QCOM Sees Meaningful Revenue

W ill deliver “multiple billions of revenue” starting fiscal ’27, which he clarified means starting this calendar year - Next five-year strategy centers on three dimensions: building a data center platform, becoming a full-stack physical AI player at the edge, and moving from silicon to integrated platform solutions. - No explicit numerical financial guidance was provided in the excerpt; management said CFO Akash Palkhiwala would later give “important updates” to Qualcomm’s financial outlook. - Data center is positioned as a major new growth pillar: Qualcomm says it has been executing a “submarine strategy” by collecting assets and now believes it has a comprehensive portfolio to enter the next phase of data center compute. - Management argues it is “never too late for Qualcomm” in data center despite a crowded market, citing rapid market evolution, inference scaling and disaggregation as creating room for technology-led entrants. - Qualcomm sees a “really big opportunity” over the next 3–5 years in industrial AI, robotics and physical AI, building on its broad edge-compute semiconductor portfolio. - Company is reframing IoT into more distinct categories including personal AI/compute, industrial networking and robotics; wearables are expected to evolve into personal AI devices. - Mobile, which had “gone out of favor” as a mature smartphone market, is expected to become more interesting again as agents drive new edge-device opportunities, including 6G. - Modular acquisition is part of Qualcomm’s push to become more software- and developer-centric, with management emphasizing hardware, software and developer ecosystem integration. - Strategic message: Qualcomm believes it has moved from a mobile-focused semiconductor company into a diversified edge leader across automotive, personal AI/compute, industrial networking and robotics, and is now starting a new execution chapter. - Agentic AI is driving “token counts” sharply higher and “CPU attach rates” “soaring through the roof,” with management saying traditional infrastructure “will not scale” to demand. - Plans to launch an "Orion" server-class compute solution in mid-2028, completing its agentic infrastructure roadmap - Will launch its third-generation AI accelerator in 2027, calling it the industry's first "near-compute" AI accelerator for inference. - Broken through the memory bottleneck" in its XPU design by putting the XPU under the DRAM stack, claiming lower power, less heat and no need for the expensive silicon interposer used in HBM solutions. - Microsoft to deploy QCOM's high bandwidth compute chip a Azure data centers - Introduced C1000, a data-center CPU fleet with cores running "greater than 5 gigahertz" and "more than 30% faster than any of the competition," plus "more than 250 cores" and over "2 terabytes" of I/O bandwidth. - Meta plans to used c1000 CPU in its data centers -First-generation 800-gig electrical and optical DSPs, including coherent light, are already in production. - Data center revenue is now expected to start in fiscal ’26; management said the business was originally modeled for fiscal ’28, then moved to ’27, then to ’26 as traction accelerated. - Revenue and EPS are going to grow “much faster than what we had told you before,” with data centers changing the business mix “radically” versus prior expectations. - Management says Qualcomm is becoming “a computing company” and is moving from a devices company to “a cloud and device company.” - Revenue and EPS are going to grow “much faster than what we had told you before,” with the business mix changing “radically” as data center ramps. - CEO said Qualcomm has moved from a fiscal ’28 data-center revenue model to ’27, and then to ’26, because traction accelerated faster than expected. - Management said capital return and capital allocation remain “consistent with what we’ve told you before.”