- Buys
- 0 of 4
- Three are needed, so Qualcomm stays out of the basket.
- Fourth-quarter operating income
- $1.36B
- Against $2.92B a year earlier, on revenue down 8.4%.
- Free cash flow yield
- 4.67%
- Trailing $8.65B over a $185.31B market cap, under the 4.95% 10-year yield.
- Licensing revenue, fiscal 2026
- $5.56B
- Down 0.5% in a year the chip business shrank.
Durable quality, long horizonThe royalty stream is excellent, but owner earnings after stock pay yield about 2.8%.
Deep value, margin of safetyThe seat's model failed, so no vote was cast.
Macro and reflexivityThe handset damage is filed; the data center upside is not.
Growth investingAutomotive is a real runway, but 21.4x free cash flow is too much for a shrinking core.
Session of November 5, 2026. Three seats voted; Investor B's model failed and no vote was cast for it. We could not open the results release, the 10-K text or the proxy today, so the tax item and the outlook range come from press reports and the holder table is missing. This is the first Qualcomm review.
Bottom line
Qualcomm's fiscal 2026 annual report, filed November 4, 2026, flatters the year. Net income of $13.63B sits above operating income of $8.7B, so the 13.6x earnings multiple is not a clean number. On free cash flow the business yields 4.67%, a little under the 10-year Treasury.
Our arithmetic says that price needs free cash flow to grow 4.6% to 9.0% a year for a decade, depending on the discount rate. All three seats that answered expect less than that by year three. Hence three Watches and no Buy.
The seats split on what would change the verdict: the phone cycle, the first filed data center revenue, or China trade. The first test is the first-quarter fiscal 2027 10-Q, the first filing with a Data Center segment inside the guidance.
The business in 60 seconds
Qualcomm has two businesses (10-K, filed November 4, 2026). QCT sells chips for phones, cars, connected devices and now data centers: $37.21B of fiscal 2026 revenue, down 3.0%. QTL licenses the patent portfolio to device makers: $5.56B, down 0.5%. Segment operating income is not tagged, so the filed figures cannot show where the profit sits.
QCT is fabless apart from some RF parts, relying on TSMC, Samsung and GlobalFoundries. The 10-K names Apple, Samsung and Chinese handset makers as the concentrated premium customer base, and calls its China exposure significant.
There is one class of common stock. The proxy's ownership table would not load for us, so we cannot say whether any holder is above 10%, and nothing we read shows a controlling holder.
| Fiscal year | Revenue | Operating margin | Free cash flow | Diluted shares |
|---|---|---|---|---|
| 2022 | $44.2B | 36.0% | $6.8B | 1.1B |
| 2023 | $35.8B | 21.8% | $9.8B | 1.1B |
| 2024 | $39.0B | 25.9% | $11.2B | 1.1B |
| 2025 | $44.3B | 28.0% | $12.8B | 1.1B |
| 2026 | $43.1B | 20.1% | $8.65B | 1.1B |
10-K fiscal year rows, SEC EDGAR. Operating margin is our division of operating income by revenue.
What changed
- The fourth quarter halved operating income. Revenue for the quarter to September 27, 2026, was $10.32B against $11.27B (SEC EDGAR, fiscal year less nine months).
- A tax item inflates the year. Press reports from November 4, 2026, describe a one-time $5.7B tax benefit from releasing a valuation allowance. We have not confirmed it in the filing.
- Data center enters the guidance. The results release footnotes say first-quarter fiscal 2027 guidance includes Data Center and pending business combinations expected to close in the quarter (8-K Exhibit 99, filed November 4, 2026).
- Handsets down, cars up. Press reports put fourth-quarter handset chip revenue down 26%, blamed on memory costs, and automotive up 89% to $1.99B.
- The customer grew while suppliers shrank. Apple's 10-K, filed November 3, 2026, shows revenue of $478.83B, up 15.1%. Skyworks, an RF supplier to the same phones, saw quarterly operating income fall to $48.5M from $110.6M (10-Q, quarter to July 3, 2026).
The bull case on Qualcomm
- The royalty stream held. Licensing fell 0.5% while chips fell 3.0%, and the press reported long renewals with Apple, Huawei and Transsion.
- Debt is not the problem. Net debt is $6.93B, with $2.00B due in the next twelve months and $10.45B after year five.
- The pivot is funded in filed cash. Capital spending doubled to $2.49B, 1.53x depreciation, and purchase commitments reached $18.60B.
Companies that are bluffing do not double capex while revenue falls.
Investor C, on the filed spending behind an unfiled story
When memory cost inflation subsides, automotive together with nascent data center silicon can drive an earnings inflection far faster than the market expects.
Investor D, Watch 7
The bear case on Qualcomm
- The cheap multiple is an accounting effect. Strip the tax item and the discount goes. On free cash flow the price is 21.4x.
- Stock pay eats the cash. Stock compensation of $3.45B is 8.0% of revenue. Taken out, free cash flow is about $5.2B, near 2.8% of market cap.
- Buybacks mostly stood still. Four fiscal years of buybacks cost $23.70B and the diluted count fell 5.6%.
- Concentration is the company's own warning. The 10-K says losing a key handset customer would have a significant negative impact.
The headline 13.6x earnings multiple is not real.
Investor A, on the tax benefit
Net today the payoff is roughly balanced, which is not the asymmetry I require, and that is why the verdict is Watch rather than Buy.
Investor C, Watch 6
Where the committee splits
Three identical votes, three different tests.
Does the phone cycle or the data center decide Qualcomm's next three years?
| Position | Seats | The figure it rests on |
|---|---|---|
| The phone replacement cycle | D | Quarterly operating income down 53.5%, SEC EDGAR |
| Memory costs and the first data center line | C | Data Center inside first-quarter guidance, 8-K footnote |
| US and China trade | A | China concentration, 10-K Item 1A |
What settles it. The segment note in the first-quarter fiscal 2027 10-Q. C breaks below roughly $1.0B of Data Center revenue for the quarter; D breaks if automotive falls under $1.50B.
From the chair, to Investor C. Investor D thinks the handset cycle, not the data center, decides the verdict. What handset figure in the same 10-Q would override a strong Data Center print?
What is the real cash yield?
| Position | Seats | The figure it rests on |
|---|---|---|
| Stock pay is a cost | A, D | $3.45B, 8.0% of revenue, 10-K |
| Reported cash flow, checked against credit | C | 4.67% yield against a 3.40% high yield spread, FRED |
What settles it. Free cash flow less stock compensation in the fiscal 2027 10-K. A breaks below $4.0B.
From the chair, to Investor A. Licensing held through a quarter when handset chips reportedly fell 26%. Why does your 2.8% owner-earnings yield charge the royalty stream for the chip business's stock pay?
Can revenue growth come back?
| Position | Seats | The figure it rests on |
|---|---|---|
| Yes, 5% base and 12% bull | D | Reported automotive growth, press, November 4, 2026 |
| Flat to 3% at best | A, C | Revenue down 0.6% a year over four years, 10-K rows |
What settles it. Fiscal 2027 revenue and its non-handset share, in the 10-K.
From the chair, to Investor D. Your bull case needs 12% a year. Which filed line supports more than the minus 0.6% four-year rate?
| Investor | Verdict | What breaks its thesis |
|---|---|---|
| A | Watch 6 | Fiscal 2027 licensing under $5.0B, free cash flow less stock pay under $4.0B, operating margin under 15%, or no net share reduction, each to Avoid. Net debt above $15B or goodwill above $25B after the deals. Critical: US and China trade, to Avoid. |
| B | No response | No vote was cast. |
| C | Watch 6 | First-quarter Data Center revenue under about $1.0B. Gross margin under 50% for two quarters. Licensing down over 10% in any quarter. Critical: memory costs and AI spending, toward Buy if memory inflation breaks, toward Avoid if it deepens. |
| D | Watch 7 | Quarterly automotive revenue under $1.50B or operating margin under 12.0% in the first-quarter 10-Q. Handset revenue down over 35% in the second quarter. Critical: the phone replacement cycle, a confirmed turn to Buy. |
Valuation
What the price requires. This is the desk's arithmetic, which the seats never saw. At 9.45%, the 10-year yield plus 4.5 points, the $185.31B market cap needs trailing free cash flow to grow 6.9% a year for ten years, then 3%. Year three lands at $10.57B. At 8.45% it is 4.6% a year and $9.90B. At 10.45% it is 9.0% and $11.20B.
Net of stock pay, a $5.20B base, the price needs 11.2% to 15.9% a year. The seats used reported free cash flow, so we do too. On enterprise value of $192.24B the yield is 4.50%, against a 4.95% 10-year Treasury yield.
Hatched rows show what the price needs at each discount rate. Each seat row is its bear, base and bull cases multiplied by its own odds.
| Investor | Bear | Base | Bull | Weighted, year 3 | Against the price |
|---|---|---|---|---|---|
| A | 30%, about $6B | 50%, $9B to $11B | 20%, above $13B | $9.4B | below the band |
| C | 30%, $5B to $6B | 50%, $8B to $9B | 20%, $11B to $12B | $8.2B | below the band |
| D | 30%, $5.5B | 50%, $9.5B | 20%, $14.5B | $9.3B | below the band |
Where a seat gave a range, the weighted figure uses its middle. A's bull case, above $13B, is counted at $13B.
All three seats sit under the lowest hurdle, and none puts fiscal 2029 back at the 2025 level of $12.8B in its base case. Investor A put the gap plainly:
This is a good business with one excellent component, at a full price. I will watch it, not buy it.
Investor A, Watch 6
| Investor | Entry discipline |
|---|---|
| A | Free cash flow yield after stock pay of 6.5% or more on five-year normal figures, or enterprise value at or under 12x five-year average operating income. |
| C | Trailing free cash flow yield near 6%, about 16x, once quarterly operating margin stops falling. |
| D | Enterprise value at or under 14x normal free cash flow, or a 7.0% yield net of stock pay. |
Comparables
| Company | Market cap | Growth last FY | Operating margin | P/E | FCF yield |
|---|---|---|---|---|---|
| Qualcomm (QCOM) | $185.3B | -2.6% | 20.1% | 13.6x | 4.7% |
| Broadcom (AVGO) | $1.55T | 23.9% | 39.9% | 40.6x | 2.5% |
| Skyworks (SWKS) | $12.5B | -2.2% | 12.2% | 43.0x | 3.5% |
| Marvell (MRVL) | $140.1B | 42.1% | 16.1% | 53.1x | 1.2% |
| Intel (INTC) | $309.7B | -0.5% | -4.2% | n/a | 1.2% |
Built the same way for all five from SEC filings, priced at the Yahoo Finance close. Intel lost money, so it has no P/E.
The custom data center chip makers, Broadcom and Marvell, grow fast on thin yields. Skyworks and Intel, tied to phones and PCs, shrank and still cost more than Qualcomm on cash. Qualcomm is priced with the handset group, and its bull case is a move into the other one. Arm and MediaTek are left out because they are not US domestic filers.
Key risks
Ranked by us, most severe first. The third column names who raised each risk.
| Risk | Why it matters | Raised by | First signal |
|---|---|---|---|
| High. Customers build their own chips | Each loss removes QCT revenue for good | Investors A, C, D; Analyst; 10-K Item 1A | Customer shares in the 10-K concentration note |
| High. China and export controls | Hits chips and royalties at once | 10-K Item 1A; Analyst; Investors A, D | New US export rules on mobile or AI chips |
| High. Memory and wafer cost squeeze | Orders fall and margin compresses together | Analyst; Investors C, D; reported by the press, not confirmed in a filing | Gross margin under 50% for two quarters |
| High. Foundry concentration in Asia | No quick substitute for leading-edge wafers | 10-K Item 1A; Analyst | Dual sourcing for flagship parts |
| Medium. Data center spending earns too little | $18.60B of commitments, $16.37B of goodwill | Analyst; Investors A, C; desk arithmetic | Goodwill above $25B or net debt above $15B |
| Medium. Licensing model and Arm | Royalties may carry most of the profit | Analyst; Investor C; forum (unverified) | The Delaware post-trial ruling |
The debt schedule is not a risk on these filings.
What to watch
| Date | Event | What it decides |
|---|---|---|
| Ongoing | First-quarter fiscal 2027 10-Q, due 40 days after the quarter ends under SEC rules | Data Center and automotive revenue, operating margin |
| Ongoing | Second-quarter fiscal 2027 10-Q, same rule | Handset revenue against D's 35% line |
| Ongoing | Delaware post-trial rulings in the Arm case, no date set | Whether a royalty suspension applies |
| Ongoing | Fiscal 2027 10-K, due 60 days after year end for a large accelerated filer | Licensing, stock pay, margin, share count |
Owning it onchain
xStocks issues QCOMx, a token tracking Qualcomm. We have not judged whether it is investable: our only quote came before regular US trading opened on November 5, 2026, when market makers quote thinly. The Backpack token, QCOM.US, is not investable because it is not verified on Jupiter.
xStocks says its tokens are not offered in the United States or to US persons. Sunrise blocks users in the United States and Canada from Backpack's tokens. This is not a recommendation to use either.
Data
| Item | Figure | Source |
|---|---|---|
| Price | $173.19 | Yahoo Finance close, November 4, 2026 |
| Shares outstanding | 1,070,000,000 | SEC EDGAR cover, as of November 2, 2026 |
| Market cap | $185.31B | Price times shares |
Method. Four investors with different philosophies, each on a different AI model, read one identical fact sheet and voted independently. A name needs three Buys to enter a basket. How the committee works.
Correction. This article first called QCOMx investable on a 0.68% round-trip cost. That quote was taken before regular US trading opened, when market makers quote thinly. We now judge that cost only during regular US market hours, and we have removed the figure, the verdict and the mint.
As of November 5, 2026, BagsCapital manages no outside capital and holds no position in Qualcomm. Nothing here is investment advice.