QCOM — QUALCOMM Incorporated

As of 2026-09-13 the tape is clearly bullish: the September 8, 2026 AWS multi-generational pact to build custom AI data-center silicon and optical networking (with warrants for up to 25M shares at $161.26 tied to deployment milestones worth up to $60B) drove an 8-9% intraday surge and lifted the stock 7.84% in a week and 11.59% in a month. Sell-side has stayed cautious - consensus rating 'Hold' with 34 holds vs 29 buys and 5 sells - but price targets cluster at $204-$213 (average $204.48, recent average $212.80), and bulls argue the ~17x forward multiple is a steep discount to Broadcom and Marvell at 28-32x. The bear side is entirely about the legacy engine: the brief flags fiscal 2027 Apple modem revenue tracking below $2B and Android chip revenue down an estimated 20% in 2026 on memory price inflation and supply constraints.

Trailing fundamentals are decelerating, and the last three quarters show it plainly. Revenue growth went from +10.0% YoY in FQ4-25 ($11.271B vs $10.244B) to +5.0% in FQ1-26 ($12.252B vs $11.669B), then -3.5% in FQ2-26 ($10.599B vs $10.979B) and -4.0% in FQ3-26 ($9.947B vs $10.365B). Gross margin slipped only modestly (53.1% in FQ3-26 vs 55.6% a year ago), but operating margin has fallen for three straight quarters: 27.5% (FQ1-26) to 21.8% (FQ2-26) to 16.3% (FQ3-26), versus 26.6% in FQ3-25 - the driver being R&D up 17% YoY to $2.607B plus the gross-margin dip. Headline EPS is not usable as a trend line: FQ4-25 printed -$2.77 (a ~$6B one-off charge, net income -$3.117B against positive operating income of $2.918B), FQ2-26 printed $6.88 on just $2.309B of operating income (a large non-operating gain), and FQ3-26 printed $1.87. Free cash flow is equally lumpy - $4.416B in FQ1-26, then $1.916B, then just $495M in FQ3-26 (operating cash flow $991M, down 65.5% YoY from $2.875B), the lowest of any of the last twelve quarters. Verdict: the underlying operating business is in a genuine slowdown, not a blip - but it is a structural mix problem (Apple modem roll-off, memory-cost squeeze on Android), not a classic semiconductor cycle trough, and the +40% six-month move is priced on the AWS/auto/data-center future rather than any of these trailing figures.

Fair with a forward-multiple kicker, not cheap on history. As of 2026-09-13, QCOM's trailing P/E of 20.73 sits about 10% above its own 5-year average of 18.89, and P/S of 4.34 is about 10% above the 5-year average of 3.93 - the deterministic score's own valuation overlay calls it 'Neutral vs. its own history'. The PEG of -1.284 is meaningless (negative trailing growth math). The interesting number is forward: next-FY consensus EPS of $10.544 puts the stock at 17.3x, versus Broadcom/Marvell at 28-32x and versus a 2.15% dividend yield and 5.5% trailing FCF yield ($10.4B TTM FCF on a $191.1B cap). Quality metrics support the multiple: 54.2% gross margin, 23.2% operating margin, 21.0% net margin, 26.1% ROE, 13.2% ROIC, 0.55 debt/equity and a 2.02 current ratio. So: fairly priced on trailing earnings, modestly cheap on forward earnings if - and only if - the $10.54 estimate survives contact with the Apple roll-off.

The stock is pricing growth that has not arrived yet, and that is the whole game here. Consensus next-FY revenue of $43.03B is below trailing-twelve-month revenue of $44.07B, yet consensus EPS of $10.544 implies roughly 20% growth off TTM EPS of $8.76 - the models are assuming a margin recovery that the last three quarters of operating-margin compression (27.5% -> 21.8% -> 16.3%) directly contradict. Practical read: at $181.97 you are paying 17.3x for a 'show me' transition. A one-dollar haircut to $9.50 EPS turns that into 19.2x with no price change. The +40% six-month move means the AWS/data-center/auto optionality is already partly in the price - the market is not underpaying for the turnaround. The unusually wide price-target spread ($120 low to $400 high across 25 analysts) is itself the signal: nobody has a settled view on what Qualcomm is worth mid-transition.

As of 2026-09-13, RSI stood at 66.29 - bullish momentum but knocking on overbought, having been pushed up by the September 8 AWS gap. Price was 9.06% above the 20-day SMA, 7.93% above the 50-day and 8.25% above the 200-day, resolving to roughly $166.9 / $168.6 / $168.1 respectively - an unusually tight cluster of all three moving averages in the $167-$169 zone. That cluster is the tell: the stock spent months consolidating around $167-169, broke out on the AWS news, and that band is now the first real support. Entry-timing implication: chasing 9% above the 20-day into RSI 66 is poor risk/reward; a pullback toward $167-170 would offer a far cleaner entry with a defined stop. Also note the shape: -14.05% over three months against +40.17% over six means this is a recovery rally off a sharp correction, not a clean established uptrend.

• Memory price inflation and supply constraints squeezing Android handset economics - the brief estimates a ~20% drag on Qualcomm's Android chip revenue in 2026, a cost-driven, not demand-driven, hit.
• US semiconductor tariff proposals raising hardware input costs, partially offset by Qualcomm scaling AI R&D and supply-chain footprint in Vietnam.
• AI data-center capex supercycle rotating toward custom ASIC/custom silicon (the AWS win is a direct beneficiary; AVGO/MRVL trade at 28-32x forward vs Qualcomm at ~17x).
• High beta (1.679) and semiconductors sector-rotation sensitivity - a single-day 8-9% gap on one headline cuts both ways if the AWS ramp disappoints.
• Handset market maturity plus China/geopolitical exposure keeps a structural discount on the legacy QCT business regardless of how the data-center pivot goes.

• FQ4/FY26 earnings in late October/early November 2026 - first hard read on FY26 close-out and, critically, concrete Q1 FY27 guidance on early data-center revenue and QCT gross-margin trajectory.
• AWS custom AI silicon and optical networking milestones - warrant vesting on up to 25M shares at $161.26 is tied to deployment, so ramp progress is both a revenue and a dilution event.
• Automotive: record $1.6B FQ3-26 revenue, +61% YoY, with BMW naming Qualcomm lead compute provider for cockpits and automated driving through the next decade.
• AI PC commercialization: Horizon Ultra with the Snapdragon X2 Elite launching mid-September 2026, plus the $40B fiscal 2029 non-handset target ($15B data center, $24B+ auto/IoT) as a re-rating framework if management delivers.
• Potential segment-disclosure or multiple re-rating if data-center revenue becomes large enough to be valued separately from handsets.

• The Apple modem roll-off is structural, not cyclical: fiscal 2027 Apple revenue is tracking below the prior $2B expectation, and that is high-margin revenue leaving permanently.
• Android chip revenue down an estimated 20% in 2026 on memory price inflation and supply constraints - a squeeze Qualcomm does not control.
• Cash conversion cracked in FQ3-26: operating cash flow of $991M (-65.5% YoY) and free cash flow of $495M, the weakest quarter in the twelve shown; a repeat would undercut the 'cash machine funds the pivot' thesis.
• Data-center margin dilution - management guides QCT weighted gross margin down 1.5-2 points near term as lower-margin custom silicon scales, so revenue growth may not translate into earnings growth.
• Earnings quality is distorted both ways by one-offs (FQ2-26 EPS $6.88 on $2.309B operating income; FQ4-25 EPS -$2.77 on a large charge), making the 20.7x trailing P/E an unreliable anchor.
• Warrant dilution of up to 25M shares (~2.4% of ~1.06B shares) tied to AWS milestones, plus US semiconductor tariff and China exposure.

Qualcomm is a fair-priced 'show me' story, not a bargain and not a broken business. The legacy handset engine is genuinely shrinking - revenue -4.0% YoY, operating margin down ~10 points YoY to 16.3%, weakest free cash flow in three years at $495M in FQ3-26 - while the stock has already rallied 40% in six months on the AWS custom-silicon pact and the raised fiscal 2029 non-handset target. At $181.97 / 17.3x forward EPS, you are paying a below-peer multiple for real optionality but also underwriting a consensus model that assumes ~20% EPS growth on declining revenue, which requires a margin recovery that has not appeared for three straight quarters. The honest position is balanced: no reason to chase an 8-9% gap into RSI 66, and every reason to wait for the late-October FQ4 report to see if data-center revenue and QCT gross margin actually begin to inflect. If they do, the discount to Broadcom/Marvell closes; if they don't, the forward EPS estimate is the thing that breaks.

Forward P/E method: 18.5x applied to next-fiscal-year consensus EPS of $10.544 gives $195, reduced to $193 for roughly 1% dilution from the 25M AWS warrants (2.4% of shares, partly offset by $161.26 exercise proceeds) and for the fact that the consensus EPS requires margin recovery not yet visible. The 18.5x is deliberate: just below Qualcomm's own 5-year average trailing P/E of 18.89, and a structural discount to Broadcom/Marvell at 28-32x forward, warranted by handset concentration and the permanent Apple modem loss. Cross-check: at $193 the stock is ~5.5% FCF yield on TTM free cash flow and ~18.3x forward EPS, in line with the low end of the analyst target range ($204-$213 average, $120-$400 spread).