Reading Financial Statements ② The Income Statement: From Revenue to EPS

This is the second installment of the "Reading Financial Statements & the Notes — A Primer" series. The income statement (under IFRS, formally the Statement of Comprehensive Income) is the statement most people look at first — and the one they most easily misread. At its core it is a flow: it takes the revenues and expenses of a period (a quarter or a year) and runs a top-down chain of subtractions, ultimately squeezing out net income and earnings per share (EPS).

In this piece, we will walk through the complete structure using the latest annual income statement of NVDA (Nvidia, US GAAP), comparing along the way with TSMC (IFRS), JPMorgan (a bank), Monster, and Costco — so you can see clearly how the same statement wears five different faces across five different businesses.

1. The Vertical Structure of the Income Statement: A Top-Down Subtraction

No matter how complex the line items, the income statement follows this top-down backbone. Let's apply NVDA's actual FY2026 figures (through January 2026) directly:

There's a trick to reading this statement: each step down answers the question, "How well does the company control this particular category of cost?"

• Gross profit reflects control over materials (raw materials / direct costs);
• Operating income reflects control over labor (personnel, R&D, selling and admin);
• Net income reflects control over charges (interest, tax, non-operating items).

The three layers are independent and should be diagnosed separately — only then can you tell exactly where a company's profitability decline is getting stuck.

The reason NVDA is a textbook example lies in its 71.1% gross margin — a hardware company achieving a margin approaching that of the software industry, driven by design capability, ecosystem, and pricing power rather than manufacturing. Remember this number; it will resonate when we compare against TSMC and Costco shortly.

2. Revenue Recognition: The Linchpin of the Entire Statement (IFRS 15 vs ASC 606)

Revenue sits on the very first line of the income statement, which is why it's called the top line. It looks simple — "the money received from selling things" — but "when can it be recognized, and how much" is the area with the greatest management discretion in the entire report, and the one most frequently tampered with.

The good news is that the biggest accounting convergence of the past decade happened right here: IFRS 15 and US GAAP's ASC 606 have converged into a nearly identical "five-step model":

1. Identify the contract with the customer
2. Identify the performance obligations in the contract
3. Determine the transaction price
4. Allocate the transaction price to each performance obligation
5. Recognize revenue when (or as) the performance obligation is satisfied (control transfers to the customer)

The core principle is "transfer of control": receiving cash does not entitle you to recognize revenue. You must wait until your obligation to the customer is genuinely fulfilled and control of the goods or services has transferred to the customer. This very principle is the fault line behind the Luckin Coffee revenue-inflation and Valeant channel-stuffing cases covered in Part Six — both recognized revenue before control had truly transferred.

Compare two examples from this series: Costco's membership fees are not all recognized at the moment of receipt, but deferred ratably over the membership period (typically 12 months) — the standard "satisfied over time" treatment under ASC 606 / IFRS 15. TSMC, by contrast, recognizes revenue when control of the wafer transfers to the customer. Different business, different timing of recognition.

3. Cost of Revenue and Inventory: The Most Practical Divergence Between IFRS and GAAP

Gross Profit = Revenue − Cost of Goods Sold (COGS). COGS fluctuates with sales volume and consists mainly of raw materials, direct labor, and inventory. Here's the problem: a sales cycle may span several accounting periods, so exactly which "batch" of inventory cost do you use as COGS? This brings in inventory costing methods — the most concrete divergence between the two standards:

This difference is very real: a US company using LIFO will report a noticeably lower book gross profit during inflation than an otherwise identical company using FIFO — not because it runs the business worse, but purely as an accounting choice. So when comparing gross margins across standards or across companies, always first confirm the inventory costing method (the notes disclose it). This is exactly the variable the CFA exam stresses repeatedly, and the first thing a practicing analyst corrects for when comparing peers.

4. Not Every Company Has a "Gross Profit": The Bank Income Statement (JPM)

Now lay open JPMorgan's income statement — and you'll find that from top to bottom there is no "cost of revenue," and no "gross profit" line. Because a bank doesn't sell things; its business is the "buying and selling of money": it absorbs deposits at a lower cost and lends or invests at a higher rate, earning the spread in between. Its income statement looks like this:

Two things manufacturing-oriented readers find most unfamiliar: First, there is no gross profit line — the bank's "gross margin" concept is replaced by Net Interest Margin (NIM), i.e., net interest income ÷ interest-earning assets, around 2.6% for JPM. Second, there is a "Provision for Credit Losses" — an expense a bank reserves in advance for "loans that may not be collected in the future," estimated under IFRS 9's Expected Credit Loss (ECL) model or US GAAP's CECL model. Note: this line item depends heavily on management's assumptions about future macro scenarios (unemployment, GDP), making it the single largest adjustment valve for bank earnings.

5. Operating Income, EBITDA, and EPS: Three Gates on the Way Down

Operating Income (EBIT): Reveals "Management Quality" Better Than Gross Profit

From gross profit, subtract operating expenses (R&D, selling, admin) to get operating income. Compared with gross profit, which easily bounces with raw-material prices, watching operating margin over the long run better reflects a company's management quality. NVDA's operating margin runs as high as 60.4%, TSMC's 50.8% — both top-tier within their respective industries; whereas Costco, a thin-margin retailer, has an operating margin in the single digits.

EBITDA: A Cash-Profit Proxy for Capital-Intensive Industries

EBITDA (Earnings Before Interest, Taxes, Depreciation, and Amortization) = Operating Income + Depreciation + Amortization. Its purpose is to add back depreciation and amortization — two non-cash deductions — to approximate a company's operating cash-generating ability, making it especially suited to industries with large upfront capital investment and long amortization periods.

TSMC is the perfect example: with annual capital expenditure as high as $41.0B, its yearly depreciation expense is enormous — this depreciation depresses book net income, yet is not cash that actually flowed out during the period. So for heavy-asset companies like TSMC, EBITDA and operating cash flow capture debt-service and reinvestment capacity better than net income alone. But remember: EBITDA is not a line item defined by IFRS or GAAP; it is a number "the company calculates itself", and Buffett and Munger have publicly scorned it for "pretending that depreciation is not a real expense." Use it, but don't be fooled by it.

Earnings Per Share (EPS): Basic vs Diluted

Net income, less preferred dividends, divided by the weighted-average shares outstanding, gives EPS. Reports present two versions:

• Basic EPS: uses actual shares outstanding.
• Diluted EPS: assumes everything that "could become common stock" (employee stock options, convertible bonds, RSUs) is fully converted, increasing the share count and showing how far EPS is diluted. This is the more conservative — and the more appropriate — version to use.

"Weighted-average" means: new shares issued mid-year are counted in proportion to the time they were outstanding. For example, 1 million shares issued only on July 1, with only half a year to year-end, count as just 500,000 shares for that year. The larger the gap between basic and diluted EPS, the more complex the company's capital structure (more options/convertibles), and the more you must watch when those dilutive conditions trigger.

6. Putting Five Companies' Income Statements Side by Side

One statement, five faces. Monster's example is especially worth savoring: its FY2025 revenue grew 10.7%, but break it apart — volume grew 13.3%, while average price per case actually fell 1.6%. This means growth came from "selling more" rather than "selling for more" — healthy volume growth, but also a signal that it currently lacks pricing power. This "volume vs price" decomposition is invisible on the face statement; it requires the notes and the earnings call — we go deeper in Part Five (industry KPIs) and Part Six (earnings calls).

In the next piece, we leave the "flow" of the income statement and step into the "stock" of the balance sheet — examining how nearly half of TSMC's total assets are locked up in plant and equipment (a heavy-asset profile), contrasted with Monster's near-zero long-term debt lightness, and the identity that strings it all together: A = L + E.

FAQ

What is the biggest difference between IFRS and US GAAP on the income statement?

The four most practical: (1) inventory costing — IFRS prohibits LIFO, US GAAP allows it, producing a large gross-margin gap under inflation; (2) recovery of inventory write-downs — IFRS permits reversal, US GAAP prohibits it; (3) development spending — IFRS requires capitalization if criteria are met, US GAAP always expenses; (4) Stock-Based Compensation (SBC) — both standards (IFRS 2 vs US GAAP ASC 718) require employee stock awards to be booked as an "expense," but the timing of recognition and the tax treatment differ: for graded-vesting awards, IFRS 2 mandates treating each tranche as a separate award, front-loading the expense (accelerated recognition), while US GAAP permits a straight-line method, so for the same award the early-period expense is typically higher under IFRS; forfeitures for departing employees must be estimated in advance under IFRS, whereas US GAAP allows a "recognize when they occur" election; the related deferred-tax measurement basis also differs. As for revenue recognition, IFRS 15 and ASC 606 have largely converged.

But for investors, more important than the standard differences is this: under either standard, SBC is a real shareholder-dilution cost, yet it is often added back by companies via Non-GAAP EPS to flatter earnings — one of the biggest traps in tech-stock valuation (NVDA's FY2026 fair value of RSUs/PSUs vested during the year reached $22.2B; see this article's "Read the Notes").

Why does a bank's income statement have no gross profit?

Because a bank sells no goods, there is no cost of revenue to deduct. Its core profit is "net interest income" (the spread of interest income on loans minus interest expense on deposits), plus noninterest revenue such as fees, less the provision for credit losses and operating expenses, yielding net income. Its "gross margin" concept is replaced by Net Interest Margin (NIM).

Can EBITDA and Non-GAAP EPS be compared directly?

Be very careful. Neither is a standard figure defined by IFRS/GAAP; both are "calculated by the company itself." Different companies may exclude different items, so before any cross-company comparison you must turn to the reconciliation in the notes and see clearly what each one added back — paying special attention to whether real dilution costs like Stock-Based Compensation (SBC) were added back to flatter earnings.

ProfitVision LAB · Shiba the Disciplined | I teach you how to think, not just what to do