Within one year, Qualcomm's chips came to sit 19 days longer on the shelf

A chip company with fuller shelves and fewer buyers

There is a good chance the phone in your pocket, or the one you replaced, runs on a Snapdragon. Qualcomm designs the chips at the heart of many Android flagships, and it collects licence fees on the wireless technology inside most phones in the world. It does not run its own chip factories. It designs, orders the silicon from contract manufacturers, and sells the finished chips to phone makers.

In the fiscal year that ended in late September 2026, Qualcomm's revenue was USD 43.1bn, down from USD 44.3bn a year earlier. A fall of 2.6% is not a crisis for a business that size. What happened on the shelves is more interesting. Inventories, the chips and materials the company had paid for but not yet sold, rose from USD 6.5bn to USD 8.6bn.

Measured against sales, that stock now covers 73 days of revenue. A year earlier it covered 54 days. Within a single year, 19 days were added to the time it would take to sell what Qualcomm already holds, at the pace it actually sold.

So the question is simple to ask: why does a chip designer facing a shrinking phone market end the year with more stock, not less? And is that stock a bet on a rebound, or a bill that has not arrived yet?

The profit line hides the strain, the stock line shows it

At first glance, fiscal 2026 looks like a triumph. Net profit rose 145.9% to USD 13.6bn, and the net margin, the share of each dollar of revenue left as profit at the very bottom, widened from 12.5% to 31.6%. A company does not usually earn that much more on a falling revenue line.

It did not, in the business itself. According to GuruFocus, fiscal 2026 net income was lifted by a tax benefit of USD 5.7bn. That is an accounting gain on the tax line, not money earned from selling chips. The operating profit, which measures what the business makes before tax and financing, tells the opposite story. It fell 29.9%, from USD 12.4bn to USD 8.7bn. The operating margin narrowed from 27.9% to 20.1%, a drop of 7.8 pp.

EBITDA, operating profit with depreciation and amortisation added back, moved the same way: down 26.3% to USD 10.3bn. Depreciation itself barely moved, at USD 1.6bn against USD 1.6bn. So the squeeze did not come from heavier write-offs on equipment or acquired technology. It came from the core of the business: lower sales and higher costs.

One of those costs is visible. Research and development spending rose to USD 10.2bn from USD 9.0bn, according to Qualcomm's annual report filed with the US Securities and Exchange Commission. That is the company choosing to spend more on new chips in a year when its old market was shrinking. Keep that in mind, because it frames the stock.

The longer view matters too. In fiscal 2024 revenue was USD 39.0bn. Fiscal 2025 was the strong year, and fiscal 2026 gave some of it back. The stock did not give anything back. It kept growing.

I would not read the 31.6% net margin as a new normal. The figure that describes how Qualcomm's chip business is doing is the 20.1% operating margin, and that one went down.

The phone market Qualcomm sells into got smaller and more crowded

The outside picture explains much of the revenue side. GuruFocus reported that revenue from handset chips in Qualcomm's chip division fell 26.0% to USD 5.1bn in the last quarter of fiscal 2026, the July to September stretch. That is the heart of the business, and it was shrinking fastest at the very end of the year, exactly when the closing stock was counted.

The market itself was weak. Counterpoint Research expects global smartphone chip shipments to fall 20.0% in 2026, as reported by Mezha. A fall that size is not a slow year. It is the kind of drop that leaves every supplier in the chain with more silicon than buyers.

Qualcomm also lost ground inside that smaller market. In the same Mezha report, MediaTek, the Taiwanese rival that dominates mid-range and cheaper Android phones, led with 28.0% of the smartphone chip market. Qualcomm held 22.0%, down from 25.0% a year earlier. The report also says Apple and Samsung gained share, which fits a broader pattern: the big phone makers want more of their own silicon, and that leaves less room for an outside supplier.

Reuters, as carried by SRN News, reported that Qualcomm's smartphone chip sales missed analysts' estimates and that the shares fell on the news. The market noticed the same thing the inventory line shows. Phone chips sold slower than expected.

Put those together and the stock build has a plain reading. A fabless chip company commits to wafers months ahead, because contract manufacturers need the orders long before the chips reach a phone. If demand falls faster than the order book can be cut, the chips arrive anyway and wait on the shelf. That is one way to get 19 days more stock in a year.

The company's own story points somewhere else entirely

Qualcomm's own account of itself has moved away from phones. In December 2025 it completed the acquisition of Alphawave for USD 2.3bn, according to its quarterly report, as reported by Futu News. Alphawave makes the high-speed connectivity technology that links chips inside data centres. The group structure now lists Alphawave Semi, LLC and Alphawave IP Inc. among Qualcomm's subsidiaries, alongside a newly listed Qualcomm Ventures LLC and Arriver HoldCo, Inc.

The company has put targets on that shift. In a current report to the US regulator, Qualcomm projected growth of more than 60.0% in revenue outside handsets and reiterated a target of USD 5.0bn in data centre revenue for fiscal 2027. Reuters reported, via SRN News, that Qualcomm valued warrants issued to Amazon at about USD 1.8bn when a deal with the cloud company was signed. Warrants are rights to buy shares later at a set price. A chip supplier hands them to a customer when it wants that customer badly enough to share some of its own upside.

These are the company's plans and the reported terms of its deals, not results. A data centre target for fiscal 2027 tells you where management is pointing. It does not show up in fiscal 2026 revenue, which fell.

This is where the stock becomes a real question. If part of the USD 8.6bn is new data centre and other non-phone products built up ahead of launches, the build is a bet that fits the strategy. If most of it is phone chips that the market did not take in the September quarter, the build is a leftover from a slump. The accounts do not break inventory down by product or by age.

Why idle stock costs more than it seems

The easiest way to see inventory is as cash in a different shape. Every chip in stock was paid for. Until a phone maker or a data centre buys it, that money does no work.

The cash line shows the pressure. Qualcomm ended fiscal 2026 with USD 4.6bn in cash, down USD 884m from USD 5.5bn. Inventories rose by more than that over the same year. I would not claim the stock alone drained the cash: an acquisition of USD 2.3bn and higher research spending also need paying for. But the shelves and the bank account moved in opposite directions, and the shelves won.

The balance sheet as a whole is not in trouble. Equity rose USD 9.2bn to USD 30.4bn, and it now funds 50.0% of total assets of USD 60.8bn, up from 42.3%. Part of that equity growth reflects the tax benefit that inflated profit. A company with that cushion can carry extra stock for a while without anyone worrying about its bills.

The risk is subtler. Chips age fast. A smartphone processor is designed for one generation of phones, and when the next generation arrives, last year's part sells at a discount or not at all. If stock cannot be sold at its cost, accounting rules require the company to write it down, and the loss goes straight through the cost of sales. That would hit the operating margin, the one that already fell by 7.8 pp.

So the stock is not only cash on hold. It is also a possible future charge, and the timing of that charge depends on how quickly the phone market recovers and how quickly the new products sell.

What would settle the question

A few things would tell us which reading is right.

The first is the inventory line in the next quarterly reports. Stock built for new products tends to fall as those products ship. Stock left over from weak phone demand tends to linger, or to shrink only through discounts. If inventory days drift back towards the 54 days of fiscal 2025, the build was timing. If they stay near 73 days or rise, it was not.

The second is the gross margin and any inventory write-down in the notes. A chip company clearing unwanted stock usually shows it there first, as a charge or as a thinner margin on each chip sold.

The third is the data centre revenue itself. A target of USD 5.0bn for fiscal 2027 is concrete enough to check. If Qualcomm reaches it, the higher spending on research, the Alphawave deal and some of the stock will look like the price of a new business. If it falls short, all of it will look like costs carried on a shrinking core.

The accounts do not settle it today, and the company has not said how its stock divides between old and new product lines.

What it means for phone makers, rivals and the rest of us

For phone makers, a supplier with full shelves in a weak market is good news at the negotiating table. Excess stock tends to make chips easier to get and easier to discount. That may help the Android brands that buy from Qualcomm hold prices on mid-range phones, at a time when MediaTek already leads the market on volume.

For the market, Qualcomm's year captures a turn many chip companies are making. The smartphone is no longer the growth story. The money is moving to data centres and artificial intelligence hardware, and suppliers are paying to get there: through acquisitions, through larger research budgets and, in Qualcomm's case, through warrants handed to a cloud customer. The phone business is being asked to fund the move while it shrinks.

For anyone running a business that holds stock, the lesson is close to home. Divide your inventory by your sales and multiply by the days in a year, as here, and compare it with last year. Then list the stock by age and by how fast each item sells, and ask how much of it you could not sell at full price within the next season. The total rarely tells you whether you are prepared or stuck. The ageing does.

My verdict: Qualcomm's record net profit is a tax story, and the business underneath earned less on slightly lower sales. The stock build is most consistent with a phone market that fell faster than Qualcomm could slow its orders, with a strategic bet on new products layered on top. I would watch the inventory days over the coming quarters, because they will show whether those shelves hold the start of a data centre business or the end of a phone cycle.