RBLX · Technology(services-prepackaged software) · 8 years of annual accounts filed with the SEC · latest fiscal year ended 2025-12-31
Roblox Corp reported revenue of $4.9 billion in fiscal 2025. Of the $4.7 billion its operations generated over 8 years, 34.9% went back into the business. On the accounting screens, it passes 6 of 9 Piotroski tests, its Altman Z'' of -2.70 is in the distress zone and its Beneish M-score is below the -1.78 line; 1 of the six cross-checks between its statements fires.
Revenue, fiscal 20254.9B
Operating margin-25.2%gross margin 78.1%
Return on invested capital-88.2%-79.0% on average over 5 years
Free cash flow after stock pay226.4M4.6% of revenue
Net debt ÷ EBITDANet cash197.5M more cash than debt
Piotroski F-score6/9tests of improvement passed
Share counts are in today's shares. The SEC's filings restate only recent years after a split, so these jumps were read as splits and the older years scaled to match — otherwise per-share figures would compare different units:
3-for-1 before fiscal 2021.
Is it growing?
Revenue and the operating income it turns into. Growth that does not reach operating income is growth bought with margin.
RevenueOperating income
-2.0B02.0B4.0B6.0B
2018
2019Revenue 508.4MOperating income -76.4M
2020Revenue 923.9MOperating income -266.1M
2021Revenue 1.9BOperating income -495.1M
2022Revenue 2.2BOperating income -923.8M
2023Revenue 2.8BOperating income -1.3B
2024Revenue 3.6BOperating income -1.1B
2025Revenue 4.9BOperating income -1.2B
20182019202020212022202320242025
Compound growth a year
3 yrs
5 yrs
7 yrs
Revenue
+30.0%
+39.6%
—
Free cash flow per share
—
+20.6%
—
Shares
+5.0%
+4.8%
—
Falling shares are buybacks: each remaining share owns more of the company.
Does it earn more than its capital costs?
Margins say how much of each sale is kept; return on invested capital says how much the business earns on the money it needs to operate. Growth only creates value when that return is above the cost of the capital.
Needs a cost of capital, which comes from the valuation below.
Return on equity
-270.0%
Return on assets
-11.1%
Asset turnover
0.51×
Research & development
32.1% of revenue
Overheads (SG&A)
11.9% of revenue
Is the profit cash, and where does the cash go?
Net income is an accounting opinion; free cash flow is what was left in the bank after investing. Stock-based pay does not leave the bank — shareholders pay it through dilution — so it is shown taken off as well.
Net incomeFree cash flowAfter stock-based pay
-2.0B-1.0B01.0B2.0B
2018
2019Net income -71.0MFree cash flow 15.9MAfter stock-based pay -1.7M
2020Net income -253.3MFree cash flow 420.2MAfter stock-based pay 341.0M
2021Net income -491.7MFree cash flow 565.8MAfter stock-based pay 223.9M
2022Net income -924.4MFree cash flow -56.9MAfter stock-based pay -646.4M
2023Net income -1.2BFree cash flow 137.5MAfter stock-based pay -730.5M
2024Net income -935.4MFree cash flow 642.7MAfter stock-based pay -373.1M
2025Net income -1.1BFree cash flow 1.4BAfter stock-based pay 226.4M
20182019202020212022202320242025
Where 8 years of operating cash went, 2018–2025
4.7B generated by the business. Each band is its share of that total.
Reinvested in the business 35%1.6B
Acquisitions 2%106.7M
Dividends 0%0
Share buybacks 0%0
Kept, or used to pay down debt 63%3.0B
Over the same years it paid 4.0B in stock.
Per share
Earnings per shareFree cash flow per shareDividend per share
$-2.00$-1.00$0.00$1.00$2.00
2018
2019Earnings per share $-0.15Free cash flow per share $0.03
2020Earnings per share $-0.46Free cash flow per share $0.77
2021Earnings per share $-0.97Free cash flow per share $1.12
2022Earnings per share $-1.55Free cash flow per share $-0.10
2023Earnings per share $-1.87Free cash flow per share $0.22
2024Earnings per share $-1.44Free cash flow per share $0.99
2025Earnings per share $-1.54Free cash flow per share $1.97
20182019202020212022202320242025
Shares outstanding
Diluted shares
400.0M500.0M600.0M700.0M
2018
2019Diluted shares 489.2M
2020Diluted shares 546.3M
2021Diluted shares 505.9M
2022Diluted shares 595.6M
2023Diluted shares 616.4M
2024Diluted shares 647.5M
2025Diluted shares 689.6M
20182019202020212022202320242025
How strong is the balance sheet?
Debt is not bad in itself; debt the business cannot service in a bad year is. Net debt is debt minus cash, and below zero the company holds more cash than it owes.
Net debt
-3.0B-2.0B-1.0B01.0B
2018
2019
2020Net debt -893.9M
2021Net debt -2.0B
2022Net debt -2.0B
2023Net debt 326.5M
2024Net debt 294.7M
2025Net debt -197.5M
20182019202020212022202320242025
Net debt ÷ EBITDA
0.2×
Interest coverage
-30× operating income ÷ interest
Current ratio
0.96 current assets ÷ current liabilities
Cash conversion cycle
— collects in 67d
Three classic screens of the accounts
Each asks a different question of the same statements — is it improving, does it look like a company heading for distress, do the accounts resemble ones that were manipulated. None is a verdict; each shows what moves it.
Piotroski F-score
Is the business improving? Nine yes-or-no tests, this year against last.
6of 9 tests passed
✕ProfitableReturn on assets above zerofailed
✓Cash from operationsOperating cash flow above zeropassed
✓Profitability improvedReturn on assets higher than a year beforepassed
✓Profit backed by cashOperating cash flow above net income (low accruals)passed
✓Less long-term debtLong-term debt as a share of assets fellpassed
✕More liquidCurrent ratio higher than a year beforefailed
✕No new sharesShare count did not growfailed
✓Better gross marginGross margin higher than a year beforepassed
✓Sells more per assetAsset turnover higher than a year beforepassed
Where it misleads: it measures change, not level. An excellent business that stood still for a year scores low; a poor one recovering from a disaster scores high.
Altman Z''-score
Does the balance sheet look like those of companies that went bankrupt?
-2.70distress zone
1.12.6
Working capital ÷ assets -0.02 × 6.56-0.16
Retained earnings ÷ assets -0.53 × 3.26-1.73
Operating income ÷ assets -0.13 × 6.72-0.87
Equity ÷ liabilities 0.04 × 1.05+0.05
Where it misleads: buybacks. A company that returns so much cash that its equity turns small or negative sinks the last two ratios without being anywhere near bankruptcy. Banks and insurers do not fit the model at all.
Beneish M-score
Do the accounts resemble those of companies that manipulated their earnings?
-3.46below the -1.78 line
-1.78
Receivables vs sales 1.08+0.99
Gross margin slipping 1.00+0.53
Soft assets 1.02+0.41
Sales growth 1.36+1.21
Slower depreciation 1.26+0.14
Overheads vs sales 1.05-0.18
Profit not in cash -0.30-1.40
Leverage rising 0.98-0.32
Where it misleads: fast growth and acquisitions. Sales growth carries a heavy weight, so a company growing 50% a year scores like a suspect without having done anything. It is a screen from a 1999 study, not an accusation.
Where the statements disagree
Cross-checks between the income statement, the balance sheet and the cash flow for the latest year, each with the benign reading and the worrying one.
The effective tax rate is -0.3%.
Benign
A favourable geographic mix, or legitimate tax credits.
Worrying
Not sustainable; projecting it forward inflates the valuation.
What is it worth, under which assumptions?
A company is worth the cash it will generate, brought back to today. This model projects revenue with growth that fades over the years, applies a free cash flow margin, and discounts the result at the cost of capital. Every assumption says where it came from, and all of them can be changed.
With a free cash flow margin at or below zero in year ten, the business never generates cash for its owners and a DCF says nothing useful. Set a positive margin for year ten to see what it would take.
What it has filed lately
The last twelve months at the SEC, most important first: annual and quarterly reports, events, large holders and insiders.
Accounts from the company's own SEC filings; lines some companies do not report are shown as a dash rather than estimated. The risk-free rate is the 10-year US Treasury yield. A DCF is a way of making assumptions explicit, not a forecast: it states what the company would be worth if the assumptions held. The scores are screens that point where to look, not verdicts. Nothing here is a recommendation to buy or sell.