QCOM and ARM earnings call read-through for the Asian technology supply chain: handset demand remains weak, while the AI server chain is stronger
AI summary card
QCOM and ARM earnings call read-through for the Asian technology supply chain: handset demand remains weak, while the AI server chain is stronger
J.P. Morgan believes that the drawdown of Chinese smartphone inventories will bring short-term sequential improvement, but underlying demand remains weak. In contrast, opportunities in ARM-related AGI CPUs and server CPUs, QCOM's entry into data-center custom silicon, and tight supply of memory, test equipment, substrates and leading-edge process wafers are more favorable for the Asian AI supply chain.
- QCOM believes Chinese smartphones bottomed in the June quarter and could achieve double-digit sequential growth in the September quarter, but management attributed the recovery to channel inventory destocking rather than a rebound in underlying demand.
- The report maintains a cautious stance on consumer electronics and smartphone-related companies, noting that demand weakness and chip de-speccing are also emerging in mid- to high-end smartphones.
- ARM has greater confidence in an AGI CPU opportunity exceeding US$1bn in FY27-28, with potential upside to more than US$2bn; supply tightness is the key variable.
- Server CPU demand is being driven by inference and agentic workloads. The report expects server CPU shipments to increase from 26 million units in 2025 to 68 million units in 2028, representing a 38% CAGR.
Report interpretation
Overview
This report summarizes the implications for the Asian technology supply chain following the QCOM and ARM earnings calls. The core view is that the smartphone chain is supported in the short term by inventory destocking, but end demand remains weak, while higher memory prices are increasing BOM pressure and prompting OEMs to de-specify toward lower-tier or previous-generation chips. The AI server and data-center chains offer more structural opportunities, jointly driven by the ARM AGI CPU opportunity, expanding server CPU demand, QCOM's entry into data-center custom silicon and supply bottlenecks.
Core views
The report remains cautious on the handset and consumer electronics chains, specifically noting that companies with high consumer exposure such as Xiaomi and Novatek face demand and margin pressure. By contrast, it is more constructive on Asian supply-chain companies benefiting from AI servers, the CPU supercycle and supply tightness, including TSMC, ASE, Unimicron, Samsung, Advantest, ASPEED, Lotes and Wiwynn. QCOM's data-center custom silicon revenue will begin contributing from the December quarter. Over the long term, QCOM may enter the data-center market through Alphawave IP and compete with existing design-service providers such as MediaTek, Alchip and GUC.
Analysis framework
The report combines earnings-call interpretation, supply-chain trend validation and cross-company read-through analysis. It links QCOM's comments on smartphone inventories, chip specifications, pricing and data-center custom silicon with ARM's views on AGI CPUs, server CPU TAM and inference demand, and then maps these conclusions to companies across the Asian semiconductor, packaging and testing, substrate, test-equipment, foundry and server supply chains.
Methodology notes
Deriving implications for the Asian supply chain from statements by QCOM and ARM management
The report does not merely repeat earnings results; it maps changes in inventories, demand, pricing, BOM pressure and AI server demand to upstream suppliers and the competitive landscape.
Tight supply of memory, test equipment, substrates and leading-edge process wafers supports the pricing-up narrative
The report believes supply-chain tightness will persist through 2027, which is more favorable for Asian semiconductor supply-chain companies with supply bottlenecks and pricing power.
ARM maintains its view that CPU TAM will exceed US$100bn by 2030
The report incorporates growth in inference and agentic workloads, raises or emphasizes the expansion path for server CPU demand, and expects server CPU shipments to achieve a 38% CAGR from 2025 to 2028.
Asset mapping & comparison
Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).
- QCOMOne of the core sources of the earnings-call analysis; its handset chips, pricing and data-center custom silicon all have read-through implications for the Asian supply chain
- Strengths
- Chinese smartphone inventory destocking supports short-term sequential improvement; data-center custom silicon revenue will begin contributing from the December quarter, and QCOM could eventually enter a larger market through Alphawave IP.
- Weaknesses
- Underlying demand in the handset market remains weak, and management believes the global smartphone market will decline by a low-double-digit percentage year over year in 2026.
- Comparison
- Over the long term, QCOM may compete with design-service providers such as MediaTek, Alchip and GUC.
- Risks
- If end demand continues to weaken or CSP customer expansion falls short of expectations, pricing and data-center contributions may be insufficient to offset pressure in the handset chain.
- ARMOne of the core sources of the earnings-call analysis; the AGI CPU opportunity, server CPUs and long-term CPU TAM underpin the AI supply-chain view
- Strengths
- Greater confidence in an AGI CPU opportunity exceeding US$1bn in FY27-28, with the view that CPU TAM will exceed US$100bn by 2030 unchanged.
- Weaknesses
- It also highlighted weakening demand for mid- to high-end smartphones, indicating that handset pressure is not limited to low-end devices.
- Comparison
- Growth in server CPU demand will compete for leading-edge process capacity alongside ecosystems including AMD, NVDA Vera, and Google's and Amazon's in-house CPUs.
- Risks
- Supply-chain tightness could constrain the pace of revenue realization, while competition for wafer allocation is intense.
- TSMC, ASE, Unimicron, Samsung, AdvantestBeneficiaries of the Asian AI supply chain
- Strengths
- Tight supply of memory, test equipment, substrates and leading-edge process wafers is expected to continue through 2027, supporting price increases and greater value for capacity.
- Weaknesses
- They remain exposed to cyclical demand and capital-expenditure timing; price support could weaken after supply expands.
- Comparison
- Compared with the smartphone component chain, AI server-related supply bottlenecks are stronger and visibility on the cycle is higher.
- Risks
- If AI server demand slows, customers adjust orders or capacity expands too quickly, the supply-tightness thesis could weaken.
- ASPEED, Lotes, Wiwynn, UnimicronBeneficiaries of the server CPU supercycle
- Strengths
- The report expects server CPU shipments to achieve a 38% CAGR from 2025 to 2028, driving demand for related connectors, server systems, BMCs and substrates.
- Weaknesses
- Demand depends on AI server capital expenditure and cloud-provider purchasing cycles.
- Comparison
- They are more exposed than traditional consumer-electronics companies to growth in inference and agentic workloads.
- Risks
- Lower-than-expected AI workload growth or customer inventory adjustments could affect order continuity.
- Xiaomi, NovatekCompanies explicitly identified as cautious positions in the report
- Strengths
- If demand recovers after channel inventory destocking, they could see short-term trading-driven improvement.
- Weaknesses
- Weakening demand for mid- to high-end smartphones, chip de-speccing and rising BOM pressure could weigh on revenue quality and margins.
- Comparison
- Compared with the AI server supply chain, they have greater consumer-electronics exposure and lower demand visibility.
- Risks
- Further deterioration in 2H26 demand, unsuccessful price transmission or continued gross-margin pressure.
- AMD, NVDA Vera, Google, Amazon in-house CPUsImportant sources of demand for leading-edge process wafer allocation
- Strengths
- The report believes these CPU projects are likely to receive most of TSMC's wafer allocation in 2027, demonstrating strong AI computing demand.
- Weaknesses
- They are highly dependent on leading-edge process capacity and supply-chain bottlenecks.
- Comparison
- Competition from these projects for wafer allocation could reduce capacity available for other AI accelerator or custom silicon projects.
- Risks
- If capacity is insufficient, product schedules are delayed or cloud providers change their in-house development needs, revenue realization across the related supply chain could fluctuate.
Key data
- JPMe global smartphone shipments in 2026Down 11% year over yearThe report says this is broadly consistent with QCOM's view that the global smartphone market will decline by a low-double-digit percentage year over year in 2026.
- Near-term trajectory for Chinese smartphonesBottomed in the June quarter, with double-digit sequential growth in the September quarterQCOM attributed the improvement to channel inventory destocking rather than a rebound in underlying demand.
- ARM AGI CPU opportunityMore than US$1bn in FY27-28, with potential upside to more than US$2bnThe report emphasizes that supply is the key swing factor.
- ARM CPU TAMMore than US$100bn by 2030The report says this long-term TAM view is unchanged and is supported by upside from inference and agentic demand.
- Server CPU shipment modelFrom 26 million units in 2025 to 68 million units in 2028, representing a 38% CAGRDriven by inference and agentic workloads.
- QCOM data-center custom siliconRevenue contribution begins in the December quarterBased on its review, the report believes the primary ASIC customer may be ByteDance, but it has not yet seen QCOM become a key design-service partner for AI accelerator projects at the four major CSPs.
Impact & implications
The investment implications are clearly bifurcated. The short-term rebound in the smartphone chain is driven more by inventory recovery, while insufficient end demand and BOM pressure may continue to weigh on margins. The AI server and data-center chains are supported by expanding CPU demand, tight wafer allocation and bottlenecks in key components, giving them stronger structural momentum. The report suggests that companies in the Asian supply chain with capabilities in leading-edge processes, packaging and testing, substrates, test equipment, and server connectivity/systems are more likely to benefit.
Risks
- End demand for smartphones could deteriorate further in 2H26; short-term sequential improvement from inventory destocking does not equate to a recovery in underlying demand.
- Higher memory prices are increasing BOM costs, while OEM de-speccing could pressure demand for high-end chips and supply-chain margins.
- QCOM's pricing increases may be constrained by weak demand; insufficient price transmission could prolong margin pressure.
- AI server and CPU opportunities depend heavily on the supply of leading-edge process wafers, substrates, test equipment and memory; insufficient supply could limit revenue realization.
- If inference and agentic workload growth falls short of expectations, the server CPU supercycle and ARM's long-term TAM upside thesis could weaken.
What to watch
- Whether double-digit sequential growth in Chinese smartphones in the September quarter can extend into an improvement in underlying end demand.
- Whether the smartphone supply chain sees further order cuts or a reversal in customer pull-ins in 2H26.
- Whether price increases for memory, foundry services, OSAT and substrates can be effectively passed through by IC design companies.
- The scale of QCOM's data-center custom silicon revenue beginning in the December quarter and progress in customer expansion.
- Whether ARM's AGI CPU opportunity can be realized from more than US$1bn toward more than US$2bn in FY27-28.
- Changes in TSMC's 2027 leading-edge process wafer allocation among AMD, NVDA Vera, Google, Amazon's in-house CPUs and other customers.