Broadcom’s AI custom silicon drives a record quarter, while Qualcomm ties growth to memory conditions and a June investor day

Broadcom and Qualcomm are both positioning for long-term growth tied to artificial intelligence, but their most recent quarters offered investors two noticeably different pictures of where the demand is showing up first. In Broadcom’s case, the company highlighted accelerating AI semiconductor revenue tied to custom chips for hyperscalers. Qualcomm’s latest update, by contrast, emphasized that it is still navigating a challenging memory environment, even as it described early progress entering the data-center market with a hyperscaler’s custom silicon engagement.

Broadcom said it delivered record results in its second quarter of fiscal 2026 ended May 3, 2026, with consolidated revenue up 48% year over year to $22.2 billion. The company also reported record adjusted EBITDA of $15.2 billion, and free cash flow of $10.262 billion. Management attributed the performance to “accelerating growth” in AI semiconductor revenue, which it defined as semiconductor revenue from AI.

In that same period, Broadcom reported AI semiconductor revenue of $10.8 billion, up 143% year over year. The company linked the acceleration to “increasing demand for custom AI accelerators and AI networking.” Those terms refer to application-specific chips and the high-speed interconnect products that move data between AI servers in data centers. Looking forward, Broadcom guided that AI semiconductor revenue is expected to grow more than 200% year over year in the third quarter to $16.0 billion, and it projected overall third-quarter revenue of about $29.4 billion.

Broadcom’s results also came alongside its software cash engine. Revenue for its infrastructure software segment, anchored by VMware, was $7.178 billion in the quarter, as Broadcom reported total segment results of $15.009 billion for semiconductor solutions and $7.178 billion for infrastructure software. The company also declared a quarterly dividend of $0.65 per share, payable June 30, 2026 to shareholders of record June 22.

Qualcomm reported results for its fiscal second quarter ended March 29, 2026. The company posted revenues of $10.6 billion and GAAP diluted earnings per share of $6.88, alongside non-GAAP EPS of $2.65. Management framed the quarter as solid execution “in line with” guidance while acknowledging a “challenging memory environment,” a reference to how semiconductor supply and pricing pressures can affect handset and related demand.

Segment results showed the continuing shape of Qualcomm’s portfolio. Qualcomm’s QCT (Qualcomm CDMA Technologies) revenue totaled $9.076 billion, with handsets at $6.024 billion (down 13% year over year), automotive at $1.326 billion (up 38%), and IoT (internet of things) at $1.726 billion (up 9%). As part of return-of-capital activity, Qualcomm said it returned $3.7 billion to stockholders during the quarter, including $945 million of dividends and $2.8 billion through repurchases of 19 million shares.

On the AI and data-center front, Qualcomm said it is entering the data center and that a “leading hyperscaler custom silicon engagement” is on track for initial shipments later this calendar year, with further growth details slated for its Investor Day on June 24. The company’s outlook also pointed to near-term sensitivity: it forecast third-quarter revenue of $9.2 billion to $10.0 billion and described an expectation that QCT handset revenues from Chinese customers will bottom in the third quarter and resume sequential growth in the following quarter. What remains unclear for investors is how quickly Qualcomm’s data-center progress will translate into meaningful revenue, because the quarter’s release did not provide shipment volumes, pricing, or customer-specific capacity details.

The market read-through is that Broadcom is currently translating AI buildout spending into reported revenue at a faster clip, while Qualcomm’s AI upside is more dependent on macro conditions for mobile and on later disclosure of how its custom data-center silicon ramps. Broadcom also appears to be leaning harder into fully custom components, while Qualcomm is balancing diversification with a narrower set of disclosed AI monetization milestones. Investors following both companies next may focus on whether Broadcom can sustain its AI revenue trajectory in coming quarters, and whether Qualcomm’s June 24 Investor Day offers more concrete economics for its data-center and “physical AI” plans, as well as any updated timeline for normalization in handset markets.