Amazon is funded by $100bn it does not own
Databoat pulled five filed figures for ten retailers straight from SEC filings and derived the cash conversion cycle. Two of the ten are funded by their own suppliers and customers. One has $12.8bn of its own cash standing in a warehouse.
- Source
- SEC EDGAR XBRL company filings, latest 10-K per company
- Period
- Latest filed fiscal year, ends ranging Aug 2025 to May 2026
- Published
- August 14, 2026
- Built with
- Databoat
-100,399.57
million dollars is what the negative cash cycle is worth — money that sits in the business, funded by suppliers and customers rather than by a lender, at no interest
Amazon
- Amazon-100,399.57
- Apple-81,029.48
- Target823.74
- Costco1,290.36
- Best Buy1,668.66
- TJX4,544.81
- Walmart5,617.23
- Lowe's12,421.23
- Nike12,764.49
- Home Depot27,080.26
Horizontal: days of inventory held. Vertical: days taken to pay suppliers. Both derived from filed balances against cost of sales.
Hover a point
Anything below the diagonal is paying for stock before selling it. Above it, suppliers are financing the shelves. Apple sits at the extreme — nine days of inventory and 115 days of credit — which is why it collects cash long before it spends it.
Two positions, same industry: Lowe’s holds 110 days of stock, Costco holds 28.
Basis — what this measures
- Sources
- SEC EDGAR XBRL company facts API. Five figures per company — revenue, cost of sales, receivables, inventory and payables — taken directly from each company’s latest annual filing.
- Period
- Latest filed fiscal year per company. Year ends range from 31 August 2025 to 31 May 2026, so the set is not a single common window.
- Grain
- One row per company. Balance sheet figures are taken at exactly the same date as the income statement period end, verified rather than assumed.
- Excludes
- Kroger and Dollar General were dropped rather than approximated — Kroger’s latest cost-of-sales tag stops in 2018, and Dollar General does not file a comparable receivables line. Year-end balances are point-in-time, so seasonal businesses are measured at their quietest moment.
Movement — why the number is what it is
Days of inventory plus days to collect, minus days before paying. Negative means cash arrives before it leaves.
- Apple-71.07
- Amazon-51.12
- Costco1.71
- Target2.87
- Walmart2.9
- Best Buy14.61
- TJX27.48
- Lowe's52.54
- Home Depot60.02
- Nike100.41
The spread runs from −71 days to +100 days. Apple and Amazon are funded by their own trading cycle. Nike finances a hundred days of its own — over three months between paying for a shoe and being paid for it.
Left mark: days of inventory held. Right mark: days of supplier credit taken. A bar running right means suppliers cover the stock.
- Apple9.45→115.4
- Amazon39.25→124.85
- Walmart40.12→42.99
- Costco27.56→30.1
- Target59.47→61.01
- Best Buy59.07→53.59
- TJX63.9→40.07
- Home Depot85.81→38.19
- Lowe's110.01→62.07
- Nike103.37→49.61
Walmart, Costco and Target sit almost exactly balanced — inventory days and payables days within three of each other. That is not luck, it is the model: negotiate terms that match how fast the shelf turns and the working capital requirement disappears.
Cycle days ÷ 365 × revenue, in millions of dollars. Negative is cash the business holds without owning.
- Home Depot27,080.26
- Nike12,764.49
- Lowe's12,421.23
- Walmart5,617.23
- TJX4,544.81
- Best Buy1,668.66
- Costco1,290.36
- Target823.74
- Apple-81,029.48
- Amazon-100,399.57
Days become money at scale. A hundred billion dollars of Amazon’s balance sheet is funded by the gap between collecting from customers and paying suppliers.
Receivables ÷ revenue × 365, across all ten.
Retailers collect in under a fortnight because customers pay at the till. Apple and Amazon collect in about 35 days because a large part of what they sell is not sold at a till at all.
Ties — the numbers reconcile
- Exact
The cycle reconstructs from its three components
ccc = dio + dso − dpo, all 10 rows - Matched, 10 of 10
Every balance figure is dated to its own income statement period
balance sheet date = period end, verified per company - Resolved, 10 of 10
Companies file the same concept under different tags
Inventory required 2 different XBRL elements; cost of sales required 3 - Dropped, not estimated
Two companies could not be reconciled and were excluded
Kroger cost-of-sales tag ends 2018; Dollar General has no matching receivables line - Not normalised
Fiscal year ends are not aligned across the set
Ends span Aug 2025 to May 2026
Method
Five figures per company — revenue, cost of sales, receivables, inventory and payables — pulled directly from SEC filings, and four derived from them: days to collect, days of stock held, days of supplier credit taken, and the cycle those three produce.
Cash conversion cycle = days inventory + days receivable − days payable. It is the number of days a business funds itself between paying for something and being paid for it. Positive means the business is lending to its own supply chain. Negative means the supply chain is lending to it.
The thing that nearly went wrong
The first run of this analysis produced a cash cycle of 34 days for Nike. The correct figure is 100.
Nike stopped filing inventory under the InventoryNet tag in 2011 and moved to a
different element. The naive read picked up the most recent value under the
original tag — a figure from 2011 — and paired it with 2026 revenue. Nothing
about the result looked obviously wrong. It was a plausible number in a plausible
range, and it was fifteen years stale.
That is the whole argument for the reconciliation panel above. The fix was to require every balance sheet figure to carry the same date as the income statement period it is measured against, and to check it rather than assume it. Two companies failed that check and were dropped rather than approximated.
Ten companies filing the same five concepts needed three different tags for cost of sales and two for inventory. This is precisely the format problem that eats an afternoon a month in every practice, occurring in the most standardised financial reporting regime in the world.
Why this belongs in a client's pack
Profit and cash are different questions and most monthly packs answer only the first. A business can be profitable and still fail, and the cash cycle is where that shows up first — earlier than the bank balance, because it is a rate rather than a level.
The practical version for a small business: every day removed from the cycle releases one day of revenue in cash. For a company turning over £3m, taking ten days out of debtor collection releases about £82,000, permanently, without borrowing it.
What this doesn't tell you
Year-end balances are a snapshot. Seasonal retailers file at their quietest point, which flatters inventory days — a toy business measured in January looks nothing like the same business in November.
The fiscal years are not aligned. Costco's runs to August, Nike's to May, most of the retailers to late January. Comparing across them assumes conditions did not move much in between, which is a real assumption and not always a safe one.
Negative is not automatically better. A deeply negative cycle usually means significant power over suppliers, which is a commercial position rather than a financial technique, and it is not available to most businesses.
Why we published this
Every figure here traces to a filed document, every derived column states its formula, and the reconciliation above records what matched, what needed resolving, and what was dropped. Two of the five ties are marked as failures, because they are.
Early access
Your files, the same treatment.
This report was built the way Databoat builds a monthly client pack — read the files, work out what is in them, publish. Tell us which report you rebuild every month.