Compute & Infrastructure
Designs the chips and modems behind most premium smartphones; expanding into on-device AI, automotive, and PCs.
{'verdict': '2 signals sit in the elevated band: organic-demand sustainability and the AI-monetization gap. This does not trip the convergence flag. Ranks 29th of 68 on composite fragility (F\xa045.85), below Salesforce and Atlassian.', 'as_of': '2026-07-11', 'source': 'engine-restatement (T1)', 'snapshot': {'composite_f': 45.85, 'n_elevated': 2, 'convergence': 'moderate', 'rank': 29, 'elevated': ['organic-demand sustainability', 'the AI-monetization gap']}}
The FY2025 10-K says Apple began using its own modem and expects to increasingly use it, with a 'significant negative impact' on QCT revenues. CEO Amon's own timeline is ~70% iPhones on QCOM this fall, ~20% next fall, ~zero fall 2027. Futurum estimates a ~$7.3–7.8B annual shortfall at full Apple exit. Auto and IoT combined are $10.574B growing at 27% combined. How does $10.6B growing at 27% close
The 10-K confirms Apple modem displacement is in progress and expects a "significant negative impact" — not speculative. Auto is $3.957B (+36%) and IoT is $6.617B (+22%), combined $10.574B (~24% of total revenue). At 27% combined growth that adds roughly $2.9B per year — not enough to offset Futurum's estimated $7.3–7.8B modem-RF cliff at full Apple exit (Futurum — attributed, not audited). Apple-specific revenue in dollars is NOT SOURCED from Qualcomm filings (10-K gives % only, ~21% FY2025). Huawei QTL royalties ended Q2 FY2025. Diversification is executing but the math is tight by the numbers the sheet can confirm.
You paid ~$2.4B for Alphawave to accelerate data-center expansion, and $9.042B in R&D in FY2025. The 10-K says data-center AI revenues may 'take many years to generate material returns.' Data-center is a nonreportable segment with approximately zero disclosed revenue. How many years is 'many years'?
The 10-K is the source: data-center business is developing and commercializing next-generation solutions with no segment revenue line disclosed. R&D was $9.042B (20% of FY2025 revenue). The Alphawave acquisition (~$2.4B enterprise value, ~$2.323B restricted cash set aside, close expected Q1 CY2026) is confirmed from FY2025 10-K Note 9. The risk factor language — "take many years to generate material returns" — is verbatim primary. Data-center AI revenue is honestly labeled ≈$0 disclosed; no counter-claim exists from filed figures.
Qualcomm returned $8.8B in buybacks in FY2025 while facing a known, 10-K-confirmed Apple modem exit starting 2027. Is that responsible capital allocation?
The $8.8B buyback and $3.8B dividends in FY2025 are confirmed from the 10-K cash flow statement ($14.012B operating cash flow). The Apple modem supply agreement runs to 2027 — the transition is ahead, not behind. The sheet flags this as a bear-case data point: "$14.8B debt + $8.8B buybacks FY2025 while facing ~$7B+ Apple revenue cliff — capital allocation could look aggressive in hindsight." The counter: pretax EBT was $12.663B (29% margin) and QTL licensing at $5.6B is a durable cash base. The question is whether reduced cushion is a material risk if the Apple cliff arrives faster than Automotive and IoT diversification scales.