Capital-Light
Compounders.
The best businesses grow their profit without needing much new capital. Buffett's own example is See's Candies: its profit rose from under $5 million to $82 million while it put in only $32 million of new capital. We run that test on every stock we track: for each extra dollar of profit, how many dollars of capital did the company have to add, and does it earn a high return on the capital it uses?
Reading the filings…
How it's calculated, exactly
| Part | What it is | Where it comes from |
|---|---|---|
| Operating profit | Profit before interest and tax, as Buffett quotes See's. The average of the first three and the last three years in the window, so one boom or write-down does not decide the answer | Income statement, "operating income" |
| Capital needed | Operating working capital (current assets less cash and short-term investments, less current liabilities that are not debt) plus net property, plant and equipment | Balance sheet |
| Full cost | The same, plus goodwill and acquired intangibles: what the company paid for acquisitions. Buffett says to judge acquisitions at full cost, so it is the default view | Balance sheet |
| Capital per $1 of new profit | (capital now − capital then) ÷ (profit now − profit then). Zero or less means profit grew while capital shrank | The lines above |
| Cash returned | Dividends plus buybacks ÷ free cash flow, over the same years | Cash-flow statement, via the Shareholder Yield data |
- Built from operating assets, not from equity plus debt. A company that buys back shares has tiny or negative book equity, which would make its capital look like nothing. Surplus cash and long-term investments are left out for the same reason.
- Customers and suppliers count. Money a company holds from customers (prepayments) or owes suppliers funds the business for free, so it lowers the capital needed. That is why some payments and software companies show negative capital.
- Pre-tax on both sides, as in Buffett's letter, so a tax-rate change does not read as a change in capital efficiency.
- Left out: banks, insurers and lenders (their balance sheet is the raw material, not capital needed to run the business) and real-estate investment trusts (the properties are the business).
What this can't see
- Research, brands and advertising. They are paid for out of profit and never appear as capital, so software, drug and consumer-brand companies look lighter than they really are. Buffett's own See's arithmetic has the same blind spot.
- Cycles. Three-year averages soften a boom or a bust at either end of the window, but a company measured from a trough looks better than one measured from a peak.
- Growth by acquisition. The tangible view ignores what was paid for other companies; the full-cost view includes it. Switch between them and see who changes place.
- A business that does not need capital is not always a good stock. This list says nothing about price. Pair it with the Margin of Safety Explorer.
- Survivors only. It covers the large US stocks we track today. Companies that grew and then failed, or shrank out of the index, are not in it.
This is Buffett's See's Candies test: how much capital a company had to add for each $1 of profit it added. On Oct 23, 2026, 58 tracked companies had at least doubled profit with a 25%+ return on capital over ten years.
Who Grew The Most
On The Least Capital?
By default: companies whose operating profit at least doubled and that earn 25% or more before tax on the capital they need, ranked by how little new capital each extra dollar of profit took. Click any row for the full stock card.
Loading…
How Buffett
Thinks About It.
Not how fast a business grows, but how much new money it has to swallow to grow. The same extra profit can cost 42 cents of capital or five dollars.
All three from Berkshire Hathaway's 2007 shareholder letter. See's pre-tax profit went from under $5M to $82M, its capital from $8M to $40M, and of $1.35B of pre-tax profit earned since 1972 all but the $32M of new capital was sent to the owner.
The dream business
See's sells a product people pay for in cash, so it has almost no receivables, and it makes and ships quickly, so it holds little inventory. Almost every dollar of profit is free to leave.
Buffett calls it the prototype of a dream business, and notes there aren't many of them (2007 letter).
Why the cash must leave
A business that earns very high returns on little capital cannot keep reinvesting most of its profit at those returns for long: there is nowhere to put it. The owner should take the cash and invest it elsewhere (2007 letter).
In 1992 he wrote that owners of such a business usually benefit when it pays out most of its earnings as dividends or buybacks (1992 letter). That is why the list shows how much of each company's free cash flow came back.
The inflation tax
In his 1983 appendix Buffett compares See's with a business needing $18M of tangible assets to earn the same $2M. When prices double, both must double their nominal earnings just to stand still, which takes another $8M of capital for See's but another $18M for the other business.
That spending buys survival, not prosperity (1983 letter, appendix).
The one-dollar test
Profit kept inside the company is justified only if each retained dollar creates at least a dollar of value for owners (1984 letter).
Capital per $1 of new profit is the same test seen from the business side: how many dollars went in for each dollar of profit that came out.
Try it with any company
Fill in operating profit and capital then and now, in any one currency unit. It starts with See's, in millions of dollars.
How To Read
The List.
One ratio is only a starting point. These are the questions to ask of it.
Capital per $1 of new profit
Under $1 is See's-like: growth that barely needed new money. Around $1–3 is good business economics. Above $3 is a capital-hungry grower, which can still be a fine investment, just not a dream.
The bands come from the three examples in Buffett's letter. They are a reading aid, not a rule he gave.
Return on capital today
Operating profit divided by capital needed (the last three years averaged), like See's 60% pre-tax when Buffett bought it. A very high figure means little capital is tied up; n/a means the company runs on negative capital, funded by customers and suppliers.
The default list needs 25% or more: what Buffett's FlightSafety earned in 2007 ($270M on $1,079M), the least he calls good. Without a floor, a huge capital base that merely held still would look capital-light.
Where did the cash go?
Cash returned is the share of free cash flow paid out as dividends and buybacks over the same years. A capital-light company that keeps hoarding the cash is not using the advantage; one that pays it out is doing what Buffett says such a business should.
See the share count and the buyback detail on Shareholder Yield.
Which companies need the least capital to grow profit?
This is Buffett's See's Candies test: how much capital a company had to add for each $1 of profit it added. On Oct 23, 2026, 58 tracked companies had at least doubled profit with a 25%+ return on capital over ten years. The leader is Fortinet (FTNT), which needed no extra capital per $1 of new profit while its profit grew 13.3 times (fiscal 2016 to 2025).
| Company | Capital per $1 of new profit | Profit growth | Profit added | Return on capital now |
|---|---|---|---|---|
| FTNTFortinet Inc | none needed | 13.3× | $1.6B | n/a |
| ITGartner Inc | none needed | 6.1× | $953M | 59% |
| LYVLive Nation Entertainment Inc | none needed | 5.7× | $867M | 170% |
| WSMWilliams-Sonoma Inc | none needed | 3.0× | $909M | 136% |
| RLRalph Lauren Corp | none needed | 3.0× | $633M | 44% |
| EXPEExpedia Group Inc | none needed | 2.3× | $807M | 52% |
| HLTHilton Worldwide Holdings Inc | none needed | 2.1× | $1.3B | 62% |
| ADPAutomatic Data Processing Inc | none needed | 2.1× | $3.0B | 319% |
| JBLJabil Inc | $0.01 | 3.2× | $1.1B | 47% |
| DECKDeckers Outdoor Corp | $0.01 | 6.1× | $941M | 185% |
See's Candies needed about $0.42 of capital per $1 of added profit in Buffett's letter. The list above has the other filters, the 15-year view and the full-cost capital switch; capital-light is defined in the glossary.
Operating profit and capital (working capital, property and equipment, goodwill) from each company's 10-K filings on SEC EDGAR. Educational content only, not investment advice.
Frequently asked questions
What is the See's Candies test?
How much capital a company added for each $1 of profit it added. In Buffett's letter See's needed about $0.42, which is why he called it a great business. See the full explanation in the blog post on the test.
Which companies are excluded?
Lenders, REITs and other companies whose capital is not measured by working capital and property, because the ratio does not mean the same thing for them.
What is a good ratio?
Lower is better: under about $0.42 is See's-like or better, and a company whose profit grew while capital shrank needed none at all. A low ratio over ten years can still come from a one-off boom, so the site averages three years at each end.
About capital-light compounders
A capital-light business grows profit without needing much new capital, the kind Warren Buffett describes with See's Candies. This list ranks stocks by how much capital each extra dollar of operating profit needed over 10 or 15 years, alongside return on capital (like ROIC, but before tax) and the share of free cash flow paid back to owners, from each company's 10-K filings. Pair it with shareholder yield and earnings quality. Read how capital-light growth is measured.
Educational content only, not investment advice. A company that needs little capital to grow may already be priced for it, and a capital-hungry one can still be a good investment at the right price. Pair this list with Earnings Quality (is the profit real?), Compounders at a Fair Price (which of these are not already priced for it?), the Sector Valuation Map (how hot is its sector?) and the Margin of Safety Explorer, and read the company's own filings before acting.