Avnet Q4 FY26 — The Upcycle Is Real. Suppliers Paid for June, the Revolver Paid for the Summer

The upcycle is real

Start with what is not in dispute. Sales of $8,295.4M came in $695.4M above the top of Avnet's own $7.30B to $7.60B outlook, adjusted EPS of $2.28 cleared a $1.70 to $1.80 range, and Q1 FY27 is guided to $9.00B to $9.30B, another 10% sequentially. Americas sales rose 55.3%, EMEA 43.7% and Asia 46.3%. The chief executive, Phil Gallagher, said Avnet delivered "record sales while growing profitability at more than two and a half times the rate of sales growth, demonstrating the operating leverage in our business model." That checks out: adjusted operating income rose to $317.9M from $142.9M on sales up 47.7%. The stock rose 5.1% on the day to $97.22 (unverified, market data rather than a filing figure).

Read the sentence again, because it is careful about what it claims. It says profitability. It does not say cash. For a distributor earning a 10.4% gross margin those are different quantities, and this year they moved in opposite directions.

The margin is leverage, not price

Gross margin did not expand. It held between 10.38% and 10.49% every quarter of fiscal 2026 and printed 10.4% in Q4 against 10.6% a year earlier. EC gross margin slipped to 9.2% from 9.5% as Asia rose to about 51% of EC sales. What moved was overhead: SG&A rose to $547.5M from $451.2M while sales rose by half, and EC operating margin climbed 2.9%, 3.2%, 3.5%, 4.1% through the year. That is the cleanest part of the print. Volume through a fixed cost base is what a cycle recovery looks like.

The volume is not all units. The CFO put memory pricing at about one-third of both the sequential and the year-over-year growth, and said about half of the quarter's inventory increase came from pricing, "substantially all" of it memory (unverified pending the official transcript). The 10-K is more reserved and calls memory pricing a contributor "to a lesser extent." Either way, this is where the warehouse is the position starts to apply to a US distributor: part of the $6.07B of inventory is a holding-period bet on memory prices, and the income statement books it before the cash statement does.

Where the $334.4M went

Net income for fiscal 2026 was $334.4M. Operating cash flow was negative $280.9M, against positive $724.5M a year earlier. Receivables absorbed $2,614.9M and inventories $918.3M; payables supplied $2,608.4M. After $73.6M of capital expenditure, free cash flow was negative $354.5M, and Avnet still returned $252.7M in buybacks and dividends. Net debt financing of $568.9M covered the difference.

The 10-K explains the $802.4M of working capital with one phrase worth reading slowly: "higher inventory purchases, the timing of payments for inventory purchases and higher accounts receivable." Timing is the admission. A payable that is high because of when it falls due is a payable that will be paid, and the payables line is the only reason the year's cash use was $280.9M and not several times that. The same filing says, correctly, that "during periods of higher growth, the Company generally uses cash to fund working capital requirements." Nobody should be surprised that a distributor growing 47.7% used cash. The question is who lent it.

The composite hides the component

Management's summary was that working capital days fell 7 to 69 "despite significant sales volume growth" (unverified pending the official transcript). It is a true number. It is also a net of three moves, and I call the pattern the composite hides the component. Recomputed from the filed balances, receivable days rose to 75.5 from 70.3 in Q3, inventory days fell to 74.3 from 77.9, and payable days jumped to 74.1 from 66.0. Customers paid slower. Suppliers were paid slower still. The cycle shortened to 75.7 days from 82.3 because Avnet stretched its vendors further than its customers stretched Avnet.

The series makes it plainer than the ratio. For two years trade working capital held between $6.05B and $6.60B. At June 27 it was $6.90B, only modestly outside the band, and just 20.8% of annualized sales against 27.0% a year earlier. On that number alone Avnet financed a 47.7% growth quarter almost for free. The reason is one ratio: payables were 99.8% of inventory, against 66.6% a year earlier. In Q4 alone payables rose $1.43B, or 30.9%, against $1.18B of added sales. The vendor becomes the credit, running upstream: the chip suppliers, not Avnet's balance sheet, were holding Avnet's warehouse on June 27.

The receivables that left the balance sheet

There is a second lender the ratio does not show. Avnet sells receivables without recourse, mainly in Asia, and books the proceeds as operating cash. Sold but unsettled receivables were $1.98B at year end against $1.56B, an increase of $0.42B. The reported receivables of $6.88B, up 59.1%, are after that. So the receivable build is larger than the balance sheet says, and part of the fiscal-year operating cash flow came from banks buying invoices. None of that is improper; the 10-K discloses it plainly. But it means the negative $280.9M already includes help that has to be renewed every quarter.

Fifty-two days

This is the line that changed my read, and it is not in the earnings release. On June 27 Avnet owed $557.0M on its revolver and $500.0M on its securitization, $1.057B together. The prospectus for the August notes gives the same two balances as of August 18: $1.575B and $700.0M. That is $2.275B, up $1.218B in 52 days, on facilities costing 4.55% and 4.49%. In between, Avnet raised the securitization limit to $700.0M, drew it to the limit, and signed a $375M term loan on August 12. Then it sold $550M of 5.650% notes due 2031 "to repay amounts owed under the Company's senior unsecured revolving credit facility" and the securitization program.

Put the two halves together. The June balance sheet showed suppliers carrying the inventory. The summer shows the bill: the revolver drew $1.218B in the seven weeks after June 27, and the CFO had already said Q1 would use cash "primarily in the form of accounts receivable" (unverified pending the official transcript). The earnings are real. They are being financed first by vendors and banks, and only later by customers.

Farnell printed once

The other headline is Farnell's 9.0% operating margin, against 5.2% in Q3 and 4.3% a year ago. Operating income rose $21.0M on $45.4M of added sales, about 46% incremental, as gross margin jumped to 29.6% from 27.2% after three quarters between 26.8% and 27.2%. The CFO credited "a better mix of higher-margin" on-the-board components (unverified pending the official transcript). I believe the direction. I don't bank the level. A mix-driven margin move is direction-agnostic, and this one has printed once; the fiscal-year margin is 5.9%, and Q4 supplied $44.8M of the $105.7M year. Farnell is the one business that could re-rate the whole company, because it is the part that earns a margin. It needs a second quarter above 5.9% before it counts, and I could not find anything in the filings that separates the mix effect from memory pricing in that gross margin.

The strongest case against me

Here is the bull argument at full strength, and it is not weak. Distributors always use cash on the way up and release it on the way down; the 10-K says so, and fiscal 2025 proves it with $724.5M of operating cash in a soft year. Payables rising with inventory is what a healthy supplier relationship looks like in an allocation market, not a sign of strain. The CFO reported gross leverage falling to 3.2x from 3.6x with a 3x target (unverified pending the official transcript), the notes were placed at investment grade, and Arrow, the closest peer, generated $318M of operating cash in its June quarter on 32% growth. On that reading the August borrowing is working capital doing exactly what it should, termed out sensibly, and the cash comes back the day the cycle cools.

I accept most of that and still land on the other side. The leverage ratio improved because earnings grew, not because debt fell; debt rose $550.8M in the year and revolver plus securitization borrowings roughly doubled after it. The covenant headroom the bull case relies on was widened by an amendment that runs only through September 2026. And the release of cash in a downturn is a promise about a period the guide says is not coming: Q1 is another 10% up. Growth this fast does not stop consuming working capital; it only changes who funds it.

What decides it

The October 3, 2026 balance sheet, due with the Q1 FY27 report around October 28 (unverified, earnings calendar). I'm right if payables fall back below 90% of inventory while total debt stays above the $3.21B of June 27: the June payables were timing, and bank debt replaced them. I'm wrong if payables hold near inventory and Q1 operating cash flow is positive: then suppliers are structurally carrying the book and the growth is funding itself. Anything else is a partial result, and I'd call it that. Two secondary checks: receivable days against 75.5, with sold receivables against $1.98B, and Farnell's margin against 5.9%.

One more thing to watch, and it is unresolved. The $650M convertible notes become convertible early if the stock trades above about $91 on at least 20 of the last 30 trading days of a quarter, and principal is settled in cash. The stock closed at $103.08 on September 25 (unverified, market data). Holders rarely convert early, because they give up the option value. I could not determine from the filings how a triggered conversion right would change where the notes sit on the balance sheet.

The trade (analysis, not advice)

I wouldn't pay for the EPS beat until the October balance sheet shows who funded the summer. The cycle is real, the operating leverage is real, and a Q1 guide of $2.80 to $2.90 of adjusted EPS says earnings are still accelerating. The problem is that the market prices Avnet on EPS, and EPS is the one line that does not show the $1.218B. If payables normalize and debt holds, the next leg of the cycle carries more interest cost and more rollover risk than the P&L currently shows. If the growth turns out to fund itself, Avnet is the cheapest way to own the component upcycle and I've been too cautious. Anyone who wants the cycle without the working capital can get it one layer up, from the component makers, through picks-and-shovels capture. Anyone buying Avnet is buying the receivables along with the recovery.

Bottom line

Avnet printed a genuine cycle quarter: sales up 47.7%, EC margin at 4.1%, a $9.15B guide midpoint. It also used $290.7M of operating cash in the quarter while suppliers funded almost all of its inventory, and then drew $1.218B more on its revolver and securitization in 52 days. The income statement shows the recovery. The balance sheet shows who is paying for it. I'm right if October's payables fall below 90% of inventory with debt above $3.21B, and wrong if payables hold and Q1 cash turns positive.