we're doing with the balance sheet.

The standard accounting equation says:

Standard Equation

Assets = Liabilities + Equity

Now we split both sides into operating and financial components:

Assets = Operating Assets (OA) + Financial Assets (FA)
Liabilities = Operating Liabilities (OL) + Financial Liabilities (FL)

Rearranging: OA - OL = (FL - FA) + Equity

The Foundational Equation

Net Operating Assets (NOA) = Net Debt + Equity

Where NOA = OA - OL, and Net Debt = FL - FA

This is the foundation of Modified DuPont. Everything builds on separating the balance sheet into these two sides: the operating engine and the financing structure.

Why cash is a financial asset, not an operating one

In theory, SOME cash is needed for daily operations — paying suppliers, making change. But in practice, most cash on a balance sheet is EXCESS. It could be used to pay down debt tomorrow.

We treat ALL cash and marketable securities as financial assets. This means they reduce net debt rather than increasing operating assets.

A company with $50B cash and $30B debt has net debt of negative $20B — it could pay off all debt and still have $20B left.

This is why NVIDIA's net financial leverage is negative (-0.25) despite having traditional leverage of 1.80x. When you move $26B of cash from "assets" to "reduces debt," the picture changes completely.

Reclassifying the Income Statement

The same logic applies to the income statement: separate operating income from financing costs.

Net Operating Profit After Tax

NOPAT = Net Income + Net Interest Expense After Tax

Removes the cost of debt from profits — the pure operating result. "What would this company earn if it had zero debt?"

The other key formulas:

Net Interest Expense After Tax = (Interest Expense - Interest Income) x (1 - Tax Rate) — the after-tax cost of net debt
Net Borrowing Cost = Net Interest Expense After Tax / Net Debt — the effective interest rate on net debt

Intel's condensed numbers walkthrough

Intel FY2023 Condensed Balance Sheet:

Operating Working Capital: -$6,872M (negative = current operating liabilities exceed current operating assets)
Net Long-Term Operating Assets: $142,086M
NOA = $135,214M
Net Debt: $25,249M
Equity: ~$109,965M (average)
Net Capital = $135,214M (matches NOA)

Intel FY2023 Condensed Income Statement:

Net Interest Expense After Tax: -$1,005M
NOPAT: $684M

Intel's operations barely generated profit in 2023. Nearly all of its net income ($1,689M) came from non-operating items. When you strip away the financing, the operating engine is barely turning.

How to reclassify any balance sheet

Step 1: Identify financial assets — cash, short-term investments, marketable securities
Step 2: Identify financial liabilities — short-term debt, long-term debt, capital lease obligations
Step 3: Everything else on the asset side = operating assets
Step 4: Everything else on the liability side = operating liabilities
Step 5: NOA = Operating Assets - Operating Liabilities
Step 6: Net Debt = Financial Liabilities - Financial Assets

Key rules:

Deferred taxes go with operating (they're non-interest-bearing long-term liabilities)
Be consistent — if you classify cash as financial on the balance sheet, interest income must be financial on the income statement

The Operating Engine: RNOA

The Pure Operating Return

Return on Net Operating Assets

RNOA = NOPAT / Average NOA

The return you'd get if the company were funded entirely by equity — no debt, no interest, no financing effects.

RNOA can be decomposed further, just like traditional DuPont decomposes ROE:

RNOA Decomposition

RNOA = NOPAT Margin x Operating Asset Turnover

NOPAT Margin = NOPAT / Sales | Op Asset Turnover = Sales / Average NOA

This is the modified version of the margin x turnover breakdown from traditional DuPont. The difference? NOPAT margin strips out interest expense, and operating asset turnover uses NOA instead of total assets (removing cash and financial assets from the denominator).

Metric	Intel	NVIDIA	TSMC
NOPAT Margin	1.3%	48%	36.9%
Operating Asset Turnover	0.43	2.49	0.85
RNOA	0.5%	119.5%	31.5%

For every $100 of net operating assets, Intel generates 50 cents of operating profit. NVIDIA generates $119.50. That's a 239x difference in operating efficiency.

Compare to traditional DuPont: Intel's ROE was 1.6%, suggesting SOME profitability. But RNOA reveals even that 1.6% was partly from financing, not operations.

Why NVIDIA's operating asset turnover is so high

NVIDIA is fabless — it doesn't own factories. That means minimal PP&E (property, plant & equipment) and minimal operating assets overall. Its net operating assets are mostly working capital and some IP.

Sales of $60.9B divided by relatively small NOA = massive turnover ratio (2.49).

TSMC has huge NOA (fabs worth tens of billions) leading to lower turnover despite strong sales. Intel has huge NOA AND lower sales, creating the worst turnover.

Key insight: Operating asset turnover is different from total asset turnover (Post 3) because it excludes cash and financial assets from the denominator. This is why NVIDIA's number jumps from 1.04 (total) to 2.49 (operating) — all that cash is removed from the denominator.

Decision Card 2

Company X has RNOA of 25%. Its ROE is 35%. Company Y has RNOA of 25%. Its ROE is 15%. Both have identical operating performance. What explains the difference in ROE?

X has positive leverage gain (borrowing helps), Y has negative gain (excess cash drags ROE down)

Company X is more profitable from operations

Company Y has worse margins

Correct!

Same RNOA (25%) means identical operating performance. The gap is entirely from financing. Company X borrows and earns more on operations than it pays in interest — leverage BOOSTS ROE to 35%. Company Y has more cash than debt — the cash earns less than operations, DRAGGING ROE below RNOA to 15%. This is exactly the NVIDIA/Intel dynamic.

Not quite

Both companies have the same RNOA (25%), which means their operating performance is identical. The difference comes from how they're financed. ROE = RNOA + Financial Leverage Gain. X's leverage gain is positive (+10%), Y's is negative (-10%). Operations are the same; financing creates the ROE gap.

The Financial Amplifier: Leverage Gain

How Borrowing Creates (or Destroys) Value

The Full Modified DuPont Formula

ROE = RNOA + Financial Leverage Gain

Financial Leverage Gain

FLG = Spread x Net Financial Leverage

Spread = RNOA - Net Borrowing Cost | Net Financial Leverage = Net Debt / Avg Equity

The Spread is the key. If you earn 30% on operations (RNOA) and pay 5% on debt (net borrowing cost), the spread is 25%. Every borrowed dollar creates 25 cents of value for shareholders.

If you earn 0.5% on operations and pay 4% on debt, the spread is -3.5%. Every borrowed dollar DESTROYS 3.5 cents of value.

Net Financial Leverage can be:

Positive when company has more debt than cash — traditional leverage
Negative when company has more cash than debt — "anti-leverage"

The interaction creates four scenarios:

Positive spread + positive leverage = leverage boosts ROE above RNOA
Positive spread + negative leverage = leverage drags ROE below RNOA
Negative spread + positive leverage = leverage destroys value (borrowing costs exceed operating returns)
Negative spread + negative leverage = excess cash partially offsets poor operations

FY2023	Intel	NVIDIA	TSMC
RNOA	0.5%	119.5%	31.5%
Net Borrowing Cost	-5.2%	6.6%	6.1%
Spread	5.7%	112.9%	25.4%
Net Financial Leverage	+0.18	-0.25	-0.21
Financial Leverage Gain	+1.0%	-28.1%	-5.4%
ROE	1.6%	91.4%	26.1%

Why excess cash drags ROE down

NVIDIA has more cash ($25.98B+) than debt, making net debt NEGATIVE. Negative net debt x positive spread = NEGATIVE financial leverage gain.

In plain English: NVIDIA is earning a low return on its cash (treasury bills, short-term investments — maybe 4-5%) compared to what it earns on operations (119.5%). That idle cash dilutes the overall return to shareholders.

If NVIDIA used that cash to buy back shares: equity goes down, cash goes down, net debt becomes less negative or turns positive, and ROE would INCREASE.

NVIDIA could literally make its ROE even higher by spending its cash pile. The fact that it doesn't means management values having financial flexibility — a safety cushion, an acquisition war chest — over maximizing a ratio.

From the source material: "NVIDIA and TSMC reserved abundant financial resources that they could potentially invest in growth opportunities or conduct stock buybacks."

In FY2021, NVIDIA spent almost $2 billion repurchasing stock — a common way of returning profits and boosting ROE.

The NVIDIA 2019-2021 paradox

In 2019, NVIDIA's ROE was 26% and its RNOA was 68%. In 2020, ROE rose to 30% while RNOA actually fell to 53%. ROE went UP while RNOA went DOWN.

How? NVIDIA took on ~$5B in additional debt, increasing net debt and shifting net financial leverage. The financial leverage gain improved enough to more than offset the RNOA decline.

By 2021, RNOA was 66% (slightly lower than 2019's 68%). But ROE was 45% — much higher than 2019's 26%.

The entire ROE improvement from 2019 to 2021 was from FINANCING decisions, not operating improvements.

Traditional DuPont would show you ROE rising and you'd conclude "great, operations are improving!" Modified DuPont reveals: operations were flat. Financing drove the change.

When should excess cash concern you?

Excess cash is NOT always bad — it provides safety, acquisition flexibility, and R&D runway. But excess cash earning 2-4% when operations earn 30%+ means shareholders are "subsidizing" the safety cushion.

The question is: does management have a PLAN for the cash?

Upcoming acquisition? (Reasonable to hold)
Building new fabs/capacity? (Reasonable to hold)
No plan, just accumulating? (Activists may demand buybacks/dividends)

Intel's situation is different: positive net leverage but low RNOA. Borrowing barely helps because operations don't generate enough return to significantly exceed borrowing costs. The spread of 5.7% is thin — any operational deterioration could make borrowing value-destructive.

Decision Card 3

A company has RNOA of 40%, net borrowing cost of 5%, and net financial leverage of 0.5. What is its ROE? And what happens if it uses excess cash to pay down all debt (net leverage goes to 0)?

ROE = 57.5%. Without leverage, ROE drops to 40%.

ROE = 40%. Paying down debt doesn't change ROE.

ROE = 45%. Paying down debt increases ROE.

Correct!

FLG = (40% - 5%) x 0.5 = 17.5%. ROE = 40% + 17.5% = 57.5%. If leverage goes to 0, there's no financial leverage gain, so ROE = RNOA = 40%. Leverage is amplifying returns by 17.5 percentage points — but it comes with risk.

Not quite

The math: Spread = RNOA - Net Borrowing Cost = 40% - 5% = 35%. FLG = Spread x Net Leverage = 35% x 0.5 = 17.5%. ROE = RNOA + FLG = 40% + 17.5% = 57.5%. Without leverage (net leverage = 0), FLG = 0, so ROE equals RNOA = 40%.

Company A vs Company B: The Cash Illusion

Same ROE, Different Reality

Consider two fictional retail companies with nearly identical financial statements. One key difference: Company B holds much more cash.

Traditional DuPont says both have ROE of ~55%. The ratios look nearly identical. "Same company, right?"

Modified DuPont tells a different story:

Company A (Less Cash)

Higher net debt (less cash)
Higher net financial leverage
Higher financial leverage gain
RNOA slightly lower

Company B (More Cash)

Lower net debt (more cash)
Lower net financial leverage
Negative financial leverage gain
RNOA higher (smaller NOA denominator)

Traditional DuPont says they're twins. Modified DuPont says they have completely different financial structures.

The Buyback Scenario

Now Company B uses its excess cash to buy back shares. Treasury stock increases, cash decreases, and equity decreases.

The result?

ROE jumps from 55% to 68% — a 13-point increase
RNOA stays the same — operations didn't change at all
The entire ROE improvement came from financial engineering (spending cash, reducing equity)

This is why some investors pressure companies to do buybacks. It boosts ROE without improving operations.

Why companies do buybacks

Reduce shares outstanding — earnings per share goes up even if total earnings don't
Return excess cash — alternative to dividends, often more tax-efficient
Signal confidence — management believes the stock is undervalued
Improve ROE — as we just saw, reducing equity mechanically boosts ROE

BUT: If funded by debt (not excess cash), buybacks increase financial risk. Taking on debt to buy back shares is a bet that operations will continue generating enough return to cover the new interest costs.

NVIDIA example: $2B in buybacks in FY2021 — using cash it didn't need for operations. A low-risk way to return value to shareholders and reduce the ROE drag from excess cash.

The Semiconductor X-Ray: Three Companies Unmasked

What Modified DuPont Reveals

NVIDIA — "Operations Are Even Better Than You Think"

RNOA 119.5% vs ROE 91.4% — cash is DRAGGING ROE down by 28 points
Negative net financial leverage (-0.25) = more cash than debt
Spread of 112.9% = enormous — operations vastly outperform borrowing costs
If NVIDIA deployed its cash into operations or buybacks, ROE could exceed 119%. The 91.5% from traditional DuPont understates NVIDIA's operational excellence.
Implication: massive financial flexibility — could acquire companies, increase R&D, or return cash to shareholders at any time

TSMC — "Quietly Cash-Rich"

RNOA 31.5% vs ROE 26.1% — also dragged down by excess cash (FLG = -5.4%)
Negative net financial leverage (-0.21) = also more cash than debt
TSMC's RNOA has steadily increased with a dip in 2023 — the operating engine is gaining strength
Strong spread (25.4%) = operations comfortably exceed borrowing costs
Conservative financial management = stability and long-term strength

Intel — "The X-Ray Reveals the Truth"

RNOA 0.5% vs ROE 1.6% — leverage is propping up ROE slightly
Positive net financial leverage (+0.18) = more debt than cash
Spread of only 5.7% = operations barely exceed borrowing costs
Intel's 1.6% ROE looked bad in Post 3. RNOA of 0.5% reveals it's actually worse. The little bit of ROE that exists is partly financial leverage, not operational excellence.
FLG of +1.0% = the only reason ROE is 1.6% instead of 0.5%

	NVIDIA	TSMC	Intel
Traditional DuPont
Net Margin	48.8%	38.4%	3.1%
Asset Turnover	1.04	0.47	0.26
Leverage	1.80x	1.44x	1.75x
Modified DuPont (The X-Ray)
NOPAT Margin	48%	36.9%	1.3%
Op Asset Turnover	2.49	0.85	0.43
RNOA	119.5%	31.5%	0.5%
Spread	112.9%	25.4%	5.7%
Net Fin Leverage	-0.25	-0.21	+0.18
FLG	-28.1%	-5.4%	+1.0%
ROE	91.4%	26.1%	1.6%

Historical trends (2018-2023): How these companies evolved

The FY2023 snapshot is revealing, but the trend tells the real story. Here's how each company's Modified DuPont ratios evolved:

NVIDIA: The Financing Paradox

From 2019 to 2020, ROE increased from 26% to 30%. But RNOA decreased from 68% to 53%. The ROE improvement came entirely from financing — NVIDIA took on ~$5B in additional debt, increasing net debt and boosting leverage gain.
By 2021, RNOA recovered to 66% (still below 2019's 68%), but ROE jumped to 45% — a massive change from financing effects, not operational improvement.
By 2023, the AI boom pushed RNOA to 119.5%, dwarfing all historical levels. Operations finally drove the story.
Key lesson: Without Modified DuPont, you'd think NVIDIA's 2019-2021 improvement was operational. It was almost entirely financial engineering.

TSMC: Steady Operational Gains

RNOA steadily increased through the period, reflecting growing manufacturing efficiency and economies of scale — exactly what you'd expect from a dominant foundry.
The slight dip in 2023 warrants investigation — likely tied to cyclical semiconductor demand normalization after the pandemic-era boom.
Consistently negative net financial leverage throughout — always more cash than debt. Conservative, stable financing approach.
Key lesson: TSMC's ROE growth was genuinely operational — the opposite of NVIDIA's 2019-2021 pattern.

Intel: The Accelerating Decline

ROE declined significantly over the prior two years leading to 1.6% in 2023 — a company in deepening operational trouble.
RNOA of 0.5% in 2023 reveals the decline is operational, not financial. The leverage gain (+1.0%) is actually the only thing keeping ROE from being even lower.
Intel is the only company with positive net financial leverage — more debt than cash — meaning it relies on borrowed money, unlike its cash-rich competitors.
Key lesson: Traditional DuPont showed Intel's ROE at 1.6%. Modified DuPont reveals the underlying operations generate only 0.5% return — the rest is leverage artificially propping up the number.

Profitability Does Not Equal Cash Flow: The Final Piece

Why Profitable Companies Can Run Out of Cash

Imagine you're a freelancer. You completed $50,000 worth of projects this month. Your income statement says $50,000 in revenue. But your clients haven't paid yet — your bank account has $2,000. You're "profitable" on paper but can't pay rent.

That's the difference between profitability and cash flow.

Two types of accruals cause the disconnect:

Current accruals: Selling on credit. Revenue is recognized but cash hasn't arrived yet (accounts receivable increases). Or buying inventory that hasn't been sold yet.
Long-term accruals: Depreciation and amortization. You recognize an expense this period, but the cash was spent when you originally BOUGHT the asset — possibly years ago.

An analyst should carefully examine the reasons for the difference between earnings and cash flows since it has clues to the future performance of the company.

Free Cash Flow — Two Flavors

Free Cash Flow to Capital

FCF to Capital = NOPAT - Change in OWC - Change in Net LT Assets

Cash available for interest payments, capital investments, and shareholder distributions. Uses NOPAT (operating perspective).

Free Cash Flow to Equity

FCF to Equity = Net Income - Change in OWC - Change in Net LT Assets + Change in Net Debt

Cash available specifically to equity holders. Uses Net Income (includes financing) and adds back net borrowing.

Intel's FCF Calculation (FY2023)

Let's walk through the actual numbers:

Intel FY2023 — FCF to Equity Calculation

Net Income $1,689M

+ Change in Operating Working Capital +$1,165M

- Change in Net LT Assets (massive capex!) -$19,368M

+ Change in Net Debt (borrowed more) +$11,524M

= FCF to Equity -$4,990M

Intel invested $19.4B in long-term assets (new fabs for IDM 2.0) — far more than it generated from operations. It funded this by borrowing $11.5B more. Even after all that borrowing, equity holders were still $5B underwater.

Why cash flow matters more than profitability for survival

Key questions free cash flow answers:

Can the company meet interest payments on debt?
Can it finance growth internally, or must it raise external capital?
Does it generate cash in excess of operating and capital expenditure needs?

"The greater the free cash flow the company generates, the greater its capacity to reward shareholders, service debt, and invest in growth opportunities."

Intel's situation: Despite positive net income, it MUST borrow or raise equity to fund operations. This is sustainable only if the investments (new fabs) eventually generate returns. IDM 2.0 is a multi-year bet — the cash burn today is an investment in future capacity.

Contrast with NVIDIA: Fabless model means minimal capex. With $29B+ in NOPAT and modest capital needs, NVIDIA likely generates massive positive free cash flow. It's accumulating cash faster than it can deploy it — hence the growing cash pile.

How to use free cash flow in analysis

Compare FCF to Net Income — large gaps signal either heavy investment or potential red flags (revenue recognized but cash not collected)
Negative FCF isn't always bad — growth companies often have negative FCF during expansion (Amazon was FCF-negative for years)
But PERSISTENT negative FCF without revenue growth = unsustainable — the company is burning cash without building future capacity
FCF is used in valuation — the Discounted Cash Flow (DCF) model projects future free cash flows and discounts them to present value

"When projected into the future, free cash flows play a critical role in determining the value of a firm." This is the bridge from financial analysis to valuation — a topic for a future post.

Decision Card 4

A company reports $2B in net income. It invested $15B in new long-term assets, working capital worsened by $1B (more cash tied up), and it borrowed an additional $8B. What's its FCF to equity?

+$5B — investing for growth

-$6B — profitable but burning cash

-$2B — slight shortfall

Correct!

FCF to Equity = $2B - $1B - $15B + $8B = -$6B. The company is profitable on paper ($2B net income) but deeply cash-negative. It's investing heavily in long-term assets ($15B) and even with $8B in new borrowing, equity holders are $6B underwater. Sustainable only if those investments generate future returns.

Not quite

The formula: FCF to Equity = Net Income - Change in OWC - Change in Net LT Assets + Change in Net Debt = $2B - $1B - $15B + $8B = -$6B. Despite being "profitable," the company is burning $6B in cash. The massive capex ($15B) dwarfs everything else.

Cheat Sheet: Modified DuPont & Free Cash Flow

Modified DuPont

ROE = RNOA + FLG
RNOA = NOPAT / Avg NOA
RNOA = NOPAT Margin x Op Turnover
FLG = Spread x Net Fin Leverage
Spread = RNOA - Net Borrowing Cost

Key Concepts

NOA = Op Assets - Op Liabilities
Net Debt = Fin Liabilities - Fin Assets
NOPAT = NI + Net Int Exp After Tax
- net leverage = more cash than debt
+ leverage + + spread = ROE > RNOA

Three Semiconductor Stories

NVIDIA RNOA 119.5%, ROE 91.4% (cash drag -28.1%)
TSMC RNOA 31.5%, ROE 26.1% (cash drag -5.4%)
Intel RNOA 0.5%, ROE 1.6% (leverage boost +1.0%)
Intel FCF profit $1.7B, FCF -$5B

Was this helpful?

Your feedback helps me improve these guides

The X-Ray Machine Separating Operations From Financing

The X-Ray Analogy

Why Traditional DuPont Isn't Enough

The NVIDIA/Intel example from Post 3

Traditional DuPont

Modified DuPont

Reclassifying the Balance Sheet

Splitting the Balance Sheet in Two

Traditional Balance Sheet

Condensed Balance Sheet

Reclassifying the Income Statement

The Operating Engine: RNOA

The Pure Operating Return

The Financial Amplifier: Leverage Gain

How Borrowing Creates (or Destroys) Value

Company A vs Company B: The Cash Illusion

Same ROE, Different Reality

The Buyback Scenario

Before Buyback

After Buyback

The Semiconductor X-Ray: Three Companies Unmasked

What Modified DuPont Reveals

NVIDIA — "Operations Are Even Better Than You Think"

TSMC — "Quietly Cash-Rich"

Intel — "The X-Ray Reveals the Truth"

Profitability Does Not Equal Cash Flow: The Final Piece

Why Profitable Companies Can Run Out of Cash

Free Cash Flow — Two Flavors

Intel's FCF Calculation (FY2023)

Practice Mode

Modified DuPont

Key Concepts

Three Semiconductor Stories

Was this helpful?

The X-Ray Machine Separating Operations From Financing

The X-Ray Analogy

Why Traditional DuPont Isn't Enough

The NVIDIA/Intel example from Post 3

Traditional DuPont

Modified DuPont

Reclassifying the Balance Sheet

Splitting the Balance Sheet in Two

Traditional Balance Sheet

Condensed Balance Sheet

Reclassifying the Income Statement

The Operating Engine: RNOA

The Pure Operating Return

The Financial Amplifier: Leverage Gain

How Borrowing Creates (or Destroys) Value

Company A vs Company B: The Cash Illusion

Same ROE, Different Reality

The Buyback Scenario

Before Buyback

After Buyback

The Semiconductor X-Ray: Three Companies Unmasked

What Modified DuPont Reveals

NVIDIA — "Operations Are Even Better Than You Think"

TSMC — "Quietly Cash-Rich"

Intel — "The X-Ray Reveals the Truth"

Profitability Does Not Equal Cash Flow: The Final Piece

Why Profitable Companies Can Run Out of Cash

Free Cash Flow — Two Flavors

Intel's FCF Calculation (FY2023)

Practice Mode

Modified DuPont

Key Concepts

Three Semiconductor Stories

Was this helpful?

Share this article

Stay Updated