Damnang
Semiconductor Engineer in Silicon Valley, AI Chip Investment Research
Qualcomm's new HBC near-memory computing architecture, unveiled at Investor Day, is technically meaningful despite dismissals that LPDDR cannot beat SRAM or HBM. The 3D IC design fits Qualcomm's long-prepared LPDDR strategy for AI, an era even Jensen Huang acknowledges has arrived. That assumes the yield ramp is executed successfully. x.comÂ
Qualcomm is being repriced for its possible entry into data-center AI. CEO Cristiano Amon said a leading hyperscaler custom silicon engagement is on track for initial shipments later this year; the customer's anonymity points to semi-custom or fully custom chips with stronger lock-in. Automotive revenue rose 38% to a record $1.326B, with about 50% growth guided next quarter as its $45B design-win pipeline converts. Handsets, down 13%, have not bottomed, and the Q3 guide was light. x.com x.com x.comÂ
Qualcomm is becoming a fuller AI infrastructure player, and the market may re-rate it if its $5B fiscal 2027 data center revenue target is met. Most of that target comes from custom chips for one US and one Chinese hyperscaler, about $2B each at mid-30s gross margins. Its High Bandwidth Cache stacks memory on logic for about 6x HBM's bandwidth per watt, easing the inference memory wall. Alphawave's 800G DSPs already earn revenue, and a Meta custom CPU deal carries a minimum share guarantee. … Show more
Qualcomm could rerate from a handset chipmaker into an AI infrastructure story now that Amazon is co-developing multiple generations of inference silicon with it. As AI shifts toward stacks built around workloads, power and cost, Qualcomm need not beat Nvidia, only help optimize around the GPU. Amazon and Qualcomm are also working on 1.6T links using Qualcomm SerDes and optical DSPs, entering a space led by Broadcom and Marvell. Deployment volumes and further design wins would confirm it. x.comÂ
Qualcomm is unlikely to rebound meaningfully this year because on-device AI needs more mobile DRAM that phones are not getting. Industry checks suggest Apple walked back full WMCM packaging for the iPhone 18, adopted to fit more DRAM for on-device AI, using it only in high-end models with no RAM upgrade. If Apple, with the strongest buying power, faces this, Qualcomm would need an extraordinary breakthrough past the memory wall. This counters @citrini's view of it as an edge-AI winner. x.comÂ
Qualcomm is shifting from handset supplier to AI infrastructure platform, which should shrink its handset discount. At its June 24 investor day it targeted data center revenue rising from $0.3 billion in fiscal 2026 to over $15 billion by 2029, with the Dragonfly C1000 CPU and AI300 inference accelerator. Micron's results show insatiable AI memory demand that its efficient compute can serve. At about 21x forward earnings versus 30-50x for AI pure plays, a 12-month target of $325 is justified. … Show more
Qualcomm is building the network layer around its HBC memory architecture as an integrated system, attacking AI data movement at several layers at once. HBC cuts movement between memory and compute, Alphawave adds SerDes, optical and die-to-die links, and per AIwire Cornelis Networks will provide networking for Qualcomm's Dragonfly HBC accelerators, with compute in the fabric to cut stalls. Qualcomm's data-center chief Tony Pialis says moving data across the rack is now as important as compute. … Show more
Qualcomm is weak in its own capabilities while facing longer phone replacement cycles and pricier memory. Its other growth narratives are a very long way off. x.comÂ
Inference's move toward edge devices and dense, memory-hungry racks should drive demand for specialized chips, links, power delivery and cooling. McKinsey projects inference at 60% of AI computing demand by 2030. Ambarella makes on-device AI chips; Astera Labs links GPUs, memory and networks; Credo supplies short copper connections; Rambus earns from memory interfaces; Navitas reduces power waste; Modine cools dense racks; and Super Micro builds inference-specific server racks at volume. x.com … Show more
Qualcomm's investor day strengthened the long case, as its CFO raised FY2029 non-handset revenue guidance 91% to $40B. Qualcomm disclosed major orders from two hyperscalers, said its Dragonfly C1000 data-center CPU launches in mid-2028 with Meta among the first customers, introduced the AI300 accelerator and data-center interconnect, and expects $6B of annualized automotive revenue in FY2026. It also agreed to buy AI software firm Modular for $3.9B. x.comÂ
Qualcomm stays a Hold, with the target cut to $186 on handset weakness and limited near-to-mid-term data-center visibility. Fiscal Q4 2026 revenue guidance of $9.7-10.5B fell short of buy-side chatter near $11B; Android faces elevated wafer inventories and Apple modem revenue is falling faster than expected. Data-center targets of $5B in FY27 and $15B+ in FY29 were reaffirmed and AI250 has four major US cloud customers, but high LPDDR5 prices hurt its competitiveness, and CPUs are years off. … Show more
Qualcomm looks undervalued among tech stocks given its opportunities in physical AI, CPUs and an emerging data-center market. x.comÂ
