Qualcomm’s Record Quarter Meets an Old Nemesis: the Memory Cycle

Record earnings overshadowed by DRAM squeeze

Qualcomm opened its fiscal 2026 with numbers that, on paper, would normally signal an inflection point. Revenues hit an all‑time high of $12.3bn, powered by record performance in its chip division (QCT) and its automotive franchise, while non‑GAAP earnings per share reached a record $3.50, at the top end of guidance. Licensing (QTL) revenue of $1.6bn came in at the high end of expectations, with an EBITDA margin of 77%.

Yet the tone of the prepared remarks was less like a victory lap and more like a weather report. Cristiano Amon, chief executive, and Akash Palkhiwala, chief financial officer, framed the quarter as a strong operational showing colliding with an external constraint: a tightening global memory market, as DRAM manufacturers divert capacity toward high‑bandwidth memory (HBM) for AI data centres.

Handset demand, they stressed repeatedly, is not the problem. Qualcomm sees robust sell‑through, particularly in premium and high‑tier devices, and a solid macro backdrop. But as DRAM supply is redirected and pricing rises, handset original equipment manufacturers—especially in China—are trimming build plans and drawing down chipset inventories. That caution is embedded in Qualcomm’s March quarter guidance.

For the second fiscal quarter, the company forecasts total revenue of $10.2bn to $11bn and non‑GAAP EPS of $2.45 to $2.65, a step down from the record December period. Within that, QTL is guided to $1.2bn–$1.4bn and QCT to $8.8bn–$9.4bn, with QCT EBITDA margins softening to 26%–28% as the unit absorbs lower scale and a modest uptick in operating expenses.

The sharpest cut lands in handsets: QCT handset revenue is expected to drop to around $6bn in the March quarter, a pronounced reset that management attributes “100%” to memory availability and pricing rather than any deterioration in end‑market demand or Snapdragon competitiveness.

Handsets: strong demand, constrained by silicon from another aisle

The contrast between Qualcomm’s handset narrative and its guidance is stark. In the December quarter, QCT handset revenues hit a record $7.8bn, riding the wave of newly launched flagship smartphones and a broader shift toward premium devices. Amon spoke of “continued expansion of the premium and high‑tier smartphone segments” and highlighted broad OEM adoption of Snapdragon platforms, including dual‑flagship strategies at major brands.

Qualcomm expects to maintain roughly 75% share in Samsung’s upcoming premium device family, in line with prior expectations—an important anchor as Chinese OEMs retrench. Licensing revenue, which provides a window into overall handset unit volumes and mix, also reflected better‑than‑expected global demand in the December quarter, particularly at the high end.

Against that backdrop, the March quarter handset guide to ~$6bn represents more than a seasonal lull. Management described a handset industry whose “overall scale” in fiscal 2026 will “be defined by the availability and pricing of memory, particularly DRAM.” As memory suppliers pivot fabs toward HBM to satisfy AI data‑centre demand, the residual DRAM pool for consumer electronics is, by Qualcomm’s reckoning, shrinking year‑on‑year.

The impact is uneven. Amon expects OEMs to prioritise premium and high‑tier devices, which have historically proved more resilient to component price inflation and can better absorb bill‑of‑materials pressure. That tilt plays to Snapdragon’s strengths and underpins Qualcomm’s confidence that its handset revenue run‑rate and growth trajectory will normalize once DRAM constraints ease.

Crucially, Qualcomm does not purchase the bulk of handset memory itself—OEMs do—limiting its direct exposure to DRAM cost swings. But as OEMs recalibrate build plans to match their own memory allocations, Qualcomm’s chip shipments inevitably follow. The company’s platform architecture, with multi‑generation memory controllers and qualification across a wide field of DRAM suppliers (including up‑and‑coming Chinese vendors), gives it some flexibility to support whatever memory is available. It cannot, however, conjure bits that do not exist.

Palkhiwala suggested investors use the March handset revenue guidance as a reasonable proxy for the June quarter under current supply assumptions, implying that the DRAM squeeze may linger at least through mid‑year.

Automotive and IoT: diversification thesis gains substance

If memory is the spoiler in handsets, it is the supporting actor in Qualcomm’s diversification story. Automotive and IoT—strategic pillars in the company’s 2029 revenue targets—delivered record and growing contributions in the December quarter and are expected to accelerate further.

Automotive revenue reached a record $1.1bn in Q1, up 15% year‑on‑year, driven by expanding adoption of Qualcomm’s Snapdragon Digital Chassis across infotainment, connectivity and advanced driver‑assist systems (ADAS). For the March quarter, management guided to more than 35% year‑on‑year growth in automotive, an acceleration that underscores how new vehicle launches and share gains are beginning to translate into tangible top‑line heft.

Qualcomm announced a letter of intent for a long‑term supply agreement with Volkswagen Group, covering multiple brands including Audi and Porsche, and spanning a range of vehicle segments and price tiers. Under the intended deal, Qualcomm would become the group’s primary technology provider for its software‑defined vehicle architecture, developed with Rivian, and collaborate on highly automated driving systems via VW’s alliance between Cariad and Bosch.

The quarter also featured expanded collaborations with a roster of Asian automakers and EV players—Hyundai, Li Auto, Zeeker, Great Wall Motor, NIO and others—bringing total design wins for Snapdragon Ride Elite platforms to 10 programs. Amon emphasised that Qualcomm’s automotive revenue trajectory is driven less by industry‑wide auto volumes and more by incremental content and share: the company is not simply “moving with the industry” but with its own pipeline of design wins.

IoT revenue, meanwhile, rose 9% year‑on‑year to $1.7bn, with growth in consumer and networking products, and is guided to grow by a low‑teens percentage in the March quarter versus the year‑ago period. Behind that number is a widening portfolio at the industrial edge. Qualcomm folded in Algenxx to augment its low‑power AI image signal processing, and launched new Dragon Wing processors for security drones, smart cameras and industrial vision, as well as the Dragon Wing IQX series to enter the industrial PC market with edge‑AI‑ready compute for PLCs and factory‑floor controls.

The company also used CES 2026 as a platform to announce a formal expansion into advanced robotics, positioning its Dragon Wing IQ 10 series as a general‑purpose robotics architecture spanning perception, motion planning and on‑device AI for household, industrial and humanoid robots. Early engagements with players such as KUKA Robotics and other robotics and humanoid platform developers suggest Qualcomm aims to transplant its automotive autonomy and safety‑grade silicon experience into what it casts as “physical AI.”

PCs and data centres: a broader AI canvas

Beyond phones, cars and factories, Qualcomm is leaning heavily into AI‑enabled computing across PCs and data centres, trying to leverage its power‑efficient compute heritage into new form factors.

In PCs, the company introduced Snapdragon X2 Plus, expanding its second‑generation platform into the enterprise and commercial segment. Powered by the third‑generation Qualcomm Orion CPU and an upgraded Hexagon NPU, the X2 Plus is pitched as delivering substantial single‑ and multi‑core gains versus prior competition, along with large NPU and GPU inferencing advantages. At CES, 18 Snapdragon‑powered PCs were unveiled by ASUS, HP, Lenovo and Microsoft; the ASUS Zenbook A16, featuring the Snapdragon X2 Elite Extreme, was singled out for its 18‑core Orion CPU, 80 TOPS NPU and claimed 21‑hour battery life in a 16‑inch ultralight form factor. Qualcomm remains on track to commercialise 150 Snapdragon X‑powered PCs this calendar year.

In the data centre, Qualcomm reiterated its intention to be a specialist in power‑efficient AI inference, particularly in disaggregated architectures where compute and memory can be tailored to specific workloads such as decode. While public customer disclosures remain sparse—Humane, a next‑generation AI device company, is the only named user at this stage—Amon said Qualcomm is engaged with “leading hyperscalers, cloud service providers, sovereign AI projects and other global partners,” and continues to receive positive feedback on its CPU and AI processing roadmap.

The company closed two notable acquisitions during the quarter: AlphaWave Semi, adding high‑speed wired connectivity for next‑generation data‑centre platforms; and Ventana Micro Systems, buttressing Qualcomm’s RISC‑V capabilities and underpinning a high‑performance RISC‑V CPU roadmap for data‑centre workloads alongside its ARM‑compatible Orion line. Both deals will nudge operating expenses higher—non‑GAAP opex is guided to approximately $2.6bn in the March quarter, with the sequential increase attributed to calendar resets, employee‑related costs and AlphaWave integration—but are framed as strategic fuel for a “multi‑billion” revenue opportunity in data centres within a few years.

Balance sheet and capital returns: discipline amid investment

Despite the growing diversification and M&A activity, Qualcomm continues to foreground capital return and cost discipline. In the December quarter, it returned $3.6bn to shareholders, including $2.6bn of share repurchases and $949m in dividends. QCT’s 31% EBT margin, above its long‑term 30% target, was achieved even as the company leaned into automotive, IoT, PC and data‑centre investments.

Palkhiwala reiterated the operating framework: reduce investment in mature businesses and redeploy into diversification priorities, while allowing acquisitions like AlphaWave to add targeted expense where they support future growth vectors such as AI data centres. Over recent years, Qualcomm has grown operating expenses significantly slower than revenue and gross profit; the company signalled no change to that philosophy even as it rides the AI capital‑expenditure wave.

Management tone: confident on fundamentals, realistic on constraints

If there is a single thread running through the prepared remarks, it is a determination to separate cyclical constraints from structural trajectory. On the one hand, Qualcomm faces a familiar semiconductor reality: a key component, made by someone else and consumed by its customers, is in short supply, and that bottleneck is likely to cap near‑term revenue in its largest end market. On the other hand, the company is starting to show more tangible scale in the businesses it has long touted as its future: automotive, industrial IoT, PCs, robotics and data‑centre AI.

Amon closed by acknowledging the frustration of having “everything in the business” firing only to be “impacted by memory,” but doubled down on Qualcomm’s 2029 diversification commitments and its emerging role in what he calls “physical AI,” where edge intelligence migrates into robots, vehicles and other machines. The next investor event, he promised, will bring more detail on roadmaps across robotics, automotive, industrial IoT and 6G, as well as the nascent data‑centre CPU and AI platforms.

For investors, the immediate calculus is awkward: record results and robust demand bumping up against an exogenous DRAM squeeze with an unclear end date. But the longer‑term question—the one Qualcomm is increasingly trying to answer with hard numbers rather than slides—is whether the company can credibly transform from a handset‑centric chipmaker into a diversified edge‑AI platform supplier across multiple, structurally growing markets. On this quarter’s evidence, the diversification is starting to move from narrative to income statement, even as the memory cycle reminds everyone that, in semiconductors, the macro context often writes its own script.