Woofun AI reports that Qualcomm held its 2026 Investor Day in New York on June 25, where CEO Cristiano Amon unveiled a comprehensive roadmap for data center AI infrastructure alongside a significant revision of long-term financial targets. The event centered on the introduction of the Dragonfly C1000 CPU, the AI300 inference accelerator, and High Bandwidth Computing (HBC) technology, marking a strategic pivot toward enterprise computing. Qualcomm simultaneously announced multi-generation collaborations with Meta Platforms, expanded partnerships with Hugging Face, and the acquisition of AI software firm Modular for approximately $3.9 billion in shares. These moves collectively signal a transformation from a mobile-centric entity to a diversified AI infrastructure provider.

The financial implications of this strategic shift were immediately reflected in market performance, with Qualcomm's stock price surging 16% in after-hours trading following the announcement. CFO Akash Palshivala raised the revenue target for the company's non-smartphone business in fiscal year 2029 to $40 billion, nearly doubling the previous long-term goal of $22 billion established in 2024. Within this expanded projection, the data center business alone is expected to generate over $15 billion in annual revenue by fiscal year 2029. Palshivala further predicted that the data center segment would contribute 'billions' of dollars in revenue as early as fiscal year 2027. This aggressive guidance stands in stark contrast to the projected revenue structure for the smartphone business, which is expected to account for only about one-third of the QCT division's total revenue by 2029.

A detailed breakdown of the revenue composition reveals the specific growth engines driving this $40 billion target. The automotive business is projected to generate $10 billion in revenue by fiscal year 2029, The Internet of Things (IoT) sector is expected to contribute over $14 billion, a figure derived from two distinct sub-segments: industrial applications, networking solutions, and robotics targeting $8 billion, and personal AI and computing solutions aiming for $6 billion. These figures represent a fundamental restructuring of Qualcomm's portfolio, where the traditional smartphone dominance is being systematically diluted by high-growth enterprise and automotive verticals.

Profitability targets were adjusted upward in tandem with revenue projections, creating a significant divergence between company guidance and analyst expectations. While the average analyst forecast for Qualcomm's adjusted earnings per share in fiscal year 2029 stands at $15.26, the company has set an internal target of over $18 per share. This gap of nearly $3 per share directly catalyzed the sharp increase in the stock price observed immediately after the event. CEO Cristiano Amon attributed this growth logic to a fundamental shift in AI application patterns, noting that the industry is moving from simple question-and-answer systems to agent applications capable of independently executing multiple task steps. Amon emphasized that these new workloads require lower-power computing capabilities, a domain where Qualcomm's mobile chip heritage provides a distinct competitive advantage. He further noted that AI computing is rapidly penetrating the automotive, consumer electronics, and robotics sectors, ensuring sustained demand for chips in these areas.

The hardware centerpiece of the announcement was the Dragonfly C1000, a CPU specifically engineered for data center environments. Built on a custom-designed Oryon core and utilizing a Chiplet architecture, the processor integrates over 250 cores and operates at frequencies exceeding 5 GHz. Performance tests conducted by Qualcomm indicate that the Dragonfly C1000 delivers more than twice the performance per watt compared to existing server CPU competitors. The chip supports PCIe Gen 7 and CXL connections, while its memory system leverages low-power technologies and includes built-in RAS functions such as ECC, fault isolation, and error recovery. The cooling solution is designed to be compatible with both air and liquid cooling methods, and the chassis design adheres to OCP ORv3 standards. A chassis configuration equipped with the Dragonfly C1000 will feature 43 TB of DRAM, with sample units expected to be available in fiscal year 2026.

Qualcomm has defined three primary application areas for the Dragonfly C1000 to maximize its market utility. The first category involves Agent CPUs, which are optimized for high-throughput agent orchestration and low-latency interactive AI tasks. The second category comprises General-purpose CPUs designed to meet the needs of both first-party and third-party users, offering optimal total cost of ownership (TCO) performance for proprietary workloads and superior vCPU performance for flexible third-party deployment. The third category includes AI head-node CPUs, which aim to minimize host processing overhead and enable XPU devices to achieve maximum performance in generative AI calculations. This segmentation strategy allows Qualcomm to address diverse computational requirements within the data center ecosystem.

The commercial viability of the Dragonfly C1000 was significantly bolstered by the announcement of a 'multi-year, multi-generation' agreement with Meta Platforms. Under this deal, Meta will deploy the Dragonfly C1000 in its next-generation server clusters, with mass production scheduled for the second half of 2028. Subsequent iterations of the CPU will also be included in this long-term collaboration. CFO Akash Palshivala noted that Qualcomm already maintains business relationships with almost all large enterprises through its smartphone chips and other existing products, stating, 'This is not a new relationship.' This comment implies that Meta Platforms is likely not the sole partner involved in these discussions, suggesting that additional customers may be in the negotiation process. When addressing concerns about entering the data center market late, Amon argued that factors such as scale, execution capability, engineering expertise, and operational supply chain capabilities remain critical, asserting that the large-scale systems engineering expertise accumulated during the smartphone era remains highly relevant.

Beyond the CPU, Qualcomm updated its roadmap for AI accelerators with the unveiling of the AI300 inference accelerator, following the previously released AI200 and AI250 models. These products are developed in a phased manner, adhering to a core concept of 'decoupled, rack-level AI inference.' Tony Pialis, Executive Vice President and General Manager of Data Center Business, explained that agent workloads require the coordinated efforts of CPUs, AI accelerators, and connection technologies rather than reliance on a single chip. The company is integrating computing, AI, memory, and connection technologies into a unified rack-level platform where memory management is paramount. To address the 'memory wall' bottleneck in data transfer between processors and memory, Qualcomm introduced High Bandwidth Computing (HBC). This technology utilizes 3D stacked silicon to closely integrate computing units and memory, enabling near-memory computing capabilities.

Woofun AI data shows that the performance metrics for HBC are substantial, with the AI250 equipped with HBC Gen 1 delivering an effective memory bandwidth of 133 TB/s per card, which is 18 times higher than the AI200 using LPDDR5X. The AI300 with HBC Gen 2 offers a bandwidth that is 54 times higher than the AI200. Compared to current mainstream High Bandwidth Memory (HBM), HBC provides six times the bandwidth at the same power consumption, while offering 200 times the storage capacity of SRAM at the same power level. These improvements directly impact the total cost of ownership for data centers by significantly increasing the amount of data processed per unit of power. Commercial samples of the AI250 are expected by mid-2027, while the AI300 samples will be available in 2028. In terms of connectivity, Qualcomm offers a robust portfolio ranging from Die-to-Die connections to campus-level solutions, supporting speeds of 800 Gbps and 1.6 Tbps over distances up to 20 kilometers. More than 35 technology partners, including Supermicro, Lenovo, SK Hynix, Micron, Samsung SDS, and Arista, have expressed support for this roadmap.

The software ecosystem strategy was equally aggressive, headlined by the acquisition of Modular for approximately $3.9 billion in Qualcomm shares. The deal, expected to close in the second half of 2026 pending regulatory approval, brings an open, AI-native software stack that allows models to run across diverse chip architectures including CPUs, GPUs, NPUs, and custom ASICs without code rewriting. Co-founded by Chris Lattner, Modular's platform is viewed as an open alternative to NVIDIA's CUDA. Amon stated that as agents become ubiquitous in data centers and edge devices, the industry requires a more open and modern software foundation, and this acquisition aims to provide customers with diverse deployment options.

Additionally, Qualcomm expanded its collaboration with Hugging Face across three key areas: introducing Hugging Face workloads into Dragonfly-powered data centers, enabling over 3 million open models to be directly deployed on Qualcomm platforms, and developing the 'Hugging Face Agent' to orchestrate workloads in hybrid on-device and cloud environments.

Regarding the Chinese market, Amon addressed the regulatory landscape by stating that while U.S. export restrictions on AI hardware exist, Qualcomm will offer data center chip versions that comply with these regulations. Although specific details were not provided, the remarks indicate that the company has not abandoned opportunities in the region. The comprehensive signals sent during the Investor Day, ranging from the Dragonfly C1000 specifications to the $15 billion data center revenue target and the $18 per share earnings goal, provide a tangible evidence base for Qualcomm's strategic plans. The company has clearly identified its customers, outlined product timelines, and established a financial model. The upcoming quarterly earnings reports will serve as the first critical test of these ambitious roadmaps. This marks a definitive shift in Qualcomm's corporate identity, moving from a mobile leader to a central player in the global AI infrastructure landscape.