Stock-Based Compensation: The Cost That Inflates Tech Margins

Nvidia (NVDA) just announced it will start including stock-based compensation in its non-GAAP earnings — a roughly ~$6.4B line item it spent years explicitly excluding. That decision is the biggest accounting honesty moment in tech since the 2003 expensing reforms, and it forces every investor to relearn how to read software margins.

What is stock-based compensation in plain English?

Stock-based compensation is when a company pays employees with shares (or options or restricted stock units) instead of cash. The employee gets equity that vests over time. The company books an expense equal to the fair value of the grant.

The accounting is the part that confuses investors. SBC is a real expense under GAAP — it hits the income statement. But because no cash leaves the company, SBC is added back on the cash flow statement, which makes operating cash flow look higher.

Most companies then exclude SBC again from their preferred "non-GAAP" or "adjusted" earnings number, which inflates the headline metric most investors actually pay attention to.

Why is SBC suddenly a 2026 controversy?

Because the dollars got too big to ignore. SBC at large software firms now equates to ~20%+ of revenue, with extreme cases running ~30-50%. The S&P 500 average is closer to ~2%. That gap matters when a sector trades at premium multiples partly because the headline margin looks high.

Nvidia's announcement that it will include SBC in non-GAAP earnings starting fiscal Q1 2027 is a watershed. Once one mega-cap accepts that SBC is a real expense, the rest of the industry is under pressure to follow — and that mathematically compresses every "adjusted" margin in software.

How does SBC actually distort earnings?

It distorts in three places at once.

First, operating margin: removing SBC from operating expense inflates operating margin, often by 5-15 percentage points. A "60% non-GAAP operating margin" company might be running ~45% on a true GAAP basis.

Second, free cash flow: because SBC is a non-cash expense, free cash flow conventionally adds it back. That makes FCF look better than the equivalent GAAP operating income. Cash flow yield calculations should subtract SBC to get a clean comparable.

Third, share count growth: SBC issuance dilutes existing shareholders. A company can post growing earnings per share even when the business is shrinking, simply by aggressively buying back stock to offset SBC issuance.

For more on related quality screens, see our explainer on free cash flow yield.

How do I calculate SBC as a percentage of revenue?

Pull the cash flow statement. SBC is usually a line item near the top under "adjustments to reconcile net income to net cash provided by operating activities."

Then divide it by total revenue from the income statement. That gives you the SBC-to-revenue ratio.

Then compare to peers. A 5% ratio is healthy. ~10-15% is meaningful. ~20%+ is a yellow flag for shareholder dilution. Anything ~30%+ should trigger a deeper investigation.

The table tells the contrarian software story: cash-rich, mature platforms run reasonable ratios. Pre-FCF-positive growth software runs eye-watering ones.

Why do tech companies use so much SBC?

Three reasons that compound.

The talent market for software engineers, AI researchers, and ML scientists is fierce. Equity is the only way to compete with venture-backed startups dangling pre-IPO upside.

SBC also has a tax advantage: under US accounting, it is a deductible expense for the company while being relatively tax-efficient for many employees holding RSUs.

Finally, SBC has no immediate cash cost. That made it irresistible during years of growth-at-all-costs investing. The hidden cost was always shareholder dilution — and during periods of falling stock prices, that dilution accelerates because the company has to issue more shares to deliver the same dollar value of compensation.

Are buybacks a real offset to SBC dilution?

Not always. Many software companies run share repurchase programs that exactly offset SBC issuance — meaning shareholders are not getting capital returned, they are simply paying for employee compensation through the buyback budget.

To test, calculate "net buybacks" = cash spent on buybacks minus SBC expense. If the result is negative or near zero, the buyback is essentially employee comp dressed up as capital return. Compare to companies like Microsoft (MSFT), where buybacks substantially exceed SBC and represent genuine return of capital.

For more context on shareholder yield versus shareholder dilution, our investment strategies hub covers capital return frameworks.

Common mistakes investors make on SBC

Three patterns trip up otherwise smart investors.

Mistake 1: Trusting "non-GAAP" margin without adjusting for SBC. Every adjusted number that excludes SBC is overstating profitability.

Mistake 2: Comparing tech companies on price-to-earnings without converting to GAAP EPS. A company at "20x non-GAAP" and "45x GAAP" is the same business at very different valuations.

Mistake 3: Adding SBC back to free cash flow without subtracting buybacks needed to offset dilution. If you want a clean owner-earnings number, treat the buyback offset as an SBC-equivalent cash cost.

Pro tips for screening on SBC

Build a simple screen: SBC-to-revenue ratio, three-year share count change, and net buyback yield (buybacks minus SBC expense). Companies that score well on all three are running disciplined dilution policies.

Watch for sudden SBC acceleration. A jump from ~10% to ~18% of revenue in one year often signals retention pressure during stock-price declines. The company is paying more shares to keep employees who saw their RSUs lose value.

Look for accounting honesty signals. Companies that voluntarily include SBC in non-GAAP earnings — like Berkshire Hathaway and now NVDA — are usually run with stronger shareholder discipline overall.

When NOT to obsess over SBC

For mature, low-growth businesses with SBC under 3% of revenue — like consumer staples or industrials — SBC is a footnote, not a thesis. Coca-Cola (KO) and Procter & Gamble (PG) run SBC ratios most software companies would call modest within a single quarter.

For early-stage growth companies pre-FCF, high SBC is partly necessary; the question is whether the company can grow into the dilution. If revenue growth meaningfully exceeds share count growth, dilution is being earned.

The danger zone is the middle: mature growth companies where SBC is high, growth is decelerating, and buybacks are large enough to mask dilution. That trio is where adjusted earnings most diverge from owner economics.

Ready to analyze these stocks yourself? Search any ticker on MainRatios to see valuations from 6 legendary investors - free.