Built by Claude Fable

QUALCOMM Incorporated | 5.2/10 | $174.71

HOLD

NASDAQ: QCOM | FQ3'26 proved the mix shift, not the reacceleration: auto grew 61% y/y and non-handset reached 40% of QCT, but handsets fell 19.6%, total revenue fell 4.0%, gross margin hit a 12-quarter low and quarterly FCF dropped 81%. The test is December-quarter data center revenue.

FQ3'26 Revenue

$9,947M

-4.0% y/y | decelerating six straight quarters

Automotive

+61.4%

$1,588M | accelerating from +15% and +38%

Data Center

$88M

Quarterly, vs $5B FY27 and $15B FY29 targets

FQ3'26 FCF

$495M

-80.8% y/y | TTM FCF down ~10%

Company overview

Qualcomm designs the Snapdragon system-on-chip and cellular modem platforms used in most premium Android phones, and licenses the cellular standard-essential patent portfolio behind 3G, 4G and 5G handsets. The chip business (QCT) is about 85% of revenue and reports Handsets, Automotive and IoT (which holds AI PC, XR and industrial). The licensing business (QTL) is about 13% of revenue at a ~70% EBT margin. Fiscal year ends late September; FQ3'26 is the June 2026 quarter, reported 2026-07-29. See Business model for the end-market map.

Price / Market cap	$174.71 / $183.4B (FMP intraday, 2026-10-08)	Enterprise value	~$194.2B (FMP TTM); net debt/EBITDA 0.8x
CEO / CFO	Cristiano Amon (CEO since 2021) / Akash Palkhiwala (CFO since 2019, COO since 2024)	FQ3'26 Revenue	$9,947M (-4.0% y/y); QCT $8,504M, QTL $1,278M
FQ3'26 Segment mix	Handsets 51% / IoT 18% / Auto 16% / QTL 13% of revenue	FQ4'26 Guide	Revenue $9.7B to $10.5B; non-GAAP EPS $2.05 to $2.25
Street FY27E	Revenue $45.0B / EPS $10.13 (below FY26E $10.54)	Forward multiples	17.3x FY27 P/E, 12.7x FY27 EV/EBITDA (peer medians 32.7x / 25.0x)
China revenue	~46% of FY25 revenue (62% to 67% on the old ship-to basis)	Ratings	~30% Buy share; consensus Hold; 8 of 9 post-print target changes were cuts
Fiscal year end	Last Sunday of September	Quality gate	1 NO (positive and growing FCF); no cap applied

Score breakdown

2.5

/ 10

Financial TrendsWeight: 25%

FQ3'26 did NOT show reacceleration at the company level. Total revenue fell 4.0% YoY after -3.5% in FQ2'26 and +5.0% in FQ1'26, GAAP operating income fell 41% YoY, and quarterly FCF collapsed 81% YoY to about $0.5B on a ~$1.0B inventory build. The one thing that did change, and it is real, is the AI-adjacent non-handset mix: Automotive reaccelerated to +61% YoY (from +15% in FQ1'26 and +38% in FQ2'26) and Auto+IoT combined accelerated for a third straight quarter (+11% to +20% to +28%). That is a segment-level inflection buried under a handset downturn (memory-cost driven smartphone contraction plus the Apple modem exit) and input-cost gross-margin pressure. Fiscal year ends late September; FQ3'26 = quarter ended June 2026.

/ 10

Thematic ExposureWeight: 35%

Qualcomm has real oligopoly positions (premium Android SoC, cellular modem, auto digital cockpit, cellular licensing, smart glasses), so the gate is passed. In FQ3'26 the AI story showed up in revenue in only one place: auto, up 61% y/y. Data center is still pre-revenue, at roughly $0.1B a quarter. The core handset franchise is shrinking: down 20% y/y, Apple is leaving, and Exynos and MediaTek are taking share. What changed in FQ3'26 is better visibility (purchase orders, wafer starts, the BMW ADAS win, a higher auto run rate), not a reacceleration in AI revenue.

/ 10

Management QualityWeight: 20%

Amon and Palkhiwala have hit or beaten the midpoint of their quarterly revenue and non-GAAP EPS guidance in 12 of 12 quarters (24 of 24 data points) going back to FQ4'23, and the FY24 Investor Day diversification promises are mostly on track, led by auto. The score stops at 7 because three things slipped: IoT is running well behind the growth its FY29 target needs, the QCT margin is below its own 30% long-term bar, and two recent forward statements were walked back within one to two quarters (FY26 growth, Apple FY27). On top of that, insiders sold more than $10M over the last 12 months and bought nothing. The FQ3'26 call raised the bar sharply: FY29 non-handset went from $22B to $40B, the auto exit run rate from $6B to $7B, and data center is now $5B in FY27. With a 3-year record behind them, that raise is believable, but they now have to deliver it.

/ 10

Investor Sentiment (Inverted)Weight: 5%

This is a real management-versus-street gap, and it sits in AI. Management is loudly and specifically bullish on data center AI inference ($5B FY27, $15B FY29) and on a $40B FY29 non-handset business. The sell side is mostly unconvinced: Buy ratings have fallen from about 55% to about 30% of coverage, and price targets were cut after FQ3'26. Most of the doubt is aimed at data center, which today brings in less than $100M a quarter. What keeps this short of an 8 to 9: no insider has bought stock in the open market since 2019, the CEO and CFO sold steadily all through 2026, and Qualcomm has tried data center twice before without success (Centriq server CPU, Cloud AI 100).

/ 10

Concerns / Catalysts / RisksWeight: 15%

QCOM trades at roughly half the peer multiple on FY+1 earnings, and the next 90 days hold an unusually dense set of hard catalysts: the Arm trial verdict (in progress now), the FQ4'26 report in early November, first data center ASIC revenue in the December quarter, and the Apple and Huawei license renewals signed in the last two weeks. Against that, China ship-to revenue is about 46% of the total, the Apple modem roll-off sped up in FQ3'26, Samsung is pushing Exynos back into the Galaxy S27, and three separate regulatory and legal overhangs remain open (Arm, China SAMR, US export controls). The rubric puts the cheap valuation and catalysts against the China and regulatory exposure. That lands at a 5.

Dimension	Score	Weight	Weighted
Financial Trends	2.5	25%	0.63
Thematic Exposure	6	35%	2.10
Management Quality	7	20%	1.40
Investor Sentiment (Inverted)	7	5%	0.35
Concerns / Catalysts / Risks	5	15%	0.75
Composite		100%	5.23

Detail pages: Financials, Thematics, Management, Sentiment, Valuation and risks.

Quality gate

Gate test	Result	Source dimension
Oligopoly (>30% share in key markets)	YES: premium Android SoC, cellular baseband ex-Apple, auto digital cockpit, cellular licensing	2. Thematic
Positive and growing FCF	NO: FCF is positive, but TTM FCF is down ~10% y/y (~$10.4B vs ~$11.6B) and FQ3'26 FCF fell 81% to ~$0.5B (OCF $991M less capex $496M)	1. Financial
Management track record	YES: 24 of 24 quarterly revenue/EPS guides at or above midpoint; 94% hit rate on promises 4+ quarters old; stable CEO/CFO	3. Management (confirmed by 4)

Gate flag: One NO (positive and growing FCF). No cap applied; scored normally with the FCF gap noted. The FCF miss comes from a deliberate ~$1.7B two-quarter inventory build ($703M + $996M), capex up 69% y/y, and acquisition-related OpEx (Alphawave, Modular) ahead of data center revenue. It clears if FCF returns to growth as inventory normalizes and data center revenue begins.

Recommendation: HOLD (5.2/10). Real oligopoly assets, a credible management team and a clear management-versus-street gap on AI are offset by a shrinking, share-losing handset core, compressing margins, falling FCF and 46% China revenue exposure. Upgrade triggers: FQ4'26 (early November) showing a double-digit sequential China handset rebound, plus first hyperscaler custom silicon revenue in the December quarter.

Data sourced from Daloopa. Market data, consensus, ratings and insider data from FMP (/stable). Transcripts FQ4'24 to FQ3'26 from FMP.

AI reacceleration verdict

Answer: not yet proven at the company level. FQ3'26 proved the mix shift, not the reacceleration. Non-handset QCT (auto plus IoT) accelerated for a third straight quarter, from +11.1% to +20.2% to +28.3% y/y, and reached 40% of QCT versus 30% a year ago. But that growth is being swamped by a handset business down 19.6% y/y. Part of that decline is caused by AI itself: memory reallocated to data center HBM is squeezing Android smartphone builds. Total revenue went from +5.0% to -3.5% to -4.0% y/y.

Market share by end market

End market	FQ3'26 rev ($M)	y/y	% of total rev	QCOM share / position	AI driver	Proven or narrative
Handsets (QCT)	5,086	-19.6%	51%	#2 in total smartphone AP at 23% (MediaTek 31%, Apple 19%, Exynos 9%; Counterpoint CQ2'26), down from 27%. Leader in premium Android; ~70% of Galaxy flagships	On-device / agentic AI raising ASP	Reversed. FY25 handset growth was ASP-led (+$2.5B price per chipset); in FY26 OEMs are trading down to prior-gen chips
Modem in Apple	Inside handsets (~$7.5B FY26 est.)	Stepping down	n/a	Share in fall 2026 iPhones "materially lower" than 20%, heading to ~0	None	Loss, pulled forward to FQ4'26
Automotive	1,588	+61.4%	16%	Leader in digital cockpit; challenger in ADAS (BMW lead compute win in FQ3'26)	Compute content per vehicle for ADAS / AI cockpit	Proven. Growth went +14.6% to +38.3% to +61.4%; exit run rate raised from $6B to ~$7B; FY29 target $10B
IoT / edge AI / industrial	1,830	+8.9%	18%	Fragmented (NXP, MediaTek, Nvidia Jetson, Intel); near-monopoly in smart glasses	Edge inference, robotics, XR	Partial. Steady high single digits, below the ~21% CAGR its $14B FY29 target needs; FQ4 guided ~flat
AI PC (Snapdragon X, inside IoT)	Not disclosed	n/a	Low single digits (est.)	~9% of US retail Windows laptops above $600 (CQ1/CQ2'25, not updated); leads Windows on Arm	Copilot+ NPU	Narrative. ~150 designs, no disclosed revenue
Data center	88	new	~0.9%	~0% versus Nvidia (dominant), Broadcom and Marvell (custom ASIC)	Inference ASICs, AI200/AI250, HBC accelerator, CPU, SerDes	Narrative backed by POs. Target $5B FY27 and $15B FY29; revenue starts in the December quarter
Licensing (QTL)	1,278	-3.0%	13%	Dominant cellular SEP licensor	None directly (tied to handset ASP)	Flat. Apple renewal (from 2027-04) and Huawei 5G cross-license signed since the call reduce the overhang

What changed in FQ3'26

Signal	Before FQ3'26	FQ3'26	Direction
Auto y/y	+14.6% / +38.3%	+61.4%, ~+60% guided for FQ4	Accelerating (proven)
Non-handset (auto+IoT) y/y	+11.1% / +20.2%	+28.3%; FY27 guided >60% including data center	Accelerating
FY29 non-handset target	$22B	$40B (Investor Day, June 2026)	Nearly doubled
Data center	"Initial shipments later this calendar year"	POs from 2 hyperscalers, wafers started, HBC Gen 1 taped out	Firmer visibility, no revenue yet
Handsets y/y	+3.3% / -13.1%	-19.6%; China OEMs said to have bottomed	Worse, bottom claimed
Apple	FY27 product revenue "a little over $2B"	Below that; step-down starts FQ4'26	Worse, pulled forward
QCT EBT margin	31% / 27%	26%, guided 23% to 25%	Compressing
Street	Buy share ~45% (Jan 2026)	~30%; 8 of 9 post-print target changes were cuts; FY27E EPS $10.13 below FY26E $10.54	More skeptical

Proven: auto is an AI-content growth engine that is accelerating, with multi-year lock-ins (BMW, Stellantis). Management has hit or beaten the midpoint of its revenue and EPS guidance 24 of 24 times over 12 quarters, so its targets deserve weight.

Narrative: data center ($88M per quarter against $5B in FY27, roughly 14x the current run rate), AI PC (no disclosed revenue) and agentic handset ASP uplift (it went backwards in FY26).

Why this still matters for the thesis: FQ3'26 widened the gap between what management says and what the street believes. Management is repeating specific, checkable AI milestones, while consensus has FY27 EPS below FY26 and barely models FY29. That is the "management is the contrarian" setup. It falls short of a high conviction call for three reasons: Qualcomm's own failed data center attempts (Centriq, Cloud AI 100), zero insider buying, and a deteriorating consolidated P&L. If December-quarter data center revenue shows up on schedule and China handsets rebound double digits sequentially as guided, the reacceleration case becomes proven. If not, the stock is a value trap.

Cross-dimension reconciliation notes

Data center scale-up: dimension 2 says "~17x" (versus ~$0.3B FY26 total) and dimension 4 says "~14x" (versus the annualized $88M quarterly run rate). Both are consistent; this report uses 14x on the run rate.
FQ3'26 buyback: $1,364M (cash flow statement, dimension 1) versus $1,400M (rounded call figure, dimension 4). Same item; the $1,364M cash figure is preferred.
Insider selling: dimension 3 cites ~$48.4M of trailing 12-month sales (including Amon's 150k-share sale on 2025-10-01); dimension 4 counts only calendar 2026 sales. Neither shows any open-market buying since 2019.
FQ3'26 handset revenue appears under two Daloopa IDs (174830187, 174829355). Both are $5,086M.

PM Review

Reviewed 2026-10-08 against pm-review and investing-principles. Thesis angle under test: Qualcomm benefits from AI, market share by end market, and whether FQ3'26 (reported 2026-07-29) proved AI-driven reacceleration.

Verdict: the conclusion is right and honestly stated ("FQ3'26 proved the mix shift, not the reacceleration"), and the 5.2 / HOLD stands, but the work over-credits the AI evidence it has. Strip out what is not organic AI revenue and the FQ3'26 proof set shrinks to one segment (auto, 16% of revenue) whose growth has not been decomposed into AI content versus price increases versus cockpit volume. Data center revenue is acquired Alphawave connectivity, not Qualcomm AI silicon. The FQ4'26 guide (about -10% y/y total revenue) means company-level deceleration extends another quarter, so there is no sentiment inversion yet. This is a "management is the contrarian" setup, not a proven reacceleration.

1. Gaps to the pm-review

Price talk in a fundamental review. The header leads with $174.71; sentiment discusses average and median price targets versus the current price, individual broker target dollar changes, and the "52-week low near $122." The rule is fundamentals only. Keep the count of target cuts (8 of 9) as a sentiment signal; drop the dollar levels and implied-upside math.
Year-over-year growth is missing on several tables. The spending table (R&D, capex, buybacks) shows y/y only in the last column; the annual 5-year table has no y/y bps rows for gross or operating margin and no y/y rows for handsets, IoT or QTL beyond a trend-column note; the China revenue table has no y/y; the guidance accuracy table never shows the guide's implied y/y (FQ3 and FQ4 guides imply about -7% and -10%, mentioned only in prose). Every metric table needs a y/y row.
Sequential comparisons crept in. "Apple revenue down roughly 50% sequentially," "China handsets double-digit sequential growth," "data center about 14x the current run rate," and the $97M to $88M data center move are q/q or run-rate framings. Restate on a y/y basis (for example, the ~$5.2B FQ4'26 handset guide versus $6,962M in FQ4'25 is about -25% y/y, worse than FQ3's -19.6%).
Internal consistency error. The promise tracker says FQ2'26 revenue was -2.2% y/y; every other table and the math ($10,599M vs $10,979M) say -3.5%. FY25 revenue is cited as both $44,284M and $44.14B without explanation. Reconcile.
Street consensus is company-level only. The thesis is about segments, yet there is no consensus for FY27 auto, IoT, handsets or data center. The single most important number for "the street does not believe management" is street FY27 data center revenue versus the $5B target. Without it, the divergence is inferred from EPS direction and analyst questions, not measured.
Acceleration is explained well for auto and margins, not for IoT. IoT has been ~+9% for four quarters and is guided flat for FQ4, with no breakdown of which sub-line (PC, XR, industrial, consumer) is offsetting which. That is the edge AI and AI PC leg of the thesis and it is unexplained.

2. Gaps of Known Skills (investing-principles)

"Leaders remain leaders / don't settle for 2nd" is applied too generously. Qualcomm is #2 in total smartphone AP (23% vs MediaTek 31%) and losing share (from 27%). The oligopoly gate is passed on premium Android SoC, but the only sourced figure is 28% of all premium chipsets including Apple, which is below the 30% bar. "Est. majority" of premium Android and of digital cockpit are unsourced. "Cellular baseband ex-Apple" is a carve-out chosen to clear the bar. The gate may still be YES, but as written it is not proven. Auto is the only end market where leadership in a growing theme is evidenced, and even there no share number is given.
Market share is a snapshot, not a trajectory. Only handset AP share has a y/y comparison. Auto cockpit, ADAS, IoT, AI PC (stale CQ2'25 data, over a year old) and licensing share have no y/y change. The user explicitly asked for market share across end markets; this is the largest gap against the ask.
AI attribution is not proven for auto. Auto +61.4% y/y is labelled "Proven" AI, but it is not split into (a) ADAS/AI compute content, (b) cockpit and connectivity volume, (c) the double-digit chip price increases the report itself calls cost pass-through, and (d) any inorganic contribution (Autotalks). Management's "order of magnitude" content claim is quoted, not measured. If a large share of the acceleration is price pass-through, the one proven AI leg weakens materially.
Data center revenue is inorganic. The report notes the $88M is "mostly Alphawave connectivity" but still presents it as the AI data center base. Organic Qualcomm AI data center revenue in FQ3'26 is roughly zero, and there is no pro forma y/y for the acquired base.
The "AI hurts the handset core" causal claim is unsourced. HBM reallocation is plausible, but there is no DRAM contract price y/y, no global smartphone unit y/y (IDC/Counterpoint), and no FQ3'26 units vs ASP split (the 10-K bridge is annual). Handsets ex-Apple y/y is never isolated, so the reader cannot tell how much of -19.6% is Apple, China destock, memory, or share loss to MediaTek and Exynos.
Sentiment inversion is not present, and the report should say so plainly. Total revenue y/y went +5.0%, -3.5%, -4.0%, and the FQ4'26 guide midpoint ($9.7B to $10.5B) versus $11,270M implies about -10%. Handsets are still worsening. Worse to better has not started; the earliest inflection is FQ1'27, which laps the toughest comp in the series ($12,252M) with the Apple step-down fully in.
Data source priority skipped. No check of internal SharePoint memos, analyst Excel models, Teams, OneNote or Outlook (Priority 1 and 2 sources). Consensus is FMP, not Visible Alpha or Bloomberg; disclosed as a fallback, but it leaves segment consensus and revision trends absent. No Snowflake alt data (smartphone sell-through, auto builds) to test the China bottom claim.
Valuation framework incomplete. Only P/E and EV/EBITDA are shown. EV/Sales and P/FCF (next year) are required, and P/FCF matters most here because FCF is the failed gate.
Inventory "strategic hedge" is management's framing, not tested. A ~$1.7B two-quarter build alongside falling revenue needs inventory days y/y and a comparison to prior cycle builds (FY22 to FY23) before it is accepted.

3. Action Items

Build the FQ1'27 bridge. Quantify what auto, IoT, data center and handsets ex-Apple must each print for total revenue y/y to turn positive against $12,252M. That number is the reacceleration test, not "data center shows up."
Decompose auto +61% into ADAS/AI compute vs cockpit/connectivity, price vs units, organic vs inorganic, using 10-Q disclosure, transcripts and design-win ramps. Only the AI-content piece counts as proof for this thesis.
Re-label data center. Show organic Qualcomm AI data center revenue (about zero) separately from Alphawave, and add Alphawave pro forma y/y.
Pull segment consensus from Visible Alpha (Bloomberg once available) for FY27 auto, IoT, handsets and data center, plus 3-month revisions. Put street FY27 data center next to the $5B target; that is the contrarian gap in one cell.
Turn the market share table into a trajectory. For each end market show current share, year-ago share and bps change, with a source: smartphone AP (have), premium Android SoC (Android only), auto cockpit and ADAS (S&P Mobility or TechInsights), Windows on Arm / AI PC (IDC or Mercury, current quarter), smart glasses, baseband ex-Apple. Mark unsourced estimates and re-run the oligopoly gate on sourced numbers only.
Isolate handsets ex-Apple y/y and add FQ3'26 units vs ASP, DRAM contract price y/y and global smartphone units y/y, to prove or drop the memory squeeze causality.
Explain IoT by sub-line (PC, XR, industrial, consumer) so the edge AI and AI PC legs are evidenced or explicitly marked narrative.
Remove price and price-target dollar discussion; keep rating mix and count of target and estimate revisions.
Add y/y rows to every table (spending, annual margins in bps, China, guidance-implied growth) and restate all q/q framings as y/y.
Fix the -2.2% vs -3.5% FQ2'26 error and footnote the $44.28B vs $44.14B FY25 tag difference.
Complete valuation with FY27 EV/Sales and P/FCF versus peers.
Check Priority 1 internal sources (SharePoint, Excel models, OneNote, Outlook) for prior QCOM work and broker color, and record what was found.

Score view: keep the 5.2 composite and HOLD. Thematic at 6 is generous for a #2 player losing share in its 51%-of-revenue segment; 5 is defensible. If the gate is re-run on sourced share data and premium Android lands below 30%, the oligopoly gate joins FCF as a NO and the composite should be capped. The upgrade trigger is organic AI revenue (auto AI content plus Qualcomm-designed data center silicon) large enough to turn total revenue y/y positive in FQ1'27 or FQ2'27.

Data sourced from Daloopa (QCOM company_id 157). Market data, consensus, ratings and insider data from FMP (/stable), 2026-10-08. Transcripts FQ4'24 to FQ3'26 from FMP. Market share inputs from Counterpoint, TechInsights and company disclosures as linked on the Thematics page. Scoring workspace: tickers/QCOM/data/review_workspaces/2026-10-08/final_score.md.