Ingram Micro Q2 2026 — The Same Cash Burn as TD SYNNEX, the Opposite Balance Sheet

A beat the tape sold

Start with what isn't in dispute. Net sales of $14,531M cleared the top of the company's own $13.6B to $14.0B range by $531M. Gross profit rose 14.2%, SG&A fell to 4.9% of net sales from 5.4%, adjusted income from operations rose to $280.4M from $200.8M, and non-GAAP net income rose 34.5%. The chief executive, Paul Bay, called it "the strongest second quarter results in Ingram Micro's history," and the income statement backs him.

The stock went from $30.40 to $28.21 the next session anyway, down 7.2% (unverified — regular-session closes, market data rather than a filing figure). The market didn't sell the income statement. It sold the cash flow statement, where cash used in operations was $533.2M, 1.79x last year's $298.0M, Nearly two months later TD SYNNEX printed the same shape at close to twice the size, negative $976M of free cash flow, and fell 9.88%. My argument is that the shape is the same and the cause isn't.

Where the cash went

The release is specific about the cause: "heavier investment in inventory in the current fiscal second quarter to support the growth of the business ahead of ongoing supply constraints and related increases in average selling prices." Read that as two claims packed into one sentence. "Support the growth" is volume. "Ahead of ... increases in average selling prices" is price. A distributor only buys ahead of a price increase when it expects to sell the same box for more later, which makes this quarter's build a position, not a backlog.

The statement of cash flows confirms the scale. Inventory absorbed $852.9M in the quarter. Payables funded $291.1M and receivables released $162.0M, but book overdrafts took back $122.7M, so $110.9M of net income became $533.2M of cash used. For the half, cash used in operations was $1,511.1M against $498.4M, and adjusted free cash flow was negative $1,489.7M against negative $421.9M.

The chief financial officer, Mike Zilis, named the position on the call: "we have done some strategic procurement of certain product categories to get out ahead of continued ASP increases and potential memory-related supply shortages" (unverified pending the official transcript). Decode it. "Strategic procurement" means Ingram chose the timing, not its customers. "Memory-related" names the input. That makes it the warehouse is the position in plain form: the inventory is a bet on the spread between what memory-heavy products cost in June and what they clear at in the second half, and the cash statement records the bet before the income statement records the payoff. He also said "the ASPs themselves are driving quite a bit of inflation in the dollar of the balance sheet" (unverified, same transcript). So part of the 16.2% sequential rise in inventory is price on the same units, not more units.

The composite, and the component

Management's summary number was flattering: "our net working capital of 26 days in Q2 2026 was three days better than the same period of 2025" (unverified pending the official transcript). A composite like that is where I usually go looking for trouble, because the composite hides the component: HPE's cycle improved a day while inventory days rose about thirty. So I rebuilt the parts from the filed balances. They hold up, which is not what I expected.

Inventory was 41.4% of quarterly net sales at the end of June, against 43.1% a year earlier. On my arithmetic, days of inventory were 40.4 against 41.9, days of receivables 66.9 against 65.1, and days of payables 79.1 against 76.5, so the three parts net to about 28 days against about 30.5. The company's own definition lands at 26 against 29; the direction and the size match. Year over year, inventory rose 9.3%, receivables 16.7% and payables 17.4%, on net sales up 13.6%. Suppliers grew their funding faster than Ingram grew its stock. That is the opposite of TD SYNNEX, where the Q3 FY26 page found payables growing more slowly than inventory. The component that did move is sequential: inventory went from 37.1% of net sales in March to 41.4% in June. That is a half-year swing, and the next section is about whether it is also a seasonal one.

Eight quarters, one range

A single quarter of working capital can't tell a lump from a level, so I pulled the series back to September 2024. Working capital dollars ran $4.34B, $4.14B, $4.31B, $4.61B, $4.95B, $3.55B, $4.42B and $4.90B. The June figure is high, but it is inside the range, just under the $4.95B of September 2025. Both year-ends in the series released cash: $4.34B fell to $4.14B into December 2024, and $4.95B fell to $3.55B into December 2025, when Q4 adjusted free cash flow was $1,630.4M.

Inventory is the line that broke out. $6.02B is the highest quarter-end in the series, above the $5.51B of June 2025. It is also the line the release explained. The same six months last year added $810M of inventory with payables roughly flat; this year added $1.05B. The pattern is the same, and it is bigger. And on a trailing basis the cash picture is better, not worse: four-quarter adjusted free cash flow through June was $30.8M, against negative $339.3M a year earlier, because the December release was so large.

Who funded it

This is the question the angle was built on, and the half-year answer is less comfortable than the year-over-year one. From December 27 to June 27, inventory rose $1.05B while accounts payable fell $168M. Supplier-finance obligations inside payables dropped to $2,591M from $3,073M, a $483M retreat. The money came from Ingram. Cash fell $1,056M to $809M, and total debt rose $596M to $3,795M: the ABL revolver went to $550M from zero and uncommitted lines of credit and other debt to $418M from $105M. Net revolver proceeds in Q2 alone were $474.1M.

Held against TD SYNNEX, that is the one resemblance that holds. The library's reference case, the ODM pays three ways, says a distributor carrying a hyperscaler program pays in working capital, debt and equity at once. Ingram is paying in one of the three, debt, and paying it seasonally. There is no customer warrant; the company is retiring stock, not issuing claims on it, with 4,713,158 shares repurchased for $105M in the half. Interest expense rose only 4.6% to $76.3M. Facilities capacity is about $7.37B, with $3.79B outstanding. Who ultimately pays for this build is the revolver until December, and then, if the pattern holds, the customers who buy the stock at the higher price.

Asia-Pacific carries the growth and the thinnest margin

The regional mix is where this print is most like the AI trade and least like the rest of the business. Asia-Pacific net sales rose 27.1% to $4,421M, and at $942M the region supplied 54.2% of the group's $1,737M sales increase. Inside it, Advanced Solutions grew 51% "driven by GPU and AI-infrastructure product sets." The region is now 30.4% of net sales, from 27.2%.

Its gross margin is the group's thinnest at 4.47%, and Zilis put a number on what the AI deals cost the consolidated line: "Excluding the impact of these deals, our Q2 2026 gross margins were 6.90%" (unverified pending the official transcript), against the reported 6.60%. That is about 30 basis points of gross margin given up for GPU volume. The operating line tells the other half. Asia-Pacific's operating margin rose to 2.13% from 1.25%, and its $94.0M of operating income is now the largest of the four regions, ahead of North America's $73.5M. Low gross margin, low cost to serve: the mix dilutes the gross line and pays at the operating line. The filing does not split inventory by region, so I cannot tie the build to Asia-Pacific, and I won't pretend to.

The guide says volume is slowing and price is working

Q3 guidance is net sales of $13.55B to $13.95B, up 7.5% to 10.7%, against 13.6% in Q2, and non-GAAP EPS of $0.72 to $0.82, against $0.72 a year ago. Management said the guide assumes a "roughly 2%-3% net revenue benefit from supply constraint puts and takes," meaning higher prices and orders pulled ahead of them (unverified pending the official transcript). Pull-forward is revenue borrowed from a later quarter, so part of the growth rate is timing.

On cash, Zilis drew the line himself: "We usually have a decent-sized free cash outflow in Q3, where we're stocking for the Q4 ... We've just pulled forward some of that stocking level. That could drive a better than normal seasonal Q3" (unverified pending the official transcript). Decode it. He is saying Q2 did Q3's stocking early, so Q3's cash should look better than usual. That is checkable. Last year's Q3 cash used in operations was $146.0M, a figure the company itself called "stronger than typical Q3 seasonal norms." At the September 15 Capital Markets Day he went further on the income statement: "for the third quarter of fiscal 2026 we expect to achieve results at the high end of the range of the guidance." He said nothing there about cash. The next print will.

The strongest case against me

The best bear argument isn't that Ingram looks like TD SYNNEX. It is that Ingram is running a price bet and calling it seasonality. Buying ahead of "continued ASP increases and potential memory-related supply shortages" is a directional position on memory prices, carried on a balance sheet that earns a 6.60% gross margin and a 0.76% net margin. If memory pricing turns before December, $6.02B of inventory is marked at lower of cost or net realizable value, and the 2%–3% of pull-forward in the guide has already been spent. Add the supplier retreat: in a shortage, vendors have the leverage, and a $483M drop in supplier-finance payables is what losing it looks like. Add the seller: the selling stockholder, Ingram Holdco, agreed on September 8 to sell 13,125,000 shares, after a 14.5 million-share secondary in May that the company said reduced its primary shareholder's stake. On that reading the 7.2% drop was early, not wrong.

I take it seriously, and the price risk is the real one. I still come down on the other side for three reasons, each from the filings. Payables grew 17.4% year over year against inventory's 9.3%, so vendors are funding more of the business than a year ago, not less; the December-to-June supplier retreat happened last year too. The build sits inside an eight-quarter working-capital range, and both year-ends in that series released cash. And the trailing four-quarter adjusted free cash flow of $30.8M is $370.1M better than a year earlier. A price bet that fails would show up as a write-down and a Q3 that stocks again. A seasonal build shows up as a Q3 that doesn't.

What decides it

Two numbers, in order. First, Q3 cash used in operations for the quarter ended September 26, 2026, reported around the end of October (unverified — last year's Q3 printed October 30): smaller than $146.0M means the pull-forward did what Zilis said. Second, and the one that decides it, working capital dollars on the fiscal year-end balance sheet, reported in early 2027. At or below $4.14B, the build was seasonal and priced, and I'm right. Above $4.95B, it has become a floor in the TD SYNNEX sense and I'm wrong. In between is a partial result, and I'll call it that. Two secondary checks: inventory against $6.02B at September 26, and gross margin excluding AI deals against 6.90%.

The trade (analysis, not advice)

The stock closed at $26.98 on September 25, about 8.3 times trailing non-GAAP EPS of $3.25 (unverified — price is market data; EPS is the sum of the last four reported quarters). My view is that the market is pricing TD SYNNEX's problem into a name that has a smaller, seasonal version of it. The asymmetry sits in the Q3 cash line: a Q3 that uses less than $146.0M with inventory at or below $6.02B says the build is converting, and the discount has nothing left to point at. The risks are the two named above. A memory price reversal would turn the position into a write-down, and continuing sponsor sales are a supply overhang whatever the filings say. Anyone who wants the AI hardware cycle without a distributor's working capital can get it one layer up, through picks-and-shovels capture. Anyone buying Ingram should know they own a memory-price position until December.

Bottom line

Ingram Micro printed its strongest income statement and a $533.2M operating cash outflow in the same quarter, and they come from the same decision: buy ahead of rising prices and a memory shortage, on the revolver. TD SYNNEX's build is program inventory with a warrant on top. Ingram's is priced stock inside a range it has held for two years. My view is that December releases it the way the last two Decembers did. The number that decides it is year-end working capital dollars at or below $4.14B, or above $4.95B.